Pontecorvo Clotilde

NATO ASI Series
Advanced Science Institutes Series

A series presenting the results of activities sponsored by the NATO Science Committee,
which aims at the dissemination of advanced scientific and technological knowledge ,
with a view to strengthening links between scien tific communities.
The Series is published by an international board of publishers in conjunction with the
NATO Scientific Affairs Division.
A Life Sciences
B Physics
Plenum Publishing Corporation

London and NewYork
C Mathematical and Physical Sciences

D Behavioural and Social Sciences
E Applied Sciences
Kluwer Academic Publishers

Dordrecht, Boston and London
F
G
H
I
Springer- Verlag
Berlin Heidelberg NewYork Barcelona
Budapest Hong Kong London Milan
Paris Santa Clara Singapore Tokyo
Computer and Systems Sciences

Ecological Sciences
Cell Biology
Global Environmental Change
Partnership Sub -Series

I. Disarmament Technologies
2. Environment
3. High Technology
4. Science and Technology Policy
5. Computer Networking
Kluwer Academic
Springer-Verlag /
Kluwer Academic
Kluwer Academic
Kluwer Academic
Kluwer Academic
Publishers
Publishers
Publishers
Publishers
Publishers
The Partnership Sub-Series incorporates activities undertaken in collaboration with

NATO's Cooperation Partners, the countries of the CIS and Central and Eastern
Europe, in Priority Areas of concern to those countries.
NATO-PCO Database
The electronic index to the NATO ASI Series provides full bibliographical references
(with keywords and/or abstracts) to about 50 000 contributions from international
scientists published in all sections of the NATO ASI Series. Access to the NATO-PCO
. Database compiled by the NATO Publicat ion Coordination Office is possible in two
ways:
- via online FILE 128 (NATO-PCO DATABASE) hosted by ESRIN, Via Galileo Galilei,
1-00044Frascati, Italy.
- via CD-ROM "NATO Science & Technology Disk" with user-friendly retrieval
software in English, French and German ( WTV GmbH and DATAWARE
Technologies Inc. 1992).
The CD-ROM can be ordered through any member of the Board of Publishers or
through NATO-PCO, B-3090Overijse, Belgium.
Series F: Computer and Systems Sciences, Vol. 160
The NATO ASI Series F Special Programme on

ADVANCED EDUCATIONAL TECHNOLOGY
This book contains the proceedings of a NATO Advanced Research
Workshop held within the activities of the NATO Special Programme on
Advanced Educational Technology, running from 1988 to 1993 under the
auspices of the NATO Science Committee. The books published so far in
the Special Programme are listed briefly, as well as in detail together with
other volumes in NATO ASI Series F, at the end of this volume.
Springer-Verlag Berlin Heidelberg GmbH
Discourse, Tools,
and Reasoning
Essays on Situated Cognition
Edited by
Lauren B. Resnick
Learning Research and Development Center
University of Pittsburgh, 3939 O'Hara Street
Pittsburgh, PA 15260, USA
Roger Saljo
Department of Communication Studies
Linkoping University, Sweden
Clotilde Pontecorvo
Dipartimento di Psicologia dei Processi di Sviluppo
e Socializzazione, Universita degli Studi di Roma
"La Sapienza", Italy
Barbara Burge
Learning Research and Development Center
University of Pittsburgh, USA
Springer
Published in cooperation with NATO Scientific Affairs Division
Proceedings of the NATO Advanced Research Workshop on Discourse,

Tools, and Reasoning: Situated Cognition and Technologically Supported
Environments, held in Lucca (II Ciocco), Italy, November 2-7, 1993
LIbrary of Congress CatalogIng-In-PublIcatIon Data
Discourse. too ls. and reason ing : essays on situated cognition /

edited by Lauren B. Resnick . . . [et al.l.
p.
em. -- (NATO ASI series. Series F, Computer and system s
sciences: vol. 1601
"Publ ished in cooperation with NATO Scientific Affairs Division. "
"Proceed ings of the NATO Advanced Resea rch Workshop on Discourse.
Tools . and Reason ing : S ituated Cognition and Techno logica l ly
Supported Environments . held in Lucca (1 1 Ciocco> , Italy. November
2-7. 1993 "--T.p. verso.
Includes b ibliographica l references and indexes.
1. Cognition--Congresses . 2. Cognitive learning theory-Congresse s. 3 . Cognit ion and culture--Congresses. 4 . Learning .
Psycho logy nf--Congresses.
I. Reslll ck , Lauren B.
II. North
Atlant ic Treaty Organization. Scient ific Affair s Divis ion.
III. NATO Advanced Research Workshop on Discourse. Tools. and
Reasoning : S i t uat ed Cognition and Technologically Supported
Environments ( 1993 : II Ciocco . Italy> IV. Series : NATO ASI serie s.
Series F . Computer and sy stems sciences ; no . 160.
BF311.D5376 1997
153--dc21
97-35966
CI P
CR Subject Classification (1991): K.3, K.4

ISBN 978-3-642-08337-2
ISBN 978-3-662-03362-3 (eBook)
DOI 10.1007/978-3-662-03362-3
This work is subject to copyright. All rights are reserved, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other way,and storage in data
banks. Duplication of this publication or parts thereof is permitted only under the provisions
of the German Copyright Lawof September 9,1965 , in its current version, and permission for
use must always be obtained from Springer-Verlag.Violations are liable for prosecution under
the German Copyright Law.
Springer-Verlag Berlin Heidelberg 1997
Originally publi shed by Springer -Verlag Berlin Ileidelberg NewYork in 1997.
So ftcover reprint of the hardcover Ist edition 1997
Typesett ing: Camera-ready by editors

Printed on acid -free paper
SPIN: 10645153
45/3142 - 5 432 1 0
Preface
Not long ago, projections of how office technologies would revolutionize the
production of documents in a high-tech future carried many promises . The paperless office and the seamless and problem-free sharing of texts and other work
materials among co-workers were just around the corner, we were told. To anyone
who has been involved in putting together a volume of the present kind, such
forecasts will be met with considerable skepticism, if not outright distrust. The
diskette, the email, the fax, the net, and all the other forms of communication that
are now around are powerful assets, but they do not in any way reduce the flow of
paper or the complexity of coordinating activities involved in producing an artifact
such as a book. Instead, the reverse seems to be true. Obviously, the use of such
tools requires considerable skill at the center of coordination, to borrow an
expression from a chapter in this volume. As editors, we have been fortunate to
have Ms . Lotta Strand, Linkoping University, at the center of the distributed
activity that producing this volume has required over the last few years. With her
considerable skill and patience, Ms. Strand and her work provide a powerful
illustration of the main thrust of most of the chapters in this volume : Practice is
a coordination of thinking and action, and many things had to be kept in mind
during the production of this volume. We are all very grateful for her willingness
to engage in this project in the midst of all her other duties .
The volume is the result of a seminar entitled Discourse, Tools, andReasoning:
Situated Cognition and Technologically Supported Environments, which the
editors organized in Italy. The seminar was made possible with financial support
from the NATO Special Programme Panel on Advanced Educational Technology .
We are grateful to the Programme Director L. Veiga da Cunha for this
opportunity to explore important issues of technology and learning .
We thank Barbara Kester of the International Transfer of Science and
Technology office for her help in the early stages of planning the seminar and
locating a suitable site.
Finally, we also thank Bruno Giannasi, the overseas conference coordinator at II
Ciocco, for his cooperation and attention to details before and during the seminar.
Contents
List of Contributors........... ............ ...................................................

Discourse, Tools, and Reasoning
Lauren B. Resnick, Clotilde Pontecorvo, and Roger Siiljo
XI
..
Part One
Distributed Cognition:
Discourse and Activity in Complex Work Environments
Chapter I Constructing Meaning from Space, Gesture, and Speech
Edwin Hutchinsand Leysia Palen.. ...... ..... ............. .............. ..... ...... .....
23
Chapter 2 Centers of Coordination: A Case and Some Themes

Lucy Suchman......................... ................................................... .....
41
Chapter 3 Animated Texts: Selective Renditions of News Stories

Christian Heath and Gillian Nicholls. ...... ..... .... ... ..... ....... ................... ..
63
Chapter 4 To Resolve a Technical Problem Through Conversation

Alain Trognonand Corinne Grusenmeyer... ....... ....... .... ........ ..... ... .... .....
87
Chapter 5 The Blackness of Black: Color Categories as Situated Practice

CharlesGoodwin..... ..................... ....... ..... .. .... ................. ..... ............ 111
VIII
Contents
Part Two
Negotiating Identities:
The Construction of Sociocognitive Communities
Chapter 6 Reasonable Uncertainties: Parents ' Talk About
Caring for Children with Chronic Renal Failure
David Middleton ...... ...... .... ......... ............ ......... ................................. 143
Chapter 7 Syncretic Literacy in a Samoan American Family
AlessandroDurantiand Elinor Ochs................. ............................ ........ 169
Chapter 8 Other Voices, Other Minds:
The Use of Reported Speech in Group Therapy Talk
Alessandra Fasulo... ... .... ........... ...... ........... ...... .................... ..... ... ... .. 203
Chapter 9 Situational Effects in Computer-Based Problem Solving
Paul Light and Karen Littleton... .... ........... ................ ............... ........... 224
Part Three
Learning in Practice:
How People and Tools Shape One Another
Chapter 10 Discourse and Development: Notes from the Field
Joseph Glick... ...................... .. ...... .............. ......... ........... ... ... ... .... ... 243
Chapter 11 Interactional Perspectives on the Use of the Computer and on
the Technological Development of a New Tool: The Case of Word Processing
Michele Grossenand Luc-Olivier Pochon.. .......... .. ..... ..... ...... .... .... ........ 265
Chapter 12 What Organizes Our Problem-Solving Activities?
Terezinha Nunes............. .. ............. ........ ....... .................. .... ... .. ..... .. .. 288
Chapter 13 Understanding Symbols With Intermediate Abstraction s:
An Analysis of the Collaborative Construction of Mathematical Meaning
Baruch B. Schwarz.i....
312
Chapter 14 Strategy-Specific Information Access in Knowledge

Acquisition from Hypertext
Wolfgang Schnotz ........ ...... ............. ................ ........... ...... .... ... ..... ... . 336
Contents
Part Four
IX
Accountable Talk: Learning to Reason
Chapter 15 Talking About Reasoning: How Important Is the Peer

in Peer Collaboration?
Stephanie D. Teasley
361
Chapter 16 Seeing the Light: Discourse and Practice in the Optics Lab
Roger Siiljo and Kerstin Bergqvist
385
Chapter 17 Learning to Argue in Family-Shared Discourse:

The Reconstruction of Past Events
Clotilde Pontecorvo and Alessandra Fasulo.... ...
406
Chapter 18 Discourse in the Adult Classroom :

Rhetoric as Technology for Dialogue
Juan D. Ramirez and James V. Wertsch
443
Author Index
459
Subject Index .... ........ ......... ........ .................. ......... ....... ... .. .... .. .. ..
469
List of Contributors
Kerstin Bergqvist, Department of Communication Studies, Linkoping University,

Sweden
Alessandro Duranti, Department of Anthropology, UCLA, CA, USA
Alessandra Fasulo, Dipartimento di Psicologia dei Processi di Sviluppo e
Socializzazione, Universita degli Studi di Roma "La Sapienza," Italy
Joseph Glick, City University of New York, Graduate Center, NY, USA
Charles Goodwin, Department of Anthropology, University of South Carolina,
SC, USA
Michele Grossen, Institut de Psychologie, Universite de Neuchatel , Switzerland
Corinne Grusenmeyer, Research Group on Communication, Nancy II University
and French National Research and Safety Institute, Nancy-Vandoeuvre, France
Christian Heath, University of Nottingham and King's College, London , United
Kingdom
Edwin Hutchins, Department of Cognitive Science, University of California, San
Diego, CA, USA
Paul Light, Department of Psychology, University of Southampton, United
Kingdom
Karen Littleton, Faculty of Social Sciences, Psychology, The Open University,
Milton Keynes, United Kingdom
David Middleton, Department of Human Sciences, Loughborough University,
United Kingdom
Gillian Nicholls, Middlesex University, United Kingdom
Terezinha Nunes, Institute of Education, University of London , United Kingdom
XII
List of Contributors
Elinor Ochs, Department of Applied Linguistics, UCLA, CA, USA

Leysia Palen, Department of Information and Computer Science, University of
California, Irvine, CA, USA
Luc-Olivier Pochon, Institut Romand de Documentation et de Recherches
Pedagogiques (IRDP), Universite de Neuchatel, Switzerland
Clotilde Pontecorvo, Dipartimento di Psicologia dei Processi di Sviluppo e
Socializzazione, Universita degli Studi di Roma "La Sapienza", Italy
Juan D. Ramirez, Laboratory of Human Activity, University of Sevilla, Spain
Lauren B. Resnick, Learning Research and Development Center, University of
Pittsburgh, Pittsburgh, PA, USA
Roger Saljo, Department of Communication Studies, Linkoping University,
Sweden
Baruch B. Schwarz, Hebrew University, Jerusalem, Israel
Wolfgang Schnotz, Department of Psychology, University of Landau, Germany
Lucy Suchman , Xerox PARC, Palo Alto, CA, USA
Stephanie D. Teasley, Colaboratory for Research on Electronic Work (CREW),
University of Michigan, Ann Arbor, MI, USA
Alain Trognon, Research Group on Communication, Nancy II University and
French National Research and Safety Institute, Nancy-Vandoeuvre, France
James V. Wertsch, Department of Education, Washington University, St. Louis,
MO , USA
Discourse, Tools, and Reasoning:

Essays on Situated Cognition
Lauren B. Resnick, Clotilde Pontecorvo, and Roger SlUjo
In 1990, we were asked by the Scientific Affairs Division of NATO I to organize a

conference on situated cognition and technologies of learning. In planning the
conference, held in Lucca, Italy, in November, 1993, and in recruiting
participants, we sought to bring together people from several scholarly
disciplines, some of whom might not yet have known each other's work. We
needed to explain to them-and, by extension, to the several scholarly
communities of which they were members-what we had in mind and why we
thought the effort was worthwhile . The terms in which we did so were these:
Recent theories of situated cognition are questioning the view that cognition can
be understood independently of the social, organizational, and material context in
which it is practiced. Sharing with Soviet-origin activity theory an
antifunctionalist point of view in which intentionality and affect are viewed as
components of activity, Western European and North American theories of
situated cognition challenge the dominant view in cognitive science that assumes
a cognitive core can be found that is independent of context and intention.
Instead, these theorists argue, every cognitive act must be viewed as a specific
response to a specific set of circumstances, and only by taking into account the
participants' construal of the situation can a valid interpretation of the cognitive
activity be made .
. . . [Because] each situation calls for and shapes cognitive processes of specific
kinds .. . the science of cognition requires studying cognitive processes in the
many specific environments in which people think , reason , and act. It is this view
that unites a diverse set of scholars who constitute a potential research community
focused on situated cognition . [These individu als] . . . come not only from several
'Our invitation came from the Panel on Advanced Educational Technology.
Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

countries but also from several scholarly disciplines, including psychology,
linguistics, anthropology, sociology, philosophy, and didactics. They use the
research tools and language of these several disciplines and publish in widely
dispersed journals that often reach only specialized audiences .
As we planned the conference, four major topics were the focus of our attention :
the role of interactive discourse (talk and, occasionally, written documents) in the
coordination of complex cognitive activities; reasoning and learning as socially
distributed activities; social and institutional constraints on cognition; and
technologies for learning. We sought to include research on cognitive processes in
a variety of social and institutional contexts--for example, families, the medical
and social service system, schools, laboratories, and factories--and asked
participants to discuss questions of methodology and fundamental research
assumptions throughout, thus heightening the prospect of future collaboration
across disciplines .
The themes with which we began our planning represented well the kind of
thinking that characterized the loosely defined field of situated cognition five years
ago. But it is striking that, although each of those themes can be found threading
through the chapters of this volume, they are no longer the most helpful
characterizations of what this book is about. Instead, the three terms in our tiUe-discourse, tools, and reasoning--came to represent the recurrent themes of the
chapters and of our discussion.
Discourse. Discourse and interaction are dominant themes throughout this
volume . These terms are in such widespread use in psychology, sociology,
linguistics, and other disciplines that our use of them here might lead readers to
overlook the theoretically radical view of the role of discourse in human cognitive
functioning that most of our authors hold. Although many scholars treat talk and
other forms of communication as a means of transmitting ideas between
individuals or perhaps of coordinating discrete cognitive actions, our authors treat
discourse primarily as constituting cognition . In this respect, they go well beyond
Vygotsky's (1934/1978) view of speech and linguistic interaction as the primary
means by which children are socialized into a community and its forms of
thought. In Vygotsky's theory, originally interpersonal exchanges become
internalized as intrapersonal modes of thought. The same concept is expressed in
George Herbert Mead's (1934) definition of thought as conversation with the
generalized other. Vygotsky and Mead both treated interaction as the source and
origin of thinking but viewed thinking itself as, eventually, activity carried out by
the individual . Our authors, by contrast, view thought and reasoning as inherently
(and throughout the life span) social activities in which talk and social interaction
are not just a means by which people learn to think, but also how they engage in
thinking . They might say that discourse is cognition is discourse. . . . One is
unimaginable without the other.
Because discourse is publicly embodied in speech, it is easy to interpret as
action . Thus the notion that discourse and cognition are coterminous helps us to
conceive of thought as situated activity. Discourse (and, therefore, all of

cognition) is situation ally specific, adapted to the material and to social
affordances of the environment. Not only words but also accompanying gestures
and physical representations are the instruments of thought. And not only ideas
but also the intentions of the actors, the roles they play (or hope to play) vis-a-vis
one another, are vital elements in cognitive activity.
Tools. A hallmark of situated theories of cognition is the proposal that people
think not with their minds alone but assisted by--indeed, in interaction with-tools. Why should a term as familiar as tools need to be emphasized with italics?
Everyone knows what a tool is: an object (physical or material, it is usually
assumed) that is used to extend the action power of humans . A screwdriver or an
electric saw is a tool. So is a calculator or a computer (with its software). A car or
a train is a tool for moving faster than one could by walking or running . Tools
are most often thought to be designed for use as extensions of physical or
intellectual power. They are artifacts, built by humans with certain purposes in
mind (cf. Simon, 1990).
The concept of the tool is expanded here beyond the conventional view of a tool
as a physical artifact. Not only physical artifacts but also concepts, structures of
reasoning, and the forms of discourse that constrain and enable interaction s within
communities qualify as tools . Vygotsky (1934/1987) originally distinguished
tools from signs, or language . However, subsequent influential developers of
theories of socially situated cognition (e.g., Cole, 1985; John-Steiner, 1995) have
suggested that many kinds of thinking, as well as physical actions, are carried out
by means of tools . For analytical purposes , we may make a distinction between
intellectual (mental) and physic al (practical) tools; but, as will be evident when
reading the chapters of this volume, this difference will not always be necessary or
even productive . Among the intellectual tools, we find a wealth of concepts,
categories, and linguistic constructions that have emerged as cultural resources for
dealing with reality .
Some of these tools , as diverse as the concepts of democracy, the zero, or the
notion of feedback, represent major breakthroughs that have affected societies and
cultures in profound ways. They have become ways of construing reality--worldmaking, to borrow Nelson Goodman 's (1976) suggestive expression. And they are
continuously used as points of reference for arguments and knowledge
construction. Other intellectual tools are more local in their use and impact and
are restricted to specific human activities such as science, politics, economics, or
sports. The concepts of elasticity of prices or of Latin square designs have little
meaning and relevance for the broad majority of people, whereas they are powerful
tools in the reasoning and practical activities of the economist and the
experimental scientist.
Reasoning. Once we take the decisive step of no longer considering
individuals' internal mental activities as fundamentally and qualitatively different
from what takes place externally, in interaction between people and between
people and tools, we move toward a conception of cognition in which thinking is
profoundly sociocultural in the Vygotskian spirit (Wertsch, 1991a, 1991b) . With

this expanded view of tools as intellectual resources of a sociocultural community
comes a revised definition of reasoning as a fundamentally social activity in which
ideas and concepts are literally constituted in interactive discourse. Furthermore,
concepts can work as communal tools of reasoning only when they form some
kind of coherent whole . The power of intellectual tools-codified in language as
concepts , explanations, and forms of reasoning-does not reside in the capacity of
single elements to explain phenomena, but in the fact that the different constructs
form part of general modes of reasoning that are productive in human activities .
The concept of a molecule, for example, is powerful in certain contexts because it
is part of a mode of reasoning in which other conceptual tools, such as atoms ,
ions, and chemical bonding , exist and are used. In a similar vein, analytical
categories, such as adjective, inflection, and passive, form part of a coherent
theory of grammar that can be used for talking about languages .
A central learning task for individuals, on this view, is to acquire both the
organizing conceptual theories and the patterns of discourse that are used by
particular reasoning communities. Developing expertise of the kind that
characterizes skilled' blacksmiths (Keller & Keller, 1993), tailors (Lave & Wenger ,
1991), pilots (Hutchins & Palen, chapter 1, this volume), chemical refiners
(Goodwin, chapter 5, this volume) , airport operations management (Suchman ,
chapter 3, this volume), or operators of a pulp machine (Trognon &
Grusenmeyer, chapter 4, this volume) requires familiarity with a broad range of
discursive constructions that are prevalent in the activities in which these
professions operate. Some of these concepts may be very general and apply to a
broad range of contexts and human activities, whereas others are likely to be local.
From this perspective, it follows that thinking should be conceived as action
rather than as abstract contemplation. To think is not to step out of social activity
and detach oneself from the world, as would the ideal prototypical Cartesian
individual. For theorists of situated cognition, the world cannot be added afterwards to a model of thinking in the form of a set of contextual variables . Rather,
it has to be there from the very beginning . The basic unit of analysis must
connect thinking to action in the world and contribute to clarifying precisely how
cognition enters into and is part of the diverse set of tasks in which people
engage. Furthermore, because virtually all activity is socially distributed, social
units rather than individuals become the appropriate unit of analysis for cognition.
The individual in socially constituted cognition. This primacy of
the social poses a set of theoretical puzzles concerning the role of the individual in
cognitive life . In analyzing these larger units of cognitive activity, how do we
account for what the individual brings to the situation and what he or she takes
away? A grounding assumption for situativity theory is that cognitive activity is
by nature interactive and therefore can only be understood when all partners in the
interaction are accounted for. This might seem a simple enough matter of taking
the perspective of one individual at a time and showing how his or her action is
affected by (and affects) what the others in the situation do. A full account-a
socioindividual account, if you wilt-could then be constructed by compiling

separate social analyses of the situation from the perspective of each individual .
Such an approach, however, would not fully respond to the challenges of
situativity. Situativity embodies a more profound epistemological challenge to
traditional accounts of the nature of knowing: namely, that knowledge is not just
jointly used in the course of cognitive activity but also jointly constituted. In the
strongest form of socioconstruction, knowledge exists only so long as the coconstructing group is present and interacting . This is because the knowledge is
constituted of the interacting contributions of all the members in their particular
situation of co-activity. It cannot, by definition, exist (or at least retain its
original form) once the group or situation is modified. Put another way, if to
know is to act in a particular situation of distributed cognition, then it would
appear that the individual cannot be said to know anything once he or she has left
the situation and the conditions of distribution. At first blush, then, the
situativity point of view seems to make the individual disappear or at least to
exist only when particular others are about. It is as if the individual were recreated
de novo in each new situation .
Although perhaps logical, such a view flies in the face both of
phenomenological experience and of what is known about the way in which the
human brain builds up historical traces of experience. Individuals experience
themselves as continuous and connected across situations. Memory traces
established in the brain constrain and enable the individual's capabilities for action
at any given moment in time . This notion of a history of experience contains the
seed of a resolution of the puzzle posed by radical situati vity . The resolution asks
us to think of individuals as passing through a series of temporally linked
situations. In each situation, the individual brings to a new interaction, with a
particu lar set of other people and artifacts, a brain that has been tuned to respond
easily and automatically to particular affordances and constrai nts. When tunings
and affordances are sufficiently matched, the individual can enter into the particular
interactive situation, both respondi ng to others and shaping their responses so that
a mutually constituted set of cognitive actions is possible (cf. Resnick , 1994).
Ge neral str uctur e of th e vo lu m e. The issues that we have been
discussing-the role of discourse in cognition, physical and mental tools as
constituents of cognitive activity , the social and world-engaged nature of
reasoning, and the ways in which individuals reconstitute knowledge in social
interaction--are addressed from various perspectives throughout the chapters that
follow . The volume is organized in four sections. The first set of chapters
examines the structure of distributed cognition in several complex work settings .
The second group of chapters focuses on the ways in which social identities are
crafted in sociocog nitive communities of discourse. In section three, the chapters
turn explicitly to learning, examining the ways in which people learn to function
in various technologically mediated environments, and the ways in which
technology itself is shaped by users . The final section of the book considers the
ways in which talk, the principal vehicle of shared cognition, is rendered
Lauren B. Resnick, Clotilde Pontecorvo , and Roger Saljo
accountable within social communities and how people acquire the capabilities for
reasoning that their communities recognize .
Part 1. Distributed Cognition: Tools, Discourse,

and Activity in Complex Work Environments
It is common to note that the conditions of the modern workplace require careful
coordination of the actions of many individuals using a complex array of tools and
technology. This observation underlies the nearly worldwide demand for workers
with multiple technical and teamwork skills . But efforts to understand the nature
of the skills required in the technological workplace often founder because they
fail to honor the extent to which people function as part of systems in which
knowledge and competent action are distributed. Consider an airplane cockpit ,
such as the one described by Hutchins and Palen (chapter 1, this volume) . A team
of three individuals work interactively with each other and with a complex array of
information displays to control the performance of the airplane. The situation is
one of high risk and responsibility; an error could mean a disaster. And it is one
in which important decisions must be made under extreme pressure of time. The
airport ground control center, described by Suchman (chapter 2, this volume) is
similar, if slightly less risky and time-stressed. Here, too, responsibility for
successful guidance, in this case of multiple aircraft as they arrive at and leave
airport gates, is distributed across a team of individuals . Information about the
position and status of the aircraft comes to team members not by their direct
observation, but through multiple information displays . The team's ability to
understand situations and act appropriately is, thus, intimately dependent on these
information tools . In both of these settings--the cockpit and ground control--it
makes sense to say that cognition is distributed not only across people but also
between people and their tools.
One could attempt to understand these complex work environments by
outlining the cognitive processes of each individual in them, and could then
attempt to overlay on that a second analysis of how these separate cognitions are
coordinated. That is how a classical task analysis, of the kind often called for in
determining job skills for employee selection or training, would proceed.
Alternatively, one can first attempt to understand the system as a whole, showing
how its various elements--individuals and their mental representations, external
representations in the form of information displays, words, physical actions-function together as an action system . That is the direction chosen by Hutchins
and Palen and by Suchman. Their analyses reveal elements of structure and
process in cognition that are invisible when one tries simply to accumulate the
knowledge and performances of individuals.
The chapters in this section generally replace the concept of coordinated

cognition with one of distributed cognition. They describe cognitive activity units
that involve multiple individuals as well as artifacts functioning as tools . As we
noted earlier, this perspective deliberately blurs the traditional distinction between
intellectual and physical tools. Although technical tools and artifacts (e.g.,
computers, texts, clocks, instruments for navigation and measurement, and other
devices) are located outside the human mind, they are nevertheless integral to
mental work. Just as in managing aircraft in flight or on the ground, when
producing a text (Heath & Nicholls, chapter 3, this volume; see also Grossen &
Pochon, chapter 11, this volume), or checking the color and texture of a chemical
measuring device (Goodwin, chapter 5, this volume), technical tools serve as
prosthetic devices that transform the nature of people's interaction with the
environment and with the other participants. To further our knowledge about
human cognition and learning in technologically complex environments, we need
to understand how people operate with such instruments in practical action.
Rather than considering mind as separate from technical tools, our unit of analysis
must be the human actors operating in practical settings with the technology.
Nothing in the computer, television screen, laboratory, or communication
technologies makes these devices powerful. It is their integration into human
practices that produces new forms of cognitive functioning.
Focusing on the distributed nature of cognitive practices leads naturally to an
interest in how collective attention is managed. Heath and Nicholls (chapter 3,
this volume) show how journalists in a newsroom recruit and monitor each
other's attention as they read aloud information flowing in from news agencies.
Strategies of attention management include voicing and animating texts as they
are read. These readings provide colleagues with cues to the character of the
information provided. Physical body orientation often signals the status of the
enunciation and communicates switches from the role of narrator to that of
commentator.
The socially shared nature of working cognition is further highlighted by
Trognon & Grusenmeyer (chapter 4, this volume). They show how a pair of
workers changing work shifts resolve the problem of a malfunctioning paper
manufacturing machine through what they call an "interlocutory logic." The two
workers start with an initial disagreement, describe experiments they have carried
out individually, and jointly produce a "theory" of the malfunction that allows
work to proceed successfully . In this situation, the need for shared cognition
derives from differential access to activity and information. The same is true in the
newsroom, the airplane cockpit, and the airport ground control center.
Specialization of labor is accompanied by strong interdependence of tasks. This
creates patterns of simultaneous participation by several actors in multiple
overlapping activities. It thus becomes difficult to draw a sharp distinction
between individual and collaborative activities.
The central role of talk in socially distributed cognition is the particular focus
of the chapter by Goodwin (this volume) analyzing activity in a chemical
Lauren B. Resnick, Clotilde Pontecorvo, and Roger Sliljo
laboratory. There an apprentice chemist is learning to prepare a fiber to be used as

a tool for measuring the distribution of water from different sources in the world's
rivers and oceans. The easiest measurement of the right level of the reaction is
provided by the color. The criterion is met when the fibers reach a certain
"blackness" (jet black), as defined in the written scientific instructions. But the
usual lab terminology is that the fiber should look like gorilla fur and not like
orangutan hair. The use of these unofficial terms incorporates many relevant
dimensions : It defines an in-group language that helps to shape the scientist's
identity ; it has an affective flavor; and, at the same time, it enriches the perceptual
distinction, providing an evocation of the distinctive texture of the fiber, which
adds an important tactile informative criterion.
The discourse analysis proposed by Goodwin suggests the need to distinguish
between language (as universal, abstract, existing in the mind, Cartesian, without
human agency) and speech (as an embodied competence, situated in the world).
Language, Goodwin argues, divorces cognition from practice because it focuses on
the abstract and universal bases of naming, whereas "human cognitive activity is
inextricably lodged within the activities and settings of the lived social world." In
the laboratory, multilayered public experiences and representations are at work: for
example, bodily experience of texture, expert gestures, actions that make the
perceptual visible, accountable talk related to the situated activity system,
legitimate judgments (when others recognize the grounds on which they are
made).
Conversation analysts (e.g., Sacks, 1992) have long observed that the way in
which something is categorized (or even named) depends on the attributed
relevance and on the interactional goals one is pursuing . Categorization is a social
endeavor based on the speaker's stance and adapted to specific discursive aims .
Divergence in assigning objects to classes can be the issue at stake between two
speakers: Arguing , reasoning, and thinking often start when assignment is
problematic (Billig, 1987). In Goodwin's chemistry laboratory, what counts as a
valid instance of the color category is established within a public, socially
constituted, world of relevant activity. The activity creates an arena for situated
apprenticeship in which the decision process requires an appropriate use of tools ,
material objects, and mental representations. This results in the creation of a
working language that is very different from what appears in the language of
scientific journals.
The chapters included in this section share several methodological challenges.
One challenge is that the cognitive scholar must have tremendous knowledge of
the technical system and of the functioning of the particular environment under
study in order to acquire an adequate understanding of the cognitive processes at
work. Whether one is studying information processing in the airplane cockpit, the
coordination of airport services, or the functioning of a machine for producing
paper, in each case the cognitive researcher must become something of an expert
in the field. Because readers cannot acquire the same expertise, they must rely
heavily on the researcher's analysis, without a great deal of opportunity to
challenge the basic interpretation . A related problem is one of sampling . In each

chapter, very brief activity episodes are subjected to intensive, detailed analysis.
These detailed analyses of very small bits of activity are meant to represent or
typify the situation as a whole. Much depends on the investigators ' apt choices of
activity sequences, but, because readers do not have direct access to the situations
under study, they must trust the investigators' judgments. This particularity of
research findings is familiar to scholars in anthropology but is less comfortable
for traditional cognitive scientists. Nevertheless, situated research of this kind is
likely to payoff if it is made useful to the people working in the situation (for
example , in better cockpit design or better apprenticeship learning processes) and
also yields some general, theoretical knowledge that extends outside the particular
situation studied.
Part 2. Negotiating Identities: The Construction

of Sociocognitive Communities
A key feature of current thinking about situated cognition is the extent to which
analysts find it impossible or inappropriate to distinguish sharply between acts of
cognition and aspects of the situation that would have, in earlier theories, been
called social, motivational, dispositional, or even personal (cf. Lave & Wenger,
1991). Our authors insist on the mutually constraining functions of thinking and
identity construction , noting that to think or to reason well in a situation is, by
definition, to take on the forms as well as the substance of a community of
reasoners and, thus, to join that community . Much of discourse, and thus of
cognition , serves to situate an individual with respect to others, to establish a
social role or identity .
Social identities are enacted as if people were on a stage co-constructing or coplaying the parts they believe have been assigned to them. According to Goffman
(1974), "Often what talkers undertake to do is not to provide information to a
recipient but to present dramas to an audience. Indeed, it seems that we spend
most of our time not engaged in giving information but in giving shows" (p.
508). The metaphor of stage and drama is not meant to imply that participants are
performing for an external audience. Rather, they are acting with respect to one
another, negotiating their identities within the group and through the group a
relationship to the outside world. Another theater metaphor helps us to understand
how children, or apprentices of any age, learn to enact the right play . As in the
teatro dell'arte, new members of the company learn to play their parts by
participation supported by others who already know the scripts. In Bruner's (1990)
words,
10
When we enter human life, it is as if we walk on stage into a play whose enactment
is already in progress--a play whose somewhat open plot determines what part we
may play and toward what denouement we may be heading. Others on stage already
have a sense of what the play is about, enough of a sense to make negotiation with
a newcomer possible (p. 34) .
A community of practice is not just a collection of individuals who happen to be

interacting. To function as a community, a group needs a common interpretation
of what is happening and what members are trying to do. Narratives of past events
and collective remembering of them (Middleton & Edwards, 1990) are easily
available resources in very close communities (e.g., families, groups of friends,
long-term work mates). Through such narrations, these groups establish
explanations and evaluations that are shared and accepted by others. Co-narration,
including sentence completion and repair for other members of the group, is a
social resource that can also be used by ad hoc groups created on the basis of
common psychological needs, as described by Middleton in this volume (chapter
6, this volume) .
Middleton describes the functioning of a support group for parents of children
with chronic renal failure. Within the group, these parents develop a sense of
being ordinary despite the extraordinary circumstances of their lives. For this
reason, an important part of the discourse is the ways in which members define
themselves as a group distinguished from others . For group members, others
include professionals engaged in their children's care, people without seriously ill
children, and even other family members who do not have primary care
responsibility for the children. Group members communicate about their world
and their psychological states using categories and interpretations that they
themselves use to construct accounts of their lives . In accounting for their
actions, they ratify forms of "ordinariness" in the extraordinary circumstances of
their caregiving obligations, the outcomes of which are uncertain , given the
seriousness of their children 's illness.
The members of the support group are, in a sense, creating a specialized
subculture, one that does not require them to renounce membership in the broader
cultural communities of which they are part, but which creates a specialized
identity within the group. Here, as in other settings, culture is embedded,
explicated, and activated through collective discourse in which assigned and
assumed roles are enacted. Individuals play multiple roles in their lives-for
example, family roles, gender roles, professional roles-with different roles
coming to the fore in different situations. However, communities of practice are
often multirole settings in which role prescriptions are not clear-cut and overtly
referred to, but are "incumbent categories" (Sacks, 1992) that are inferred from the
ways in which participants deliver and accept (or refuse) suggestions, advice,
proposals, and evaluations .
The simultaneous assumption of multiple roles is particularly important in the
capacity of immigrant populations to function successfully in new cultural
11
surroundings. In syncretic literacy, as Duranti and Ochs (chapter 7, this volume)

frame the homework activity carried out in a Samoan American family, the
simultaneous playing of different roles by the same person suggests a complexity
that is not adequately explained by the current views of multiculturalism. For
example, an aunt, asked by a child for help in completing a homework
assignment, assumes both the Western role of a caregiver who accommodates the
child's request and the Samoan role of acting on behalf of the grandmother.
Samoan tradition requires that younger members accommodate to adults' needs,
not adults to those of children. More broadly, although the entire homework
transaction is carried out in English and is done in response to demands from an
American school, several aspects of the transaction--the particular ways in which
the help is negotiated, the location in the house in which homework is done, the
tools that are used-suggest that what is being formed within this family is a nonconflictual syncretic culture that allows members to function simultaneously as
Samoans and Americans.
The assumption of multiple roles seems also to be a factor in the functioning
of therapeutic groups, as studied by Fasulo (chapter 8, this volume) . Here,
members of the group use reported speech to dramatize roles. This is an example
of the heteroglossia described by Bakhtin (1991), which serves to illustrate the
polyphonic nature of self, that is, the social person seen as a crossing of possible
discourses. In the therapy setting, the enacting of multip le roles operates as a
mechanism for acquiring a new identity. The new identity is defined through
fictional reported speech: mainly candidateutterances that one of the participants
could say or had said that were first offered by the therapist. This helps to defme a
possible self as a socially constructed "costume" that is displayed publicly and can
be picked up and "worn" by any group member.
Overt shared discourse is not the only way in which adopted social roles are
displayed. The representation of self in relation to others can also affect individual
performance on a problem-solving task. Even an apparently neutral computerbased task can be affected by the presence or absence of other people, particularly
when they are marked by their gender role (Light and Littleton , chapter 9, this
volume). A well-known body of research shows that peer interaction can enhance
the cognitive level of individual children. This facilitation is usually assumed to
flow from the conflicting points of view that come to light and that require
children to construct a new, higher level of understanding. Light and Littleton
argue that peer facilitation may come not just from conflicting points of view,
but also from the changing meaning of the task, which is affected by social norms
and representations. Although working collaboratively with computers can confer
benefits, facilitation also occurs in the absence of direct verbal interaction. That
is, the mere presence of peers can also increase performance. However, this
facilitation does not occur for girls when the non-interacting partner is a boy .
Without a chance to interact directly with their opposite sex partners, girls appear
to take on traditional gender-based interpretations of the task in which they, as
girls, are not expected to be interested or to do well. Thus, according to Light and
12
Littl eton, "The social context of a cognitive task embraces not only direct
interpersonal interactions but also the social norms, expectations, representations,
and comparisons that condition such interactions."
Part 3. Learning in Practice: How People and Tools

Shape One Another
The theme of learning has only recently become central in research on situated
cognition. At first, just finding the language and analytic frameworks to describe
small slices of activity in complex environments occupied virtually all of the
energy and publication space available to an emerging field. Now, however,
because certain basic ideas can be taken for granted, more attention is being
devoted to questions of how people enter new environments of distributed
cognition, how they learn to act using new tools and in cooperation with new sets
of people. The chapters in this section offer several attempts to formulate theories
of learning that share the developing assumptions of situativity theory: distributed
cognition, tuning to situations, the central role of discourse in learning. The y also
explore the ways in which technologies designed specifically for learning function
to both enhance and constrain learning, as do other cognitive and material tools.
Theories of situated cognition are casting in a new light an old debate among
psychologists: the relationship between learning and development, or, otherwise
stated, the extent to which people grow into competence as opposed to being
taught particular skills or bits of knowledge. The metaphor of growing, which has
dominated developmental psychology, focuses attention on the ways in which
people mature within an environment. Proper development enables one to occupy
a particular environmental niche. At the same time, it systematically leaves one
unprepared for some other niches. The language of learning and teaching , by
contrast, stresses prepar ation for future activity. In most theories of learning , the
most general, abstract, and nonsituated forms of knowledge are privileged . It is
assumed that these will be useful in the widest variety of future situations. The
teaching-learning process, as one of transmission and recepti on, has long been
rejected by constructiv ist theorists, from Piaget forward, who have insisted that to
learn is to appropriate material from the environment and to make it one's own by
active and personal acts of construction and interpretation. In educational theory ,
constructivism has been used as the basis for attacks on traditional forms of
teaching in which a teacher (or text) presents information that students learn to
recite . Constructivism instead calls for various forms of active learning in which
students interact directly with environmental phenomena for which they are
expected to build their own meanings.
Situ ativity theory goes beyond constructivism to insist that the constructive
interactions that constitute learning are, in fact, sociocultural exchanges and that
13
newcomers to a particular environment both adapt to and modify the environment.

Glick (chapter 10, this volume) quarrels with the classical developmentalist focus
on individual progress from younger to older as well as with the concept of
preparation for some general future. He favors instead a view of development as
participation in organizations and their technologies . Through participation, one
is guided toward efficiency of function and displayable (as opposed to testable)
competence. Students of socialization into the workplace (e.g., Scribner, 1984)
have documented how jobs, including the technical language that marks one as an
insider, are learned through these kinds of informal processes (see also Lave,
1991; Greenfield, 1984). This is a view of learning as apprenticeship that
contrasts sharply with established school practices. In apprenticeship learning
environments, the learner is, by definition, a person who has not yet mastered the
culture and is not yet a full member of it, but who nevertheless is viewed as a
legitimate participant (Lave & Wenger, 1991). The novice participant, not yet
fully competent in the new activity system, typically learns by practicing with
people who are already part of the group and functioning competently.
In some respects, the conditions of schooling--in which a whole cohort of
individuals (students) who have not mastered a new activity structure come into
contact with one or two individuals (the teachers) who are experts--are mirrored
when radical new technologies are introduced into a workplace. As Glick points
out, in these cases, the cultural practices and assumptions of an ongoing activity
system are disrupted, and normal forms of apprenticeship by which newcomers
can be absorbed into a working community are not available. At the same time ,
because a new technology often requires a new way of thinking about and talking
about work processes , its introduction calls for the creation of a new culture in
which the people and the technology interact.
Grossen and Pochon (chapter 11, this volume) focus explicitly on this new
cultural construction . They note that, as computers and their software become
partners in activity, it is necessary to understand how they are responded to and
appropriated by users unfamiliar with technology in general or with any particular
program or application. They refer to a field of research that they call
ethnotechnology, the study of the interactions between technologies and societies.
It assumes that the use planned by the designer does not always coincide with the
user 's actual use (see also Perriault, 1989). To understand what constitutes a tool
in use, one must analyze both how various categories of users interact with the
tool as it is first introduced and how some of the interpretations of non-expert
users are incorporated back into the tool design. What eventually emerges has, in
effect, been co-created by designers and users. Designers and users have mutually
appropriated each other's ways of thinking . Perriault (1989) calls the emergent
rationality resulting from the interaction between the user and the machine the
logic of usage. Within the Vygotskian tradition, this would be interpreted as a
social process of creation and the transmission of new semiotic tools.
Grossen & Pochon show how the word processor was initially designed for
technical users and then moved into general use as a modification of the
14
typewriter . But the changes created in the general world of office machines flowed
back into computer design to create what is today a historically unprecedented set
of mental processes for composing and using texts and an accompanying set of
institutional changes . From a sociocognitive point of view, texts are now viewed
as less permanent, and the use of word processing has enlarged the possibilities of
co-authoring. Word processing has also had an impact on the organization of
work, dramatically changing secretaries' jobs and distributing "keyboarding tasks"
throughout the organization, thereby changing their perceived meaning with
respect to status, knowledge, and power.
Word processing is but a single example of the ways in which tools participate
in the definition of culture and competence. Any environment-that is, any
activity system that must be learned by a newcomer-contains a set of tools
(technologies in the broadest sense of the term) and practices for using them that,
in effect, define the culture that the newcomer must enter. Nunes (chapter 12, this
volume ; see also Nunes, Schliemann, & Carraher, 1993) shows that these tools
need not be highly technological. Treating arithmetic procedures as cultural
artifacts, she shows how such procedures guide and constrain reasoning in ways
that are adaptive to the particular situations for which they have been designed.
Different procedures invite different kinds of errors, and Nunes argues that
procedures that have grown up in certain social activity systems protect against
errors that would have important social consequences. In this discussion, although
arithmetic procedures exist only as ideas and verbal rules, they are as constraining-and supporting--as the physical artifacts that are more commonly recognized as
technologies. In another example of the way in which symbols structure thinking
in relation to particular activities, Nunes shows that the use of tiles in learning to
measure area is more effective than the use of rulers, because tiles carry the
meaning of area. This is so even when there are not enough tiles available
physically to cover the space being measured. The tile as symbol for a unit of area
nevertheless continues to function to support reasoning.
Another example of how artifacts and systems of representation affect cognitive
functioning is offered by Schwarz (chapter 13, this volume), who shows how a
particular set of computer-based representations of quantity and functional
relations can affect children's conversation and reasoning about mathematics.
Schwarz's representational software is explicitly designed for pedagogic purposes.
It is an instance of an intermediate abstraction (cf. White, 1993) for learning
concepts, a representation that embodies formal mathematical relations yet is
familiar in form and thus cognitively accessible on the basis of everyday
experience. Such intermediate tools make possible the negotiation of shared
meaning among children working together to solve word problems. By talking
about the objects in the representation system, and by modifying the
representations in order to test hypotheses, children "talk math" even before they
have learned the formal language of the field (in this case, algebra). Children have
to attune to constraints and affordances by negotiating the situated environment
established by the symbolic representation system . In so doing, they develop
15
explanations of why the objects behave the way they do, in conversations that are
similar to the transactive ones described in a later chapter by Teasley (chapter 15,
this volume).
In another use of techn ology to support learning , readers effectively become codevelopers of the texts from which they learn . Hypertexts present the material to
be communicated in modularized sections; learners choose the order in which they
want to read the sections, presumably in accord with their preferred learning styles
and information needs. Schnotz (chapter 14, this volume) examines the ways in
which learners use hypertext resources . He considers the possibility that the
modularization of hypertexts interferes with the flow of thought that ordinary
texts help readers to develop and to which they are accustomed. Hypertexts may ,
like word processing, be a domain in which a new technology can eventually
change the learner' s usual way of functioni ng. At the sam~ time, hypertext
technology, if it is to be widely used, will need to be transformed and adapted to
the user' s logic. This process will be aided by careful study of how people use
hyper texts as a function of both the problem-solving task and the type of guidance
that is given to the subjects (Pontecorvo, Cesareni, & Romanelli, 1995) .
Schnotz's study of the processes students used in learning about time differences
in cities around the world is an example of the kind of research likely to be useful
in this process.
Part 4. Accountable Talk: Learning to Reason

The final sectio n of the volume turns to questions of how people learn to reason
and specifically how reasoning capabilities and habits develop in the course of
social interaction. Although they study very different situations of learning-problem solving at a shared computer screen, high school and adult classrooms,
the family dinner tab le-vall of our authors show how particular forms of reasoning
develop through participation in what we term accountable talk. Accountable talk
occurs in any situation in which individuals are required to defend and justify their
observations and conclusions . Several kinds of accountability may be called for by
the communities of practice in which people learn to reason. At the simplest
level, people in any ordinary conversation are held accountable to the others with
whom they are talking. Thus, participants in accountable conversatio ns must
show, in speech , gesture, or cultura lly appropriate forms of eye contact, that they
are attending to what others have prop osed. In most conversations, people are also
held accountable for facts, evidence, and shared knowledge. For example , one
cannot make factual claims that fly in the face of what is commonly known or
physically evident without being challenged by other participants. Participants in
accountable conversation s are also held to communal standards of reasoning. So ,
premises must follow logically from conclusions, and it is acceptable to challenge
16
a conclusion by challenging the veracity of a premise, even an unstated premise

(cf. Resnick, Salmon , Zeitz, Wathen, & Holowchak, 1993).
Accountability as allocation of responsibility is a central aspect of the social
meaning of actions (Hill & Irvine, 1993) and is accomplished through specific
linguistic forms . However, accountable social participation is not a matter of
language alone. The whole of what is said, how it is said, how it is meant to be
heard, and what is referred to are elements in a process of offering a reasonable
account of one's behavior and words in order to be accepted by the other
participants. Because discourse is so central to situated human action ,
accountability is an implicit theme in many chapters of this volume. In this
section, the ways in which people learn culturally appropriate forms of
accountability take center stage. The chapters each offer an extension of the classic
Vygotskian perspective in which thought is understood as the individual
internalization of initially social activity (Vygotsky, 1934/1990).
The development of language--from the language of and for others to inner
verbal thinking--is the paradigmatic example of this evolution (Wertsch, 1985).
Child language studies have shown the ways in which young children and
caregivers jointly accomplish the production and interpretation of talk (Jacoby &
Ochs, 1995) through repetition, reframing of utterances, contributions of
linguistic elements, and sentential cooperation. In a more general view, children's
language socialization (Ochs & Schieffelin, 1984) can be seen as a socially
distributed practice that is instantiated in culturally diverse ways: in the childcentered manner typical of middle-class Western families (Pontecorvo & Fasulo,
chapter 17, this volume), or in a more situation-centered way, which is well
represented by the Samoan culture (Duranti & Ochs, chapter 7, this volume;
Ochs, 1988).
Teasley 's studies (chapter 15, this volume) of students working together to
solve scientific and logical problems show how joint construction of a problem
solution engages students in a process of explicit analysis and public justification
that leads to higher levels of problem solution than when students work alone.
She shows that explicit talk about theories and evidence is what produces
solutions. Talk about strategies serves to make explicit a cultural tool: testing
hypotheses by experimentally controlling variables. She also shows that the
natural tendency when working with a partner is to engage in these explicit forms
of reasoning, suggesting an explanation for the frequently claimed power of a
collaborative problem as an educative form.
A more complex environment for the development of accountable knowledge
and reasoning, one with practical consequences in the real world, was examined by
Goodwin in an earlier chapter (chapter 5, this volume). Goodwin showed how, in
a chemical laboratory charged with creating a fiber that can accurately measure
water content, a web ofaccountability is created. An apprentice in this laboratory
must learn a complex form of discourse in which her judgments and related
actions come to be recognized as appropriate and reliable by the competent
17
practitioner. Perception itself--the "blackness ofblack"--is shaped by the discourse

that serves both to instruct the novice and to create accountability for the product.
The socially mediated nature of perception is further explored by Saljo and
Berqvist (chapter 16, this volume) in the setting of a school optics laboratory, an
environment whose sole purpose is instruction. The idea of perception as a
strictly physiological process is called into question in their historical and
epistemological analysis. This is reinforced in their empirical study showing that
young physics students cannot rely on what they "naturally" see in order "to see
the light." Not knowing the theory of light that the physics teachers take as
presupposed background, the students literally cannot perceive the behavior of
light that the teachers think is fundamental and obvious. When they are told what
to see, they comply, but it is evident in the examples given that the students are
reporting what they are supposed to see, not what they actually do perceive.
Students do not think that the observation that light does not go through a solid
object, but instead produces a shadow, needs any explanation . Furthermore, the
analogies that are offered to them do not work because the students do not yet
know the concepts that are used to create the analogies. These concepts are
available to teachers because of the discursive community that they are part of,
but not to students who have not yet appropriated this knowledge.
This is a common observation in science teaching laboratories. Students often
see "nothing" under the microscope or in an experiment. Our perceptions are so
conditioned by our theories that we are very nearly incapable of learning from
experience, except when it is mediated by cultural practices of explanation and
initiation into the proper talk of a knowledge community. It is easy, then, to
suggest that these students will come to truly "know" the physicist's explanation
of light only if they remain students of physics long enough to participate in
something like ''physicist' s talk." Whether such authentic discourse can be
organized in classrooms is one of the great challenges that situated cognition
theory brings to education.
There is no problem of authenticity when the site of learning is the family
dinner table. There children are socialized into accepted forms of discourse and
reasoning in a situation that is not explicitly focused on reasoning, and yet it
demands accountability for claims and interpretations. Participating in such
discussions, children acquire both strategies of argumentation and an identity as a
family member, with a particular set of expectations and responsibilities.
Pontecorvo and Fasulo (chapter 17, this volume) examine problem solving in the
context of family talk. The problems to solve in these discussions have to do
with judgments of behavior, attributions of blame and praise, and, above all,
definitions of roles and relationships. As in research on more strictly intellectual
problem solving in ill-structured domains, most of the participants' attention
appears to be devoted to problem definition rather than, strictly speaking, to
problem solving . In problem definition, appropriate resolutions are prefigured by
participants. This phenomenon appears clearly in family talk, where the debates
18
Lauren B. Resnick, Clotilde Pontecorvo , and Roger Saljo
are often about who or what is counted as problematic. Frequently, the issue is
whether there is a problem, exactly what the problem is, or who is the focus of it.
In the family discussions analyzed by Pontecorvo and Fasulo, parents often treat
their children 's behavior and ideas as problematic. They then appeal to categorical
rules and values to educate the children in the categories that function to define
proper behavior. This socialization is managed through specific discursive forms:
reproaches, suggestions, imperatives, and examples . Demands for accountability
are often addressed to children, and children are questioned and challenged more
than parents . By giving explanations and successful arguments, children develop
situation ally appropriate reasoning skills . Particularly striking is the way in
which children use a strategic appeal to a superordinate normative rule (e.g., being
punctual , telling the truth, respecting the rights of children, saving food for
others) to justify an action that has been challenged. Both authorities (e.g., aunt,
grandmother, school teacher, books) and normative values are used, even by very
young children, to challenge others.
The passage from local topics and strategies of discourse to more general and
superordinate ones is shown in another context by Ramirez and Wertsch (chapter
18, this volume), who studied women in an adult literacy program. Analyzing
group discussions on topics relevant to the women (e.g., children's education,
unemployment, drug abuse, and work), the authors show transformations in the
language produced by these adults while acquiring literacy. The issues discussed
by the women shift from reference to their personal daily experiences to more
general social considerations. As they advance in their literacy training, the
women also shift from discussing specific cases to "arguing about categories"
(Billig , 1987). At the same time, they begin to engage complex forms of
expression that combine two discursive genres: the colloquial working class genre
and the rhetorical devices of public debates. This genre combination is aimed at
persuading the audience to identify with their point of view through a kind of
dramaturgical expression . We have here evidence that familiarity with written
texts and literacy practices invokes new forms of oral expression . There is, thus, a
cultural evolution at work, in which mastery of one set of tools (the alphabetic
code and texts embodied in that code) yields a new cultural tool (formal oral
rhetoric). This process of reconstitution of tools in use has been, like discourse
and reasoning, a persistent theme throughout our collection of essays on situated
cognition.
19
References
Bakhtin, M . M. (1991). The dialogic imagination: Four essays . (Ed. by M. Holquist,
Trans . by C. Emerson & M. Holquist) . Austin : University of Texas Press.
Billig, M. (1987) . Arguing and thinking: A rhetorical approach to social psychology .
Cambridge, England: Cambridge University Press .
Bruner, J. S. (1990). Acts of meaning. Cambridge, MA: Harvard University Press.
Cole , M. (1985) The zone of proximal development: Where culture and cognition
create each other . In J. V. Wertsch (Ed.), Culture, communication, and cognition:
Vygotskian perspe ctives (pp. 146-161). Cambridge, England : Cambridge
University Press .
Goffman, E. (1974) . Frame analysis. New York: Harper and Row.
Goodman , N. (1976) . Languages of art. Indianapolis, IN: Hackett.
Greenfield, P. (1984) . A theory of the teacher in the learning activities of everyday
life. In B. Rogoff & J. Lave (Eds.), Everyday cognition (pp. 117-138). Cambridge,
MA: Harvard University Press.
Hill , J., & Irvine , J. (Eds.). (1993). Responsibility and evidence in oral discourse.
Cambridge, England: Cambridge University Press.
Jacoby, S., & Ochs, E. (1995) . Co-construction: An introduction. Research on
language and social interaction, 28(3) , 171-183 .
John-Steiner, V. (1995) . Cognitive pluralism: A sociocultural approach . Mind, Culture
and Activ ity, 2(1), 2-11.
Keller , C., & Keller, J. D. (1993) . Thinking and acting with iron. In S. Chaiklin and J .
Lave (Eds.), Understanding practice: Perspectives on activity and context (pp. 125143). Cambridge, England : Cambridge University Press .
Lave, J. (1991) . Situating learning in communities of practice. In L. B. Resnick, J. M .
Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cogn ition (pp. 6382). Washington DC: American Psychological Association.
Lave, J., & Wenger, E. (1991) . Situated learning: Legitimate peripheral participation.
Mead , G. H. (1934) . Mind, self, and society. Chicago: Universi ty of Chicago Press .
Middleton, D., & Edwards, D. (1990). Collective remembering. London : Sage.
Nunes , T., Schliemann, A. D., & Carraher, D. W. (1993) . Street mathematics and
school mathematics . New York : Cambridge University Press .
Ochs, E. (1988) . Culture and language development : Language acquisition and
language socialization in a Samoan village. Cambridge , England : Cambridge
University Press .
Ochs, E., & Schieffelin, B. B. (1984) . Language acquisition and socialization: Three
developmental stories . In R. A. Shweder & R. A. LeVine (Eds.), Culture theory:
Essays on mind, self, and emotion (pp. 276 -320) . Cambridge, England : Cambridge
University Press .
Perriault, J. (1989) . La logique de l'usage . Essai sur les machines a communiquer [The
logic of use. An essay on communication machinery] . Paris: Flarnmarion.
Pontecorvo, C., Cesareni, D., & Romanelli, P. (1995). Apprendere esplorando
[Learning by exploring] . In M. A. Garito (a cura di), Multimedia and distance
learning for science and technology (pp. 217-237) . Roma: Garamond .
20
Resnick, L. B. (1994) . Situated rationalism: Biological and social preparation for

learning . In L. Hirschfield & S. Gelman (Eds.), Mapping the mind: Domain
specificity in cognition and culture (pp. 474-493). Cambridge, England : Cambridge
University Press .
Resnick, L. B., Salmon, M. H., Zeitz, C. M., Wathen, S. H., & Holowchak, M. (1993) .
Reasoning in conversation. Cognition and Instruction, 11, 347-364.
Sacks , H. (1992) . Lectures on conversation . Vols . 1 & 2. (Edited by G. Jefferson .)
Oxford , England : Basil Blackwell.
Scribner, S. (1984) . Studying working intelligence. In B. Rogoff & J. Lave (Eds .),
Everyday cognition: Its development in social context. Cambridge, MA: Harvard
University Press.
Simon, H. A. (1990) . The sciences of the artificial. Cambridge, MA: MIT Press .
Vygotsky, L. (1978) . Mind in society. Cambridge, MA: Harvard University Press .
(Originally published in 1934.)
Vygotsky , L. (1987). Thinking and speech. (N. Minick, Trans.). New York: Plenum.
(Originally published in 1934)
Vygotsky, L. (1990) . Myslenie I rec '. Psichologiceskie issledovanija. [Pensiero e
linguaggio Thought and language.] Bari, Italy : Laterza. (New Italian critical edition
by L. Mecacci, based on the first Russian edition ; originally published in 1934.)
Wertsch, J. V. (1985) . Vygotsky and the social formation of mind. Cambridge, MA :
Harvard University Press .
Wertsch, J. V. (1991a). A sociocultural approach to socially shared cognition. In L. B .
Resnick, J. M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared
cognition (pp. 85-100). Washington DC: American Psychological Association.
Wertsch, J. V. (1991b). Voices of the mind: A sociocultural approach to mediated
action . Cambridge, MA: Harvard University Press .
White, B. (1993). ThinkerTools: Causal models, conceptual change, and science
education. Cognition and Instruction, 10, 1-100.
Part One
Distributed Cognition: Discourse and Activity in

Complex Work Environments
Chapter 1
Constructing Meaning from Space, Gesture,

and Speech
Edwin Hutchins' and LeysiaPalen'
'Department of Cognitive Science, University of California, San Diego, USA
2Departrnent of Information and Computer Science, University of California, Irvine,
USA
Abstract
Face-to-face communication in the workplace is often conceived of as consisting
mainly of spoken language. Although spoken language is clearly a very important
medium for the creation of representations, in complex work settings, it is one of
several such media. Gestures and the space inhabited by speakers and listeners are
normally thought of as providing context for the interpretation of speech. In this
chapter we show how space , gesture, and speech are all combined in the
construction of complex multilayered representations in which no single layer is
complete or coherent by itself. We examine a brief explanation given by one
worker to two others. We show how the meaning of the explanation is carried in
the coordination among the spatial organization of specilized artifacts, the
positioning of gestures with respect to those artifacts, and the words that are
spoken .
Face-to-face communication in the workplace is often conceived of as consisting

mainly of spokenlanguage. Although spoken language is clearly a very important
mediumfor the creation of representations, in complex work settings, it is one of
several such media. Gestures and the space inhabited by speakers and listeners are
normally thought of as providing context for the interpretation of speech. In this
24
Edwin Hutchins and Leysia Palen
chapter we show how space, gesture, and speech are all combined in the
construction of complex multilayered representations in which no single layer is
complete or coherent by itself. We examine a brief explanation given by one
worker to two others. We show how the meaning of the explanation is carried in
the coordination among the spatial organization of specialized artifacts, the
positioning of gestures with respect to those artifacts, and the words that are
spoken. Our inspiration for this analysis comes from the work of Charles
Goodwin on situated seeing and the cognitive uses of spatial organization
(Goodwin, 19943., 1994b; Goodwin & Goodwin, in press) and of Eleanor Ochs
and her colleagues on the layering of speech and gesture over graphic displays
(Ochs, Gonzales, & Jacoby, in press).
The computational properties of a cognitive system are in part determined by
the patterns of communication within the system. This is true whether the system
is contained in the mind of an individual or distributed across a number of
individuals (Hutchins, 1991, 1995). The representations that are created depend on
the resources available for their creation. What can be represented? How can it be
represented? When a team is engaged in joint reasoning activity, communicative
resources can be seen as media for creating the representations that move
information around inside the system. Communicative behaviors are the
representations by which a socially distributed cognitive system does its work.
Data Collection
The setting for our study is the cockpit of a commercial airliner. This is a
complex high-technology work setting in which the crew engages in event-driven,
high-stakes activities. The quality of the crew's performance depends on their
ability to coordinate their actions with one another and with the dynamic behavior
of the airplane (Hutchins & Klausen, in press; Hutchins, in press). The data were
obtained from a videotape of a simulated flight. The simulation was performed in
a Boeing 727-200 high-fidelity simulator in the Manned Vehicle Simulator
Research Facility (MVSRF) at the NASA-Ames Research Center in Mountain
View, California. Flight in a high-fidelity simulator is very close to the
experience of flying a real airplane. The simulator used in this study provided full
visual displays with dusk lighting conditions and six-degrees-of-freedom hydraulic
motion . A real airline crew composed of pilots employed by a major air carrier
flew a simulated flight approximately one hour in duration from Los Angeles to
Sacramento. The flight was designated NASA 900 in order to hide the identity of
company with which the pilots were employed. As part of the scenario, a
dangerous fuel leak occurred midway in the flight.
There are three cockpit crew members on a 727-200: a captain, a first officer
(FlO), and a second officer (S/O) who is sometimes called a flight engineer. Either
Constructing Meaning from Space, Gesture, and Speech
25
the captain or the FlO actually flies the plane, typically alternating with each
flight leg . This person is designated as the pilot flying (PF) . The pilot not flying
(PNF) handles radio communications. For the flight examined in this chapter, the
captain was the pilot flying , and the FlO was the pilot not flying . The SIO
monitors systems such as the fuel and the hydraulic systems on the SIO panel.
The SIO also troubl e-shoots and refers to the airplane operations manual for
procedures in non-normal situations when necessary.
A low-light infrared camera was positioned behind the crew facing forward. All
three crew members could be seen, as well as most of the main flight instrument
panel and some of the second officer's controls and instruments, including the fuel
panel.
The SIO in the NASA 900 flight discovered the fuel leak by monitoring the
fuel gauges located on his panel. The fuel panel is described at length in the
following section. The interactions that ensue between the crew members upon
the SIO's notification of the problem are the focus of this chapter.
The Arrangement of Pilots in the Cockpit

In a three-person cockpit like the 727-200, the captain 's and the FlO 's seats face
the front windows of the aircraft. The main flight instrument panel is directly in
front of the captain and FlO. Additional controls are located on an overhead panel.
The SIO's seat is mounted on a swivel behind the FlO's seat. The SIO can sit
facing forward or can turn the seat to face the SIO's panel, which is on the righthand side of the cockpit behind the FlO's seat. The main panel is just close
enough to the SIO so that he can physically reach the center portion of it between
the FlO and the captain . The captain cannot reach the SIO's panel , although the
FlO can with difficulty. Both the captain and the FlO can see most parts of the
S/O's panel.
The 727-200 Fuel System

The System
The three engines on the 727-200 are fed by three main fuel tanks and an aft
auxiliary fuel tank. The main fuel tanks are located in the wings and the wing
center section. Tank one is located in the left wing, and tank three is located in the
right. Tank two is in the center of the plane between the wings . The aft (rear)
auxiliary tank is located in the forward section of the aft cargo compartment. Each
engine has a corresponding fuel tank : Tank one has a direct feed line to engine
26
one, tank two to engine two, and tank three to engine three. The aft auxiliary tank
also has a direct line to engine two.
Fuel from each tank can be fed to other engines as well. Cross-feed valves
control the flow of fuel between tanks and engines in the fuel lines. There are
three cross-feed valves, one for each main tank and engine combination. The fuel
from the tank feeds a manifold, and from there the fuel goes to the engine. If the
corresponding cross-feed valve is open, the fuel also feeds into another manifold,
the cross-feed manifold. This manifold can supply all the engines with fuel,
depending on the configuration of the other cross-feed valves (see Figure 1). When
all three cross-feed valves are open, the fuel is free to flow from all tanks to all of
the engines. Direction of flow is determined by the pressure in the fuel lines and
by check valves that permit fuel flow only in one direction. When the cross-feed
valves are closed (the default setting), the configuration is called tank to engine,
because each engine is fed by only its own tank. The crew can control how much
fuel is burned from which tanks by using the cross-feed valves to direct the fuel
flow.
= '.... .
.-
ax=-..
Fig. 1.1. A model of the fuel system as it appears in training and operations
manuals.
Boost pumps also control fuel flow by supplying the pressure necessary to move
fuel to the engines. Tanks one and three and the aft tank each have a pair of boost
pumps located in the tanks. Tank two has two pairs of boost pumps because the
27
tank is divided into sections, two of which are located over the root of each wing .
Boost pumps can be turned on or off. When the boost pumps are on, they extract
fuel from the tank and feed it into the fuel manifold and to the engine. When the
pumps are off, the fuel remains isolated in the tank.
Fuel Instrumentation
The fuel system gauges and switches are located on the lower left section of the
S/O panel on the 727-200 . The fuel panel in the simulated airplane displays four
fuel quantity gauges: one for each of the main tanks and one for the aft auxiliary
fuel tank. The corresponding boost pump toggle switches, the low pressure
indicator lights , the cross-feed valves, and the engine shutoff valves are also
displayed on the panel (see Figure 2).
Fig. 1.2. The fuel panel. The second officer constructed his gestures in coordination
with the spatial organization of this panel. The number three tank gauge and boost
pump control switches are on the right side of the panel. The fuel quantity test button
is at the far left.
28
Signs of a Fuel Leak
A crew uses the cockpit instrumentation and cues from flight controls and other
aspects of the environment to monitor the status of the aircraft. The following
are some of the signs that may be available to a crew when there is a fuel leak in
tank three.
Flight Controls. A leak in tank three, which is located in the right wing,
should cause the plane to roll to the left as the right wing becomes lighter than
the left wing. This condition is called a lateral weight imbalance. In order to
maintain a wings-level attitude with such a lateral weight imbalance, the control
yoke would have to be tipped to the right. The need to adjust the control yoke is a
cue to the pilots that something may be amiss. If the aircraft is on autopilot,
however, the autoflight system would make the required control correction
without notifying the crew. Although the autopilot system still physically tips
the yoke, this visual cue is subtle, because the amount of displacement of the
yoke may be small.
Instruments. Another sign of a fuel problem would be a greater decrease in one
of the fuel tank quantity gauges relative to the other gauges as indicated by the
gauge needle levels. On the 727-200, this information is available most readily to
the S/O because the instrumentation is located on his panel.
What to Do in Case of a Fuel Leak
When a fuel leak is suspected, a typical response for the S/O is to press the fuel
quantity test switch to confirm that the fuel quantity gauges themselves
are operating properly. This test confirms that an irregular gauge indication is a
result of the physical state of the fuel system and not a result of a malfunctioning
gauge. Pressing the fuel quantity test button moves each of the fuel tank gauge
needles simultaneously to different positions to test for responsiveness. When the
fuel quantity test button is released , the needles return to their original positions.
Once the S/O confirms that the gauges are working properly, the next step is to
locate the fuel leak. A leak could be in one of two places: in the tank itself or
somewhere in the fuel line.
To determine if the leak is in the tank, the fuel must be isolated in the tank by
turning off the boost pumps in that tank. If the gauge still indicates a decline, the
leak is in that tank. If there is no decline in fuel quantity when the boost pumps
are turned off, the leak may be somewhere in the fuel line. This is an even more
dangerous situation than a leak in the tank, because the fuel may be escaping into
the fuselage where it could ignite and destroy the airplane. An additional action
that can confirm a fuel leak is to perform a visual check outside a cabin window
29
to look for fuel escaping from the wing. In a simulator, a visual check is
simulated by asking an experimenter if fuel can be seen from the wing.
Another action must be taken before the diagnostic of turning off the tank three
boost pumps is performed. An alternate fuel supply to engine three must first be
established so that the engine will not stop working when the fuel it normally
receives from tank three is no longer available. To do this, the cross-feed valves in
the fuel lines between the new fuel source and engine three must be opened. Once
this step is taken, the tank three boost pumps can be safely turned off.
The Second Officer Explains His Diagnosis

From the detection of the problem with the fuel system to the safe landing of the
aircraft, the crew engaged in many kinds of activities. We examine the 24 seconds
during which the SIO notified the captain and the FlO of the problem
and explained how he had diagnosed the problem . The following is a transcription
of the verbal behavior of the crew during this brief episode.
Transcription symbols:
\2\ Indicates a pause (here, a 2 second pause)
xxx Indicates an uninterpretable utterance

12 .00 .43
8 /0
well it l ooks huh like a funny situation . we

have a fuel leak or something \2 \ in number
three tank
12 .00 .50 Capt hnmn

12.00.51
FlO
FlO
12 .00 .56 8 /0
ohhhh
xxx
1 don't know we must be losing it very quickly

you see right now 1-\2\1 turned the pumps off ok
1 tried to feed from number one to both engine
one and three but we're still losing in number
three quite a bit
In a previous analysis of the SIO's announcement of the problem and explanation

of his actions regarding it, we began with the transcript of the verbal behavior and
tried to show how the gestures supported the speech. It became clear to us,
however, that this separation of speech from gesture and the removal of the
gestures from the space in which they were performed distorted the phenomena. In
the following analysis, we therefore try to show how space, gesture, and speech
interact with each other, giving none of them precedence over the others.
30
The S/O turned in his seat to face the front of the airplane while addressing the
captain and FlO . No gestures other than body orientati on accompanied this
announcement.
12 .00 .4 3 S/ O
well it l ook s huh like a funny situation . we

have a fuel leak or s omething \ 2 \ in number
three tank
The S/O's opening announcement was a call for the attention of the other crew
members . The language the S/O used was explicit but indicated some uncertainty.
A fuel leak is a potentially flight-threatening situation and requires the immediate
and coordinated attention of all the crew members. After the S/O's announcement ,
the crew members collectively knew what the S/O suspected (a fuel leak) and
where he thought the problem was located (in fuel tank three). With that
information, the crew members prepared to attend to the problem.
Fig. 1.3. A frame from the video tape showing the 727 cockpit. The captain,
first officer, and second officer are all attending to the fuel panel that is on the
instrument panel at the bottom right.
By focusing their attention collectively, they created an environment that enabled

them to collaborate and develop a shared understanding of the fuel problem. We
31
assume that the crew members' mental models of the fuel system were similar
because they all received similar company training for the 727-200. Additionally,
they all have access to the same manuals that describe the operation of the fuel
system and fuel system procedures. Still further, typical airline career trajectories
start in the SIO position and move to the FlO position and then finally to
captain . In most cases all the crew members have had SIO experience.
A salient part of a pilot's understanding of a fuel leak is that it is a situation
that must be dealt with quickly. In response to the SIO's announcement, the
captain and the FlO turned in their seats to face the SIO and the SIO's panel
(Figure 3). Each of them also produced a contentless verbalization with a rising
intonation.
12.00.50
Capt
hrnmn
12.00.51
FlO
FlO
12 .00 .56
S IO
ohhhh
I don't know we must be losing it very quickly you see
The very act of the captain and the FlO turning around to face the SIO and the fuel
panel indicates that they heard the SIO's announcement and realized that their
attention was needed. Once the captain and the FlO were situated, the SIO began
his explanation of the problem without further prompting. As the SIO spoke , he
turned in his chair to face the fuel control panel.
SPEECH
GES'IURE IN SPACE
right now
placed index finger on, but did

not depress the fuel quantity test
switch
With the fuel system, there is always a question of whether what is observed is
really the behavior of the fuel system or if it is simply a gauge malfunction. The
SIO began by gesturing to (placing his finger on, but not depressing) the fuel
quantity test switch while saying "right now."
There was nothing in the S/O's words about the fuel quantity test button .
Pressing it in the context of a suspected fuel leak would have been a meaningful
action. But the S/O did not press it. He only touched it. We believe that the other
crew members interpreted this as an indication that the S/O had already tested the
gauges (in fact, he had) .
The words "right now" gave a sense of immediacy to the situation. They place
something in the present time, but what it is not yet clear. The speech and the
32
gesture seem to be working independently of each other here, each conveying a

different sort of information about different topics.
SPEECH
GES'IURE IN SPACE
1-\2\1 turned the

pumps off
ok
brought index and middle fingers down

onto the tank three boost pump
switches Which were in the off
position
The S/O next made a motion over the number three tank boost pump switches
that mimicked the motion used to turn the pumps off. The switches were already
in the off position . The combination of the gesture and the state of the panel and
the knowledge that boost pumps are normally on in flight made this action
unambiguous. The words redundantly expressed that which the gestures had
described, but the words also included information about temporal relationships
that cannot easily be represented in gesture. Speech marked the gestures as a
reenactment of what the S/O had already done. The verbal statement did not
indicate which pumps had been turned off, but the fingers did. The location of the
gesture in the space of the fuel panel resolved an ambiguous reference in the
verbal stream . The verbal component provided temporal markings that were
lacking from the gesture, and the gesture provided aspects of indexical reference
that were ambiguous in the S/O's words.
If the pumps were off, one may wonder where the fuel for engine three was
corning from. The topology of the panel facilitates certain inferences about the
functional behavior of the fuel system, and the S/O next moved to demonstrate
these inferences to the other crew members.
SPEECH
GES'IURE IN SPACE
I tried
moved hand from the tank

three boost pump switches to the area
of the tank one boost pump switches
The S/O changed topics at this point and his gesture directed attention to the other
side of the fuel panel where subsequent events would be described. He was now
beginning to explain how he established an alternative fuel source for engine
number three. The use of the past tense placed the action referred to in the past
with respect to the present course of action.
Constructing Mea ning from Space, Gesture, and Speech

SPEECH
GESWRE IN SPACE
to feed from
number one
moved index and middle fingers up and

down between the tank one quantity
gauge and the boost punp low pressure
indicator l i ght s
33
Here the gesture and the speech were almost completely redundant. The gestures
indicated the states of the controls that feed fuel from tank number one as the
fingers moved along the lines painted on the panel that depict the pipes in the
system that move fuel from the number one tank, through the boost pu mps ,
and to the engine one fuel feed valve.
SPEECH
GESWRE I N SPACE
to both
hand raised away f r om the sufrace

of the fuel panel
The S/O's hand lingered a moment near the controls for tank number one.
SPEECH
GESWRE IN SPACE
engine one
moved hand across the fuel panel

t o the area of the controls for
tank and engine number three
The S/O pointed to the area of the engine number three cross-feed valve and main
fuel supply valve while saying "engine one."
SPEECH
GESWRE IN SPACE
and three
moved hand back across t he f uel

panel to the area of t he controls
for t ank and engine number one
In the brief statement, "I tried to feed from number one to both engine one and
three," the S/O explained that he had remembered to feed fuel to engine three
before he turned the tank three boost pumps off. The gesture accompanying this
section was complex and quickly executed. The S/O pointed to the tank one
gauge, to the tank one pumps (which were on), then to the engine three cross-feed
valve controller, and to the engine one cross-feed controller. These gestures drew
attention to the controllers that indicate that the valves were open and supplying
34
fuel to engines one and three from tank one. Some of the motions of the hand
also followed the flow of fuel through the system .
SPEECH
GES'IURE IN SPACE
but
pointed with index finger to the

engine three fuel gauge
Having established the alternate source of fuel for engine three, the S/O pointed to
the engine three fuel gauge. This was the locus of the problem. The S/O marked
with gesture a return to the topic of the fuel level in tank three and, with speech, a
return to the present tense. But signals a logical disjunction . The elements that
stand in disjunction are not yet clear but will be made clear by what follows.
SPEECH
GESTURE IN SPACE
we're
flicked the face of the engine

three fuel gauge with middle finger
The S/O flicked the gauge with his finger. This is a common technique among
pilots to free a gauge needle that is believed to be stuck. From a strictly
functional point of view, this is a useless action. The S/O detected the fuel leak
by observing the rapid movement of the fuel gauge needle. The fact that it was
possible to detect the fuel leak is evidence that the needle is not stuck.
This flick was not performed in the S/O's original diagnosis and was not a
report of a previous action. Rather, it was a new action performed while the other
crew members looked on. Because this action was not functional, we might ask
what other kind of role it might be playing here . For one thing, it returned the
narrative to the temporal present. It was a way of emphasizing that the fuel level
shown by the number three tank gauge is the salient problem. At a more abstract
level of description, flicking a gauge is a way to produce an expected reading when
an unexpected reading has been encountered. In that sense, this action could also
be read as a assertion by the S/O that he would have liked the behavior of the
gauge to be other than it was.
SPEECH
GES'IURE IN SPACE
still losing
repeated jabbing motions at the face

of the tank three gauge with the
index finger
35
The S/O then emphatically gestured to the tank three gauge, while he said, "still
losing ." This last gesture drew attention away from the function of the needle
(that which the prior gesture, the flick, demanded) to the actual fuel quantity level
that the needle was indicating.
SPEECH
GESTURE IN SPACE
in number three quite

a bit
moved hand away from panel and

into lap
Finally, the S/O returned his hands to his lap indicating that his turn was
completed.
The Multilayered Representation

The actions of the S/O produced a multilayered representation. Gesture was
superimposed on the physical structure of the fuel panel itself, and the S/O 's
verbal account was superimposed on the gesture . If we want to understand what
the crew members do, we must take into account the production and use of such
complex structures. We will try to show what each layer contributed and why we
cannot entirely separate the layers from one another.
None of the layers was completely coherent by itself. The panel provided a
coherent depiction of the fuel system, but it was neither a representation of what
the S/O had done nor even a representation of the state of the airplane . The fact
that fuel was leaking from the wing tank was not represented in the instantaneous
state of the fuel panel. It could only be inferred by comparing the rates of change
of the tank indications over time.
The gestures performed on the panel nearly provided a complete account by
themselves . They certainly formed a more complete and meaningful description of
what was done than the S/O's words did. How can this be?
The Panel
First, the spatial organization of the panel is a central element of the usefulness of
the panel as a communicative resource. The spatial layout of the panel is
topologically (but not metrically) identical to the spatial layout of the fuel system
that it depicts. Table 1 shows correspondences between components on the fuel
panel and components in the fuel system.
36
IN THE SYSTEM
ON THE PANEL
Fuel tank
Fuel line
Quantity gauge
Painted line
Purrp control switch
Pressure indicator
Valve control switch
Position of valve control with
respect to painted lines
Punp
Pressure sensor
Valve
Fuel flow established
by valve position
The topological relations among panel components (e.g., the quantity gauges,
painted lines, and pump control switches) are the same as the topological relations
among the system components (e.g., fuel tanks, fuel lines, and pumps). The
actual mapping of the space of the fuel system onto the space of the panel
is complex. Components that are higher on the panel generally correspond to fuel
system components that are forward in the airplane. Components that are to the
right on the panel generally correspond to fuel system components that are on the
right of the airplane. But there are exceptions. The gauge and pump switches for
the rarely used aft auxiliary fuel tank have been placed out of the way so that they
do not interfere with the depicted relations among the main tanks and the engines.
The panel is further simplified by omitting depictions of check valves that cannot
be controlled from the panel.
The topology of the painted lines and switch positions creates a representation
that permits the crew to do conceptual inferences with simple and robust
perceptual skills. For example, figuring out where fuel will flow can be
accomplished by visually following lines on the panel. The valve controller has a
line painted on its top surface. When the controller is in the cross-feed position,
this line appears to connect the painted lines that depict the fuel line arriving at
and departing from the valve. The rotational action of the cross-feed valve
controllers, combined with the shape of the controller knob, makes the open and
closed states of the valve "look like" flow through or blocked flow. These may
seen to be trivial design features, but they have important cognitive consequences.
Imagine valve switches of a different kind (e.g., toggle switches) and a readout
that lists the name of the valve and its state in text format. With such a
representation, it would be impossible to use simple perceptual skills to reason
about the behavior of the system.
The simplified topology of the panel as a representation of the fuel system
itself permits the pilots to reason about the state and behavior of the fuel system
by "seeing" the panel in a particular way (C. Goodwin, this volume; Goodwin &
Goodwin, in press). The fuel system itself as a collection of physical components
cannot actually be seen from any real vantage point, but the pilots can "see" the
fuel system by seeing through the fuel panel. In fact, only through seeing fuel
panels and diagrams such as Figure 1 do pilots have any experience of the
topology of the fuel system. As with any materially instantiated symbolic
37
representation , it is possible to see either the representation itself or to see the

thing that is represented. Sometimes it is possible to see both at once.
Understanding the SIO's performance requires several shifts in seeing. How do
gesture and speech guide these shifts between the perceptual stance in which the
panel is seen as a thing in itself and the perceptual stance in which the panel is
seen as a representation of the fuel system?
Meaningful Gestures
The gestures superimposed on the space of the panel can be read as meaningful
actions and courses of action on the fuel panel itself, or they can be seen as events
in the fuel system . Seeing each of these things requires a different stance with
respect to the panel. To see the gestures as actions on the panel, one must see the
panel as a panel. To see the gestures as representations of events in the fuel
system, the panel must be seen as the system that it represents .
The first meaningful gesture in this sequence is the SIO placing his finger on
the fuel quantity test switch. The fuel quantity test switch differs from all other
elements of the panel. All the other elements are in some sense "about" the fuel
system, but the fuel quantity test switch is "about" a set of components , the
quantity gauges, on the panel. This gesture must be read as being about the panel
rather than about the fuel system. In order for the captain and FlO to interpret the
SIO's gesture to the fuel quantity test switch, more than a shared understanding of
its function was necessary. It was not enough that they all have a similar model
of the switch's function. They needed to know that the others had a similar model
of the function as well. This kind of intersubjectivity underlies all of the
meaningful actions on the panel.
The procedure for diagnosing the fuel leak involves two distinct courses of
action. The first course of action establishes an alternate fuel supply for engine
three. The second course of action is to turn off the pumps in tank three and to
monitor the gauge for continued fuel loss. These courses of action were executed
in this order by the SIO before he notified the crew of the potential problem. The
explanation he gave of his action, however, interwove the two courses of action,
placing all of the second course of action in between the elements of the first.
We find it interesting that, although the order in which the actions are reported
is not the same as the order in which they were executed, it is the same order that
would be encountered in a traditional problem-solving account. The goal of
turning off the boost pumps for tank three cannot be accomplished directly
because it will cause engine three to flame-out (quit running) . This leads to the
creation of the subgoal of establishing an alternate fuel supply for engine three.
Once this has been accomplished, the pumps can safely be turned off and the
gauge monitored for further fuel loss.
The gestures acquire their meaning by virtue of being superimposed on the
meaningful spatial layout of the fuel control panel. The same gestures produced in
38
the absence of the panel would, of course, be quite meaningless. Enacted over the
panel, though, these gestures take on meanings such as "turning off the pumps"
and "the newly established path along which fuel is flowing ." The functional
consequences of the actions re-enacted in gesture by the S/O are easily seen by the
other members of the crew.
Speech
The verbal layer of the representation does things that cannot be done in the other
layers . For example, it uses tense markers and other linguistic devices to indicate
temporal relationships among actions . Gesture by itself is always action in the
present. The verbalizations place the actions in a temporal framework. This is
what makes it possible for the S/O's actions to be seen as a re-enactment of
action already taken rather than as a proposal for action to be taken.
Speech is also used to indicate the S/O's relationship to the actions and to
belief states derived from the actions. The use of personal pronouns is interesting
here. The S/O speaks of (1) his own state of knowledge, "I don't know," (2) a
condition shared by them all, "we must be losing it very quickly," and (3) a
relation between the captain and F/O to the shared condition, "you see, right
now ." Responsibility for actions and even for the flight in general are often
implicitly expressed in the use of pronouns in such settings.
In this excerpt, speech is used to control conceptual and temporal relationships.
Consider the words, "but we're still." There is a conceptual disjunction in the
"but" and a temporal disjunction in the "still." The conceptual disjunction marked
by the use of "but" is between the expectation that the level in tank three will not
decrease if the boost pumps are turned off (in normal operation) and the fact that
the level is decreasing. The temporal disjunction is between the past action that
should have put an end to the decrease and the present fact of continued decrease.
These disjunctions, together with the S/O 's action of flicking the gauge with his
finger , move the discourse back into action in the present time. In addition to
expressing the S/O 's relation to the gauge reading and shifting attention
momentarily from the fuel system to the panel , the flick gesture brings the
account back into the present tense. It is the only action taken on the panel during
the explanation. All of the other gestures depict an idealized set of movements
that the S/O 's hands might have made in doing the diagnosis.
Discussion
Does gesture support speech? Clearly it does, but no more so than speech
supports gesture . This example demonstrates the creation of a complex
39
representational object that is composed through the superimposition of several

kinds of structure in the visual and auditory sense modalities . Granting primacy to
anyone of the layers of the object destroys the whole.
The physical layout of the fuel panel and its relations to previously encountered
representations of the fuel system permit the crew to see the panel as an object in
itself and as the fuel system it represents. This allows the gestures performed over
the panel to be interpreted as actions taken on the panel, or as events in the fuel
system , or both . The speech is used in part to manage relations that are not easily
expressed in gesture and also to move from one interpretive mode to another. In
this way, the whole is a complex interwoven performance. The properties of the
crew and cockpit as a cognitive system are in part determined by the patterns and
richness of communication among them. The space of the panel, the placement of
the crew with respect to the panel, and the availability of hands for gesture all
have consequences for the communicative possibilities in the cockpit. In order to
understand the operation of such systems, it is not enough to understand the
properties of individual cognition or even of individual decision making with
decision aids (although such knowledge may certainly be helpful) . Real world
decision making often involves the creation and use of the sort of complex
multilayered public representation described here.
Acknowledgments
Funding for the research described here was provided by grant NCC-2-591 to
Edwin Hutchins from the Ames Research Center of the National Aeronautics and
Space Administration in the Aviation Safety/Automation Program. Everett
Palmer served as technical monitor.
40
References
Goodwin, C. (1994a) . Professional Vision. American Anthropologist 96(3) , 606-633 .
Goodwin, C. (1994b). Seeing in depth: Space. technology and interaction on a
scientific research vessel. Unpublished manuscript. Univers ity of South Carolina,
Columbia, South Carolina.
Goodwin , C., & Goodwin, M. (in press) . Formulating Planes : Seeing as a situated
activity. In Y. Engestrom & D. Middleton (Eds.), Cognition and communication at
work. New York : Cambridge University Press .
Hutchins, E. (1991) . The social organization of distributed cognition. In L. B.
Resnick, J. M. Levine, & S. D. Teasley (Eds.), Perspectives on Socially Shared
Cognition (pp . 283-307) . Washington, DC: American Psychological Association
Press .
Hutchins, E. (1995). Cognition in the Wild. Cambridge, MA: MIT Press .
Hutchins, E. (in press). How a cockpit remembers its speeds . Cognitive Science.
Hutchins, E., & Klausen, T. (in press) . Distributed cognition in an airline cockp it. In
Y. Engestrom & D. Middleton (Eds.), Cognition and communication at work . New
York: Cambridge University Press .
Ochs, E., Gonzales, P., & Jacoby, S. (in press) . When I come down, I'm in the domain
state . In D. Slobin, J. Gherhardt, A. Kyratzis , & J. Guao (Eds.), Social interaction ,
social context, and language: Essays in honor of Susan Ervin-Tripp.
Chapter 2
Centers of Coordination:
A Case and Some Themes
Lucy Suchman
Xerox PARC, Palo Alto, California, USA
Abstract
This chapter identifies a class of worksites characterizable in terms of
participants' ongoing orientation to problems of space and time, involving the
deployment of people and equipment across distances according either to a
timetable or to the emergent requirements of a time-critical situation. To meet
simultaneous requirements of mobility and control, centers of coordination must
function as centers to which participants distributed in space can orient, and which
at any given moment they know how to find. At the same time, to coordinate
activities distributed in space and time, personnel within the site must somehow
have access to the situation of co-workers in other locations. One job of
technologies in such settings is to meet these requirements through the
reconfiguration of relevant spatial and temporal relations. This general
characterization is explored through ethnographic materials from an investigation
of the work of airline ground operations at a metropolitan airport on the west
coast of the United States.
Introduction
Recent work within anthropology and sociology includes a growing body of
ethnographically based analyses of the place of technologies in human activity and
42
Lucy Suchman
of human activity as grounds for the significance and functionality of

technologies . These analyses differ from traditional treatments of "humanmachine interaction" within cognitive psychology and computer science in their
interest in locating technologies within the socially organized activities and
settings of their production and use. Research has proceeded through the study of
technology intensive forms of practice that include medicine, science, navigation,
office work, machine design, maintenance and repair, specialized worksites such as
a city dealing room, and a class of worksites that might be called centers of
coordination.t Studies of the latter include the research of Whalen (1992, 1993)
and Whalen and Zimmerman (1990) on call taking and dispatch in 9-1-1 public
safety centers, of Harper and Hughes (1993) and Harper, Hughes, Randall,
Shapiro, and Sharrock (in press) on air traffic control centers, of Heath and Luff
(1992) on line control centers in the London Underground, of Filippi and
Theureau (1993) on control rooms in the Paris Metro, and of our own research on
airport ground Operations rooms.l
Centers of coordination are characterizable in terms of participants' ongoing
orientation to problems of space and time, involving the deployment of people
and equipment across distances, according to a canonical timetable or the emergent
requirements of rapid response to a time-critical situation . In this chapter, I
outline a set of themes that organized and were developed by the analysis of one
particular center of coordination. By examining ethnographic materials from this
worksite, we begin to get detailed access to the phenomenon of coordinated work
activity as the practical reasoning and action involved in maintaining an
institutionally accountable spatio/temporal order. My goal is both to provide a
guide to the analyses produced within that particular project and to contribute to a
collective framework within which a larger body of studies might be related .
1Anyone familiar with this growing literature will recognize the difficulty of an
adequate listing . On medicine , see for example Barley, 1986; Dugdale & Fujimura, in
prep ; Hartland, 1993; Jordan, 1992a. On science , see Clark & Fujimura, 1992 ;
Goodwin , this volume, 1994; Knorr-Cetina & Mulkay, 1983; Latour & Woolgar,
1979; Lynch , Livingston, & Garfinkel, 1983; Lynch & Woolg ar, 1990; Pickering,
1992; Star, 1989; Traweek, 1988. On navigation, see Hutchins, 1990, 1991; on office
work, Blomberg, 1987; Suchman, 1983; Heath & Luff, 1993; Luff & Heath , 1993; on
machine design, maintenance and repair , Bowers, 1994; Forsythe, 1993; Orr, 1990 ;
Sharrock & Anderson, 1993; Star & Ruhleder, 1994; Suchman, 1987; Suchman &
Trigg, 1993; on the city dealing room, Heath, Jirotka, Luff, & Hindmar sh, 1994 . For
more on the relation of the phrase "centers of coordination" to Latour's "centres of
calculation" (1987), see Suchman & Whalen, 1994.
2Participants in this research included Francoise Brun-Cottan, Kathryn Forbes, Charle s
Goodwin, Marjorie Goodwin, Brigitte Jordan, and Randy Trigg . It should be obvious
that this chapter is not only indebted to them but is also meant largely as a reference to
their work. Paul Drew, Christian Heath, and Emanuel Schegloff all served as invaluable
consultants. (See Brun-Cottan, 1991; Brun-Cottan et al., 1991; Goodwin , C., 1991;
Goodwin, M., 1995, in press ; Goodwin & Goodwin , in press; Jordan , 1990, 1992b ,
1992c ; Suchman , 1993, in press ; Suchman & Trigg , 1991.
Centers of Coordination
43
The Site
From 1989 through 1991, we carried out an extended study of the work of ground
operations at a metropolitan airport on the west coast of the United States.' We
began the project with a general interest in contributing to analyses of the
dynamic structuring of people's interactions with each other and with their
material environments, and in exploring the relevance of such analyses to
problems in design.
We were initially drawn to the airport as a study site by two considerations.
First, as a workplace, an airport includes a diverse collection of professions and
activities, which in some respects occupy clearly delineated, largely separate
territories . At the same time, the coordination of the work requires interaction
across territorial boundaries at certain critical junctures. Second, during the course
of our study period, a new terminal building was completed and opened. This
made it possible for us to learn from the changes involved in moving operations
from one work setting to another. We were interested to see how the relations of
work and technology established in one facility would be reproduced, or
transformed, as they were re-established in a new setting.
Our study focused on the two airlines scheduled to move into the new terminal
and, within those airlines, on the work of ground operations: that is, all of the
work involved in servicing arriving and departing airplanes." Within ground
operations, we took as a further focus a particular "backstage" area, called the
Operations room, charged with coordinating the work of the gates and the ramp .
The Operations room proved to be ideal as a setting for our interests. Within the
room, we were able to see the working practices of a small group of people copresent to each other and closely attuned to each other's actions. At the same
time, the division of labor was such that each person within the room was
assigned the task of maintaining communications with some other relevant
location outside the room, via various technologies. We therefore were able to see
the structuring of these distributed, technology-mediated interactions as well.
Finally, we were able to see how these two working orientations - to co-workers
within the room and to those outside - were organized in relation to each other.
3The research, under the heading The Workplace Project, was funded by Xerox and
Steelcase Corporations. Brun-Cottan et al. (1991) presents a video final report on the
study.
4The work of ground operations does not include air traffic control, but rather all of the
activities involved in servicing an airplane while it is at the gate, for example, loading
and unloading passengers and baggage, ensuring that connecting baggage from an
incoming plane is moved to the appropriate outgoing plane , replenishing the plane 's
store of food and fuel. For analyses of air traffic control as a center of coordination, see
Harper and Hughes (1993) and Harper et al. (in press) .
44
Lucy Suchman
Themes
Our project, as analysts of the work of Operations, was to see just what that work
comprises within this particular site. At the same time, the materials from the
site contributed to a developing understanding of the social and material
organization of skilled practice within complex, technology-intensive worksites
more generally. The analyses that resulted from the project explore a set of themes
that both organized our looking at the site and were developed and elaborated by it:
Technologies as material practice: The inseparability of technologies and
the activities of their use. This includes locating the functionality of technological
artifacts not in particular devices, but in densely structured courses of action
involving the assembly of heterogeneous devices into a working information
system.
Reading a scene: How competency in these settings involves learning how to
read a scene, through the juxtaposition and interpretation of verbal reports, visual
images, and various forms of text, in real time, into provisional assessments of
an emerging situation .
(Re)producing a normal order: How, through their management of everyday
contingencies , workers are able to maintain an accountable spatio/temporal order.
Structures of participation: How participants in a multi-activity setting
structure their focus of attention and engagement from moment-to-moment.
Constituting workspaces: How workspaces are dynamically configured
through interactions across visible and invisible boundaries .
Acquiring competency: How the identity of competent practitioner is
acquired through progressive rounds of increasingly demanding work, supported in
situ by experienced co-workers.
Authoritative knowledge: Relations between participants' access to
technologies and the distribution of knowledge taken to be consequential for the
work at hand.
Designing for change: Implications of the analysis for relations between
professional design and design-in-use.
45
Techn ologi es as Ma ter ial Practice

A central theme of our studies is the intimate relationship between work
environments and the structuring of work activities . We take the work
environment to include architectural features and furnishings, telephone lines,
radio frequencies, computer screens, video monitors, documents, and the like .
These objects, moreover, assume multiple identities according to their relevance
for practice; for example, an airplane may be for one person at one moment a
specific aircraft, whereas for another, it is an instance of a flight, a container to be
loaded, a machine to be repaired, and so forth (Brun-Cottan, 1991; Goodwin &
Goodwin, in press; Suchman, 1993). There are no uni-directional effects between
these elements . Rather people are engaged in a continuous process of making the
environment work for the activities at hand. In doing so, they leave the mark of
their activities on the environment in ways that set up the conditions for
subsequent actions . Along the way, the workspaces, furnishings, technologies,
and artifacts are experienced as more and less focal or contextual, negotiable or
resistant, enabling or constraining of the work that needs to be done.
Their function as centers of coordination requires that Operations rooms
comprise a stable site to which participants distributed in space can orient and
which at any given moment they know how to find. At the same time, to
coordinate activities distributed in space and time, personnel within the site must
somehow have access to the situation of co-workers in other locations. One job of
technologies in such settings is to resolve this problem through the
reconfiguration of relevant spatial and temporal relations. That is to say,
information and communications technologies make it possible to maintain one
site as central by providing connections from that place to activities located
elsewhere, and by tracking those activities against a standardized temporal order.t
The following view of an Operations room shows some of these technologies:
5JoAnne Yates (1989) provides a history of the place of information and

communications technologies in maintaining centralized coordination and control
over increasingly distributed and time critical operations within U.S. railroads and
manufacturing enterprises, from the middle of the 19th century .
46
Lucy Suchman
Fig. 2.1. View of the Operations room
Along one wall of the room is a row of video monitors that feed images from
cameras located at each of the airline 's gates into the ro om. By seeing the image
of an airplane in a particular mon itor, mapping the monitor to a gate, and
mapping the gate in turn to a flight number , workers can track the status of
arriving and departing flights.
The association of gates with flights is supported through additional
representational devices. Located just below the row of monitors is the complex
board. The complex board is a whiteboard on which is drawn a space/time grid,
each cell of which is filled with an arriving and departing flight number, along
with its point of origin or destination. As Latour has argued (1990), the creation
and use of a device like the complex board involves the alignment, in a twodimensional, manipulable array, of a number of different spatial relations,
temporal orders and heterogeneous objects distributed across great distances.
Through its matrix structure, the complex board provides a graphic surface on
which operations workers can juxtapose and relate gates, times , flight numbers,
and other sites in the airline network . The further association of a clock then
makes it possible to relate the order represented on the board to the observable
order that can be viewed in the video monitors. Finally, computer terminals
provide access to a network of databases shared among the distributed sites of the
airline 's operations , representing the scheduled order from a variety of views, and
continuou sly updated to reflect the actual course of unfolding events.
47
Reading a Scene6
A central finding of our analyses concerns the extent to which the work of
Operations involves the assembly of knowledge about past, present, and future
events through the juxtaposition and relationship of a diverse range of
technologies and artifacts. Access to information and its timely communication to
relevant others involves interaction not with a single technology but rather with
multiple technologies (e.g., forms, computer screens, video monitors) held in
relation to each other and read off in ways specifically structured by the task at
hand.
The reading of these technologies does not proceed in a unidirectional sequence:
rather, they can be juxtaposed as needed in order to add to or elaborate on whatever
information already happens to be in hand (see M. Goodwin, 1995; Goodwin &
Goodwin, in press .) Workers address inquiries to these information resources not
from some neutral starting place, but always from a position within the midst of
an ongoing situation. So, for example, experienced Operations room workers,
knowing the schedule for a particular flight , can locate the associated plane on a
video monitor and read the activities around it for the flight's progress.
Alternatively, with a given plane in view on the video monitor, they can associate
that plane with a particular flight in order to make sense of the image they see.
Competency in the Operations room involves learning how to read a scene
through the juxtaposition and interpretation of verbal reports , visual images , and
various forms of text, in real time, into provisional assessments of an emerging
situation.
The example that follows is drawn from a chapter by Charles and Marjorie
Goodwin (in press) where it is extensively and elegantly analyzed. It concerns the
work of a position in the Operations room called the Flight Tracker. The Flight
Tracker is charged among other things with tracking arriving and departing
airplanes, maintaining communications with pilots on the ground via a radio, and
clearing their arrival at designated gates.
On the simplest account, the incident begins with a report from an incoming
pilot that another plane is already parked at the gate to which he has been
assigned . The report implicitly identifies a problem, and the question locates the
solution to that problem with the Flight Tracker":
6This phrase is due to Goodwin and Goodwin (in press). "Reading a scene" is closely
related to what Heath and Luff (1992) describe under the heading of "overseeing the
local environment" (p. 83).
7In transcript segments, colons U:" indicate prolongation of the immediately
preceding sound; italics mark stress . A dot in paretheses "(.)" indicates an untimed
pause ; numbers in parentheses indicate elapsed time in seconds . An equal sign "="
indicates "latching," that is, the beginning of one utterance following directly on the
end of the prior utterance with no gap.
48
Pilot :
Flight
Tracker:
Lucy Suchman
I understand gate fourteen is occupied?
Do you have any instructions for (it)?
Uh : . m,
(0 .3)
(0 .1)
Should've left ten nlinutes ago .=

Hopefully : ,
(1.0)
Pilot :
They have pulled the passenger stairs.

They should be leaving momentarily .
O:kay, thanks .
In their analysis of this sequence, the Goodwins observe that, to understand the
skills that the Flight Tracker deploys in doing her work, we need to begin with
her actions as hearer during the course of the pilot's calL We can see this clearly
by considering a transcript of the same sequence, this time with an indication of
the Flight Tracker's orientation during the exchange with the pilot, as indicated by
her gaze:8
Pilot :
I understand gate [Flight Tracker's gaze

goes to video monitors] fourteen is
occupied?
Do you have any instructions [Flight
Tracker's gaze moves to the radio log,
listing flight arrival and departure
times, beside her workstation] for (it)?
Flight
Tracker:
(0.3)
[looking through radio log] Uh: : m , [gaze

to a Flight Information Display monitor
above her workstation ] ( 0 . 1 )
[gaze back to radio log] Should've left
ten nlinutes ago .=

[gaze
back
to
video
monitors]
Hopefully: ,
(1.0)
Pilot :
They have pulled the passenger stairs .

They should be leaving momentarily .
O:kay, thanks.
8Charles and Marjorie Goodwin among others have developed innovative means of
indicating non-vocal activity within transcripts (e.g ., see Marjorie Goodwin's
transcription of this sequence in Goodwin & Goodwin, in press, and also Heath & Luff,
1993, and Luff & Heath, 1993). Here I have adopted the simple strategy of inserting
textual descriptions of the Flight Tracker's orientation within brackets, in boldface,
into the transcript.
49
With the pilot's mention of gate, the Flight Tracker is already orienting to the
video monitors . In her look to the monitors , we see her beginning a course of
action that involves the juxtaposition and relation of multiple, partial perspectives
onto the scene. Specifically, she finds the location referred to by the pilot as it
appears on a video monitor within her work site, maps the plane that she sees
there to an associated flight number by means of a paper record of scheduled
arrivals and departures, compares the observable plane and current time with the
represented status of that flight in the record and on a flight information display
screen in front of her, finds an unreconciled disparity between actual and
represented events, gazes again at the monitor, and there finds the actual situation
to have changed in such a way (the workers on the ramp having pulled the
passenger stairs away from the plane) that she can project a resolution (the plane's
imminent departure .) Her vantage point in operations does not provide her with a
single master perspective, in sum, but rather with a range of partial information
resources with which she can assemble a coherent view.
(Re)producing a Normal Order
The Flight Tracker's projected resolution to the problem makes reference to an

expected sequence of events and brings us into the presence of a theme that is
central, particularly within ethnomethodological studies, to analyses of
organizational settings, namely, the accountable (re)production of normal orders.
The term accountability has a useful ambiguity within ethnomethodology ,
referencing two senses of the moral grounding of everyday activity. In the most
basic sense, our viability as members of the social world turns on our mutual
intelligibility, that is, on our ability to make sense of the actions of others and
make ourselves sensible to them. In this sense our accountability with respect to
our actions means just that we are responsible for their intelligibility in relation
to relevant circumstances . This is not to say, of course, that we always do act
sensibly, or that questions of sense and significance will necessarily arise, or that
meanings cannot remain highly uncertain and/or contested. It is to say that as
social actors we are unavoidably implicated in such practices of sense making.
At the same time, we are also enjoined within more specific, historically
constituted orders of accountability, enforced through more and less explicit
regimes of administration and control. Airline operations is one such
institutionalized , power-differentiated social world. Within the work of operations,
organization members are accountable for the relation of normative rules to
observable/reportable events (Garfinkel & Sacks, 1970). On this premise,
traditional sociological theorizing takes normative rules, such as the order of
flights prescribed by the airline schedule, as explanations for the social order
observed in airline operations . Ethnomethodology, in contrast, takes the positing
of normative rules not as a resource for sociological explanation but as a (often
problematic) resource for members in their interpretation of the activities in which
50
Lucy Suchman
they are engaged. So we see here that the Flight Tracker makes recourse to the
schedule and to an expected order of events at the gate, while at the same time she
implicitly acknowledges the kind of order that it is. Specifically, she notes in her
hopefully and in her use of the modal should that there is no necessary relation
between schedules and the timely movement of airplanes, or between stair-pulling
and momentary departure. Rather, the normal order is contingent on its faithful
reproduction through the artful practices of personnel dedicated to its accomplishment, on each next occasion and across the unique and unpredictable circumstances
that each occasion provides . It is on that artful accomplishment that the Flight
Tracker's expectation of a projected resolution to the problem relies .
Jordan (1992c) looks at the management of trouble with an interest in the
relation between routine and improvisation in the work of Operations. In
particular, she develops the notion of a typified action sequence as a way of
talking about the projectability that organizational routines provide . Her analysis
demonstrates that the reproduction of organizational routines is not provided for
from outside but rather is a continuous accomplishment from within the local
group . At the same time, within the group, members orient to a normal course of
events, as the framework into which unexpected contingencies can be absorbed and
through which they can be managed.
M . Goodwin (in press) explores in detail the workings of such action sequences
as a matter of resources used by Operations workers to design their talk for the
specific environment that the room affords, in particular, for the troubles that it
presents for hearers . Taking as a case in point the routine delivery of informings
and announcements, Goodwin's analysis shows how workers employ various
forms of stylized intonation reminiscent of ritual speech in order successfully to
convey these utterances through the "sonic soup" from which their intended
recipients must pick them out.
Suchman (1993) takes up the theme of the accountability of practical action and
relates it to the use of technologies within the work of Operations. The phrase
technologies of accountability is meant to take advantage of the term's ambiguities to suggest two senses in which we might pursue the investigation of
technologically mediated practice. In the first sense, the competent use of
technologies in Operations supports the production of an accountable course of
intelligible and effective action. In the second sense , the technologies and actions
of Operations incorporate regimes of control designed to administer a particular,
institutionalized spatio-temporal order.
Structures of Participation
We noted that the social organization of work within the Operations room
involves a division of labor that assigns to each position different responsibilities
for communication, via various technologies , with other relevant locations. One
consequence of this organization is a differentiated structure of attention among
51
workers within the room. The most compelling evidence for this differentiation is
the ability of participants to disattend sights, sounds, and events that draw the
attention of their co-workers , and to keep on with their business at hand. At the
same time, members of the operations room are engaged in the highly integrated,
joint task of getting planes smoothly in and out of gates according to specified
schedules. This collective responsibility means that they are in some very
important sense "in it together," both in the sense of sharing responsibility for
the work's success or failure and sharing the facilities through which the work
gets done. The coordination required within the center means that Operations room
work is characterized by a strong mutual orientation among co-present workers to
each other and to developing situations .
The interactional order of the Operations room is characterized as well by a
continuous state of incipient talk. The peripheral monitoring made possible by
their co-presence in a shared auditory and visual space means that at any time
something overheard in the work of another may be assessed for its relevance and
taken up as the business of the hearer. So, for example, a question asked "of the
room" or even of a specific co-worker may be answered by anyone who has an
appropriate response (Jordan, 1992c). Similarly, people are oriented to the
possibility that they or their co-workers may initiate an interaction at any time
without any marked pre-announcement or inquiry into the others' availability for
that talk (Brun-Cottan, 1991). The multi-activity nature of the setting means that
the one who would initiate an interaction needs to do so in a way that is sensitive
to the engagement of others, for example, through the utterance's placement or, as
an alternative, through its intonation (M. Goodwin, in press).
Another sequence analyzed by Marjorie Goodwin (1995) involves an incident
similar to that of Goodwin and Goodwin (in press); that is, an arriving pilot has
been assigned to a gate that is already occupied. Again we see the Flight Tracker
(Fl') receive a call from the pilot and consult the video monitors and complex
board to establish, first, that there is an airplane at the assigned gate, and then its
flight number and scheduled departure time. In this case, however, the projected
resolution of the problem is provided not by what the Flight Tracker can read of
the activities around the plane but by a reading done for him by the Ramp Planner
(RP), a co-worker in the Operations room responsible for directing activities
around the plane . The Ramp Planner is retrospectively found to be attending to
the call-in from the pilot of Flight 1091 and to the Flight Tracker's resulting
problem:
Pilot :
RP :
(via radio) Operations.

Atlantic ten ninety-one's on the ground for gate
seven.
[Turns gaze towards monitor bank]
(2.5)
52
FT :
RP:
FT:
Pilot :
FT :
RP :
FT :
Lucy Suchman
[Looks to monitors]
(To pilot vi a radio ) Roger, ten ninety-one .
Charlie- Alpha : seven:? uh : : :, (0 .2) Shoo : :
(0.9)
Hold on one second, ten ninety-one .
(Off radio, into the room) Alpha seven,
[Shifts gaze from monitors to complex board]
That plane should be pu shin.
[Re-orients back to radio] (To pilot) That aircraft
should be off the gate shortly .
St and by until seven clears, Ten ninety-one .
Roger . Could you tell them we're gonna need
ground power please .
That's affirm . [Gaze toward RP] Did you catch that,
Joe?
Yeah .
Okay .
In this case, it is the Flight Tracker who, in response to the radio call from the
pilot of Flight 1091, discovers, through a look to the video monitors, that
another plane is occupying the gate to which Flight 1091 has been assigned. The
Ramp Planner evidently sees this problem as well. Moreover, the division of
labor in this case provides the Ramp Planner with a different history and
consequently different ability to interpret the state of affairs at the gate; this
difference is a resource here because he, like the Flight Tracker in the previous
example, offers a projected solution based on his reading of the scene. The Flight
Tracker, in turn, evidently reasons that the Ramp Planner, having just displayed
his attention to the call, will have continued to monitor its course and will hear at
its close a request from the pilot for an electrical power source. He figures that,
but he is not sure. Given this uncertainty, the Flight Tracker fulfills his
responsibility to pass on the pilot's request to the Ramp Planner not by
delivering it as news, but by asking the Ramp Planner to confirm that it has, in
fact, been heard.
Workers in the Operations room are sensitive to the environment for
communication that the room provides and to changes in that environment from
one moment to the next. Much of the time, workers are able to track the activities
of their co-workers or to be enlisted into them at the same time that they respond
to the demands of their own position . By listening not only for events to which
they need to respond but for the responses of others as well, minor troubles in
communication are routinely identified and resolved? In some situations, the
supervisor effectively acts as an additional pair of eyes and ears, charged
specifically with watching for the possibility that events may be missed (see
Suchman, in press) .
9See Heath and Luffs discussion (1992) of surreptitious mon itoring and rendering
activities visible for a related analysis .
53
Constituting Workspaces
The activities of Operations are distributed across spaces separated by distances

ranging from several feet to many thousands of miles. With the advent of new
information and communications technologies , the composition of the working
group no longer correlates just with physical proximity. So, for example, two coworkers seated side by side may at one moment be more closely engaged with
parties in other locations than with each other. Our focus on ground operations
allowed us to see how a work group distributed in space is tied together through
architectural, technological, and interactional resources, as well as the obstacles
that such a group must face.
A starting premise for our analyses is that workspaces are not simply physical
locations but are also actively constituted fields of perception and interaction,
continuously maintained over the course of the day's work. Through the spaces
they occupy, people identify themselves and their place within the organization .
Within Operations, each role in the division of labor is mapped to a specific
location in the room, tied in turn to other locales outside. What are called
positions capture nicely this double sense of location in the organization of the
work and in the arrangement of space. In this sense, the social relations of the
work both defme the visible and invisible boundaries of the workplace and
are defined by them.
The division of labor in Operations is reflected in an arrangement of habitual
workspaces, each oriented toward the equipment that ties personnel within the
room to their co-workers in other relevant locations. The stability of these
workspaces means that members are able to project the location both of their
colleagues and of specific pieces of equipment. The projectibility of habitual
spaces contributes to the ease with which a task can be taken up or handed off
from one co-worker to another.
Although the division of labor in Operations differentiates people's attention,
the absence of interior walls or other fixed boundaries within the room maximizes
mutual access. And while Operations personnel inhabit habitual workplaces, those
places are not in any sense owned by the individuals who occupy them .
Equipment associated with one position may be borrowed by an Operations
worker at another position. Alternatively, one worker may take up the position of
another if the latter is temporarily absent or otherwise occupied. Within the
Operations room, boundary markers that mark the line between adjacent territories
are notably absent. Boundaries are defined more by the placement of people and
equipment and by the dynamic structuring of activity, than by explicit
designations of ownership .
54
Lucy Suchman
Through the application of previous work on multi-party interaction , 10 we have

engaged in detailed, systematic analyses of just how members of the Operations
room coordinate their separate and joint activities . Suchman (in press) examines
the interactional constitution of shared workspaces. In the complex ways that
members organize their respective workspaces both individually and in relation to
the ongoing activity of the room , the Operations room as a single, shared workspace can be more adequately understood as a place for the successive divergence,
convergence, and re-alignment of multiple, shifting lines of activity. This is the
process we propose to call the constitution of sharedworkspaces. Talk, gaze, body
position, gesture, space, furnishings , and equipment can all be viewed as resources
in its accomplishment.
Acquiring Competency
Our interest throughout the project was to analyze interactions within Operations
not by recourse to a master plan or external order, but rather with reference to the
activity's ordering from within, through the sensitivities of the participants to
each other and to their joint situation . However, to say that social order is
produced from within, through the local interactions of participants, is not to say
that it is produced without any reference to previous activity . One way in which
we can see the presence of a historically and culturally constituted community
into which people come is in the process of learning the work (Lave & Wenger,
1991) . We can take as an example a routine task within ground operations: that
is, the work of establishing what are called an airplane's weights and balances.
Weights and balances involve a relation between the plane's total complement of
passengers, baggage, fuel, and other cargo and the settings required for the
stability of the aircraft (e.g., the wing flaps). On many newer aircraft, the final
weights and balance calculation is taken by the pilot directly from a computer in
the cockpit. For older aircraft, however , the numbers are called up on a computer
screen in the Operations room, then radioed out to the pilot in an exchange called
a radio close-out, a routine part of the Flight Tracker's work.
Charles Goodwin (1991) analyzed a case involving an apprentice in the work of
Flight Tracking doing her first radio close-out , supported by an experienced coworker. In his analysis, Goodwin draws attention to the exquisite coordination of
apprentice and teacher in the course of the call. Specifically, he shows how the
more experienced co-worker momentarily "redesigns" the computer screen for the
apprentice in response to her demonstrated difficulty in finding her way through,
by directing her reading of it with his finger (see Figure 2). Goodwin's analysis
10Por example, the constructs recipient design (M. Goodwin, 1980; C. Goodwin,
1981 ; Schegloff, 1972) and part icipation structures (Goodwin & Goodwin, 1989)
underwrite almost all of our analyses. (See also Atkinson & Heritage, 1984; Heath,
1986 .)
55
shows how, through their precise attunement to each other's place in the course of
action and its possible problems, apprentice and teacher together achieve a
successful reading.
Fig. 2.2. Flight Tracker and apprentice
Competent participation in the work of operations involves learning how to see

one's environment in an informed way. Whether in situations of explicit
instruction or embedded within interactions among co-workers identified as peers,
members' ability to bring their differentiated expertise to bear on the situation at
hand is tied to their access to each other's activities and interactions. In this sense,
the overall effectiveness of the working group is tied to the relations between its
members, and to the opportunities for teaching and learning that the common
environments of their work provide .
Authoritative Knowledge
Through a comparative analysis of an American obstetrics ward and the
Operations room , Jordan (1992c) explores the relation between access to and
control over technologies, and the forms of knowledge taken as authoritative: that
56
Lucy Suchman
is, as "grounds for legitimate inference and action" (p. 1) within a given situation .
In contrast to the highly specialized and power-differentiated order of the hospital
setting, she finds familiarity with relevant technologies to be distributed across
participants in Operations, with a corresponding distribution of knowledge
required to get the work done.
As an example, Jordan traces the work on a particular afternoon in Operations
required to effect a switch of airplanes between gates, in order to enable
mechanical repairs to one of the planes. This involves towing an airplane that has
arrived at one gate to a different gate for departure, as well as a complex reassignment of crew and transfer of passengers and baggage. Jordan points to the
relatively "horizontal" distribution of knowledge with respect to the work
required to accomplish this rearrangement, including the use of relevant
technologies. She argues that this distribution of knowledge is both produced by
and helps to explain the frequency of "out louds," or apparently undirected
comments on the situation, and questions asked "of the room," over the
maneuver's course (see also M. Goodwin, in press). That is, workers recognize
more or less implicitly that information about events may be relevant to anyone
or more of their coworkers at any given time, just as one or more of their
coworkers may have answers to a question, given their distributed access to
available information resources and social networks. Jordan concludes by
proposing an agenda of comparative analysis of the distribution of authoritative
knowledge across different settings, involving differently privileged or accessible
technologies.
Designing for Change
Our findings on relations of work and technology imply that , in order to design
anyone aspect of a working order, one needs to understand that aspect's
relationship to the extended system of activities and technologies of which it is a
part. At the same time, products of professional design will always be based in
partial, specifically situated and historically constituted projections of the
circumstances of an artifact's use. As a consequence, professional design needs to
be understood not as an end point but as a starting place, or platform, for the
ongoing processes of design-in-use that are both inevitable and necessary for an
effective working environment.
The specific processes of change that we observed in the Operations rooms
illustrate more general characteristics of workplace design . Rather than developing
according to a single master plan, changes in facilities and work practices arise
from participants' emerging and, to some extent, idiosyncratic appreciation for the
problems and possibilities that a particular setting affords. First impressions of a
facility change as one goes to work in it. Requirements are discovered through the
contingencies of everyday use.
57
At the same time, a change in anyone aspect of a working environment will

usually have repercussions elsewhere. At the airport of our study, moving ground
operations to a new location implied transformations in other sites, in
technologies, and in the structuring of work activities. In the case of the airline's
Operations room for short range, commuter flights, for example, the move to the
new terminal involved a shift from the management of arriving and departing
airplanes within a large, open parking area to their maneuver through a relatively
narrow roadway into and out of the gate area. The requirements on moving planes
through this roadway led to a change in the schedule of flights for the entire
airline, as it became necessary to reschedule in order to avoid simultaneous
arrivals and departures for commuter flights at this particular airport. Brun-Cottan
et al. (1991) provide further examples.
The interconnectedness of systems means that design of anyone aspect of a
particular worksite potentially implicates an open horizon of other aspects of that
site, as well as that site 's relations to a more extended network of settings and
activities . However comprehensive and well-founded our analyses as researchers or
designers, new understandings and new requirements will continue to emerge for
those who actually do the work. It follows that they are the ones best qualified to
carry forward and fine tune the design of the workplaces they inhabit. Design in
use is always a process of improvisation, of making the best of what you have .
The role of professional design is to lay the groundwork for that process, to
provide the tools that support it, and to learn the lessons that are taught by it.
Conclusion
The analyses reported here draw on a rich set of materials that make evident the
dynamic , moment-to-moment structuring of coordinated work activities in a
complex, distributed, technology-intensive workplace. The accumulation of
studies that locate regularities of everyday interaction within specific sites of work
practice provide the materials from which richer tapestries of research and
theorizing can be woven . Moreover, others engaged in the study of what I am
calling here centers of coordination have proposed generalizations that intersect
with those that I have presented. There is every indication that comparative
analysis across such sites would more than repay our efforts . For example, we
might compare and contrast workers' relative expectations regarding the identity
and circumstances of callers into the center in the cases of the Operations room
and a 9-1-1 public safety center, and the consequences for the problem of turning
calls into organizationally relevant objects (Whalen, 1992, 1993; Whalen &
58
Lucy Suchman
Zimmerman, 1990; Suchman & Whalen , 1994).11 An orientation to the problems

of turning calls into organizationally relevant objects , correspondingly, might lead
to different requirements on the design of communications and documenting
technologies in the two settings . Similarly, common practices across all of these
sites include overhearing , outloud comments and mutual monitoring, the recipient
design of announcements, what Heath and Luff (1991) term rendering tasks
visible, various ways of structuring one's own attention and directing the
attention of others, and the intricate structuring of talk and other forms of activity .
We might also compare the central artifacts of these sites, for example, forms of
various kinds and the dynamics of their creation and use across paper and digital
media. We might explore how workers in these sites employ available artifacts to
maintain an accountable relation between normal orders as represented by
schedules, protocols, and the like and the contingencies of actual events . Finally,
we might compare processes of change in technologies and associated divisions of
labor across these sites and their implications for those interested in professional
design.
The listing offered here is meant only as a preliminary suggestion of what the
fruitful lines of comparison might be; it is a listing that, I trust, could easily be
elaborated by any readers familiar with the research. A wider reading of these
studies in relation to each other could yield a reconciliation of topics that organize
the cumulative findings. I hope, by enumerating the themes of one particular
project, to contribute at least indirectly to that larger synthesis .
Acknowledgment
I am grateful to Charles Goodwin, Marjorie Goodwin, and Randy Trigg for their
reading and comments on earlier versions of this paper. Lauren Resnick and
Clotilde Pontecorvo provided further thoughtful review and suggestions .
IlThanks to Randy Trigg for suggesting the interest of this particular line of
comparison .
59
References
Atkinson, M., & Heritage, J. (Eds .), (1984). Structures of social action. Cambridge,
England: Cambridge University Press.
Barley, S. (1986) . Technology as an occasion for structuring: Evidence from
observations of CT scanners and the social order of radiology departments .
Administrative Science Quarterly, 31, 78-108 .
Blomberg, J. (1987). Social interaction and office communication: Effects on user's
evaluation of new technologies. In R. Kraut (Ed.), Technology and the
transformation of white collar work (pp. 195-210). Hillsdale, NJ: Lawrence
Erlbaum Associates .
Bowers, J. (1994). The work to make a network work. In R. Furuta & C. Neuwirth
(Eds .), Proceedings of the Conference on Computer-Supported Cooperative Work
(pp. 287-298) . New York: ACM Press.
Brun-Cottan, F. (1991) . Talk in the workplace : Occupational relevance. Research on
Language in Social Interaction, 24, 277-295 .
Brun-Cottan, F., Forbes , K., Goodwin , C., Goodwin, M., Jordan, B., Suchman, L., &
Trigg, R. (1991) . The workplace project: Designing for diversity and change
(Videotape). Palo Alto, CA: Xerox Research Center.
Clark, A., & Fujimura, J. (Eds.). (1992). The right tools for the job: At work in
Twentieth-Century life sciences. Princeton, NJ: Princeton University Press.
Dugdale, A., & Fujimura, J. (Eds.). (In prep) . Making sex, fabricating bodies: Gender
and the construction of knowledge in the biomedical sciences.
Filippi, G., & Theureau, J. (1993) . Analyzing cooperative work in an urban traffic
control room for the design of a coordination support system. In G. Michelis, C .
Simone, & K. Schmidt (Eds.), Proceedings of the Third European Conference on
Computer-Supported Cooperative Work (pp. 171-186) . Dordrecht, The Netherlands :
Kluwer.
Forsythe, D. (1993). Engineering knowledge: The construction of knowledge in
artificial intelligence. Social Studies of Science, 23, 445-77 .
Garfinkel, H., & Sacks, H. (1970). On formal structures of practical actions. In J .
McKinney & E. Tiryakian (Eds.), Theoretical Sociology (pp. 337-366). New York:
Appleton-Century Crofts.
Goodwin, C. (1981) . Conversational organization: Interaction between speakers and
hearers. New York: Academic Press.
Goodwin, C. (1991) . Situated literacy in the workplace. Colloquium, University of
California, Los Angeles, Department of Applied Linguistics, February 21, 1991.
Goodwin , C. (1994) . Professional vision . American Anthropologist, 96(3), 606633 .
Goodwin , C., & Goodwin, M. (1989) . Context, activity and participation. In P. Auer
& A. diLuzo, (Eds.), The contextualization of language (pp. 77-99) . Amsterdam:
Benjamins.
Goodwin, C., & Goodwin, M. (in press) . Formulating planes : Seeing as a situated
activity . In Y. Engestrom & Middleton, D. (Eds.), Communication and cognition at
work. New York : Cambridge University Press.
60
Lucy Suchman
Goodwin, M. (1980). Processes of mutual monitoring implicated in the production of

description sequences. Sociological Inquiry, 50, 303-317.
Goodwin, M. (1995 ). Assembling a response . In P. ten Have, & G. Psathas (Eds .),
Situated order: Studies in the social organizationn of talk and embodied ac tiv ities
(pp . 173-186 ). Washin gton , DC: University Press of America.
Goodwin, M. (in press). Announcements in their environm ent: Pro sody within a
multi- activity work se tting. In E. Couper-Kuhlen & M. Selting (Eds.), Prosody in
conversation: Interactional studies. Cambridge, England: Cam bridge University
Pre ss.
Harper, R., & Hughes , J. ( 1993). "What a F-ing System ! Send 'em all to the same
place and then expect us to stop ' em hitting": Making techn ology work in air
traffi c control. In G. Button (Ed.), Technology in working order: Studies in work,
interaction and technology (pp. 127-144). London : Routl edge.
Harper, R., Hughe s, J., Randall, D., Shapiro, D., & Sharrock, W. (in pre ss). Ordering
the skies: Sociology, CSCw, and air traffic control. London : Routledge.
Hartland, J. (1993) . The use of "intelligent" machines for electrocardiograph
interpretation. In G. Button (Ed.), Technology in working order: Studies in work,
interaction and technology (pp. 55-80). London: Routled ge.
Heath, C. (1986) . Body movement and speech in medical inte raction. Cambridge,
Engl and: Cambridge University Press.
Heath, C., Jirotka, M., Luff, P., & Hindmarsh, J. (1994) . Unpacking collaboratio n:
The intera ctional organization of trading in a city dealing room. Journal of
Computer-Supported Cooperative Work (CSCW, 3(2), 147-165.
Heath, C., & Luff, P. (1992) . Coll aboration and contr ol: Crisis management and
multimedia techn ology in London Undergound line control room s. Journal of
Computer-Suppo rted Cooperative Work (CSCW), 1(1), 69-94.
Heath, C., & Luff, P. (1993) . Disembodied conduct: Interaction al asymmetri es i n
video -mediated communicatio n. In G. Button (Ed.), Technology in work ing order:
Studies in work, interaction and technology (pp. 35-54). London : Routledge.
Hutchins, E. (1990). The technology of team navigation. In J. Gallagher, R. Kraut, &
C. Egido (Eds.), Intellectual teamwork (pp. 191- 220). Hillsdale, NJ: Lawrence
Erlb aum Associates.
Hutchin s, E. (199 1). The social org anization of distributed cog nitio n. In L. B .
Resnick, J. M. Levine , & S. D. Teasley (Eds .), Perspectives on socially shared
cog nition (pp . 283-307). Washington, DC: American Psychologi cal Association .
Jordan, B. (1990) . Teamwork in an airlines operations room: A skills scenario . In
Identifying and describing the skills required by work (pp. A5-A7) . Report of The
Secretary's Commission on Achieving Necessary Skill s (SCANS). Washington,
DC: U.S. Department of Labor.
Jord an, B. (1992a). Birth in fo ur cultures: A crosscultural investigation of childbirth
in Yucatan, Holland, Sweden and the United States (4th expanded edition, revised by
Robbie Davis-Floyd .) Prospect Heights , IL: Waveland Press.
61
Jordan, B. (1992b). New research methods for looking at productivity in knowledgeintensive organizations. In H. V. D. Parunak (Eds.), Productivity in knowledgeintensive organizations: Integrating the physical, social, and informational
environments (pp . 194-216). Ann Arbor, MI: Industrial Technology Institute.
Jordan, B. (1992c). Technology and social interaction: Notes on the achievement of
authoritative knowledge in complex settings (IRL Technical Report No. IRL920027) . Palo Alto, CA: Institute for Research on Learning .
Knorr-Cetina, K., & Mulkay, M. (1983) . Science observed. London: Sage.
Latour, B . (1987). Science in Action. Cambridge, MA: Harvard University Press.
Latour, B. (1990) . Drawing things together. In M. Lynch & S. Woolgar (Eds.),
Representation in scientific practice (pp. 19-68). Cambridge, MA: MIT Press.
Latour, B., & Woolgar, S. (1979). Laboratory life. Beverly Hills, CA: Sage.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation.
New York: Cambridge University Press .
Luff, P., & Heath , C. (1993) . System use and social organization: Observations 0 n
human-computer interaction in an architectural practice. In G. Button (Ed .),
Technology in working order: Studies in work, interaction and technology (pp .
184-210). London: Routledge.
Lynch, M., Livingston, E., & Garfinkel, H. (1983) . Temporal order in laboratory
work . In K. Knorr-Cetina & M. Mulkay (Eds.), Science observed (pp. 205-238).
London: Sage.
Lynch, M., & Woolgar, S. (Eds.). (1990). Representation in scientific practice.
Cambridge, MA: MIT Press .
Orr, J. (1990) . Talking about machines: An ethnography of a modem job. Doctoral
Dissertation, Cornell University. (PARC Technical Report SSL-91-07 [P9100132]). Palo Alto, CA: Xerox PARC o
Pickering, A. (Ed.) (1992) . Science as practice and culture. Chicago : University of
Chicago Press.
Schegloff, M. (1972) . Notes on a conversational practice: Formulating place . In D.
Sudnow (Ed.), Studies in social interaction (pp. 75-119) . New York: Free Press .
Sharrock, W., & Anderson, B. (1993) . Working towards agreement. In G. Button
(Ed.), Technology in working order: Studies in work, interaction and technology
(pp. 149-161). London : Routledge.
Star, S. L. (1989) . Regions of the mind: Brain research and the quest for scientific
certainty. Stanford, CA: Stanford University Press .
Star, S. L., & Ruhleder, K. (1994) . Steps towards an ecology of infrastructure. In R.
Furuta & C. Neuwirth (Eds.), Proceedings of the Conference on Computer-Supported
Cooperative Work (pp. 253-264). New York: ACM Press .
Suchman, L. (1983) . Office procedures as practical action: Models of work and system
design. ACM Transactions on Office Information Systems, 1(4), 320-328.
Suchman, L. (1987) . Plans and situated actions: The problem of human-machine
communication. Cambridge, England : Cambridge University Press .
Suchman, L. (1993) . Technologies of accountability. In G. Button (Ed.), Technology
in working order: Studies of work, interaction and technology (pp. 113-126).
London : Routledge.
Suchman, L. (in press) . Constituting shared workspaces. In Y. Engestrom & D.
Middleton (Eds.), Communication and cognition at work. New York : Cambridge
University Press .
62
Lucy Suchman
Suchman, L., & Trigg, R. (1991) . Understanding practice : Video as a medium for
reflection and design . In J. Greenbaum & M. Kyng (Eds.), Design at work :
Cooperative design of computer systems (pp. 65-89) . Hillsdale , NJ: Lawrence
Erlbaum Associate s.
Suchman, L., & Trigg, R. (1993) . Artificial intelligence as craftwork. In S. Chaiklin
& J. Lave (Eds.), Understanding Practice (pp. 144-178). New York: Cambridge
University Press.
Suchman, L., & Whalen, J. (1994, November) . Standardizing local events and
localizing standard forms. Paper presented at the annual meetings of the Society for
Social Studies of Science, New Orleans, LA.
Traweek, S. (1988) . Beamtimes and lifetimes : The world of high energy physicists .
Cambridge, MA: Harvard University Press .
Whalen, J. (1992) . Technology and the coordination of human activity : Computeraided dispatch in public safety communications. (Unpublished report) . Eugene, OR:
Univers ity of Oregon , Department of Sociology .
Whalen , J. (1993, August) . Accounting for "standard" task performance in the
execution of 9-1-1 operations. Paper presented at the annual meetings of the
Ameri can Sociological Association, Miami, FL.
Whalen , M., & Zimmerman, D. (1990) . Describing trouble: Practical epistemology in
citizen calls to the police . Language in Society, 19, 465-492 .
Yates , J. (1989) . Control through communication : The rise of system in American
management. Baltimore and London: Johns Hopkins University Press .
Chapter 3
Animated Texts:
Selective Renditions of News Stories
Christian Heath' and Gillian Nicholls'
'University of Nottingham and King's College London, UK
2Middlesex University, UK
"Shared Agreement" refers to various social methods for accomplishing the

member 's recognition that something was said according to a rule and not the
demonstrable m atching of substantive matters. The appropriate image of a
common understanding is therefore an operation rather than a common
intersection of overlapping sets.
H. Garfinkel (1967, p. 30)
The fact is that when the listener perceives and understands the meaning (the
language meaning) of speech, he simultaneously takes an active , responsive
attitude towards it. . . . and understanding live speech, a live utterance, i s
inherently responsive, although the degree of this activity varies extremely .
M. M. Bakhtin (1976, p.68)
Abstract
The following chapter is concerned with the ways in which journalists, working in
an international news agency, coordinate the production of news stories with each
other . In particular, it explores how journalists animate stories on which they are
working and thereby render their own activities visible to colleagues within the
newsroom. Although the analysis is princip ally concerned with the ways in which
textual embedded stories are voiced and shaped in and through talk and interaction,
it also considers , more generally, how tools and technologies feature in everyday
collaborative work .
64
Christian Heath and Gillian Nicholls
Introduction
In recent years, we have witnessed the emergence of a growing corpus of
naturalistic research concerned with the organization of work and communication
in complex technological environments. Although, in part, such research has
emerged in the light of rapid technological developments and the transformation of
the workplace, it has also been driven by a sophisticated critique of more
traditional analytic orientations to the interaction between the individual and the
computer (Winograd & Flores, 1986; Suchman, 1987). In particular, the growing
dissatisfaction with conventional plan-based, goal-oriented models of human
conduct and its concomitant assumptions concerning the nature of human
cognition and action have led to an interest in exploring the in situ organization
of work and in explicating the tacit and indigenous resources through which
individuals accomplish and coordinate their tasks and activities.
Despite the contributions of such research and, in particular, their substantial
body of findings concerning work and communication in settings such as air
traffic control (Harper & Hughes, 1993), ship navigation (Hutchins, 1990),
airport operation centers (Suchman, 1993; Goodwin & Goodwin , 1996),
emergency dispatch centers (Whalen, 1995), urban transport control rooms (Heath
& Luff, 1992), and financial dealing rooms (Heath, Jirotka, Luff, & Hindmarsh ,
1995), we still have relatively little understanding of the ways in which
technologies "mediate" the interaction between organizational personnel. It is clear
from these studies that new technologies not only reconfigure tasks and the
division of labor within the workplace, but also introduce unanticipated
asymmetries to the communication between organizational personnel. Although it
has long been recognized that working environments are founded on and embody a
wide variety of forms of communicative asymmetry, especially between
organizational personnel and their clients (see, for example, Markova & Foppa,
1990; Drew & Heritage, 1992), the ways in which real time, co-present
interaction is configured to manage differential access to specialized information
and tasks (accessed and performed through new technologies) remain to be
explored.
Take , for example, a document such as a medical record, which is held in
electronic form. During the course of the consultations, it is not unusual to find
patients attempting to coordinate their actions, such as utterances, with the
doctor's use of the computer as he enters data into the record (cf. Heath, 1986;
Greatbatch, Luff, Heath, & Campion, 1993). The patient may be unable to to see
the screen and the text itself and may only have a passing understanding of the
sorts of data that the doctor may be entering, such as diagnostic and prognostic
information. All the same, we find patients attempting to infer the character and
progression of the activity from such "visible" elements of the activity as the
rhythm of the key stokes , the movement of the hands, and slight shifts in the
doctor's gaze . These forms of peripheral sensitivity to the activities of others and
Animated Texts
65
their relevance to the coordination of technologically mediated conduct are being

given increasing attention, not only by those concerned with understanding
technologically mediated conduct, but also by designers involved in the
development of new systems. Indeed, in a recent and important paper, Brown and
Duguid (1994) have argued that both the centre and the periphery should be
considered in the design process and that phenomena such as "the skeletal rattle of
the keyboard" may be critical to the use of personal computers and their
coordination with the activities of others within the local domain.
How individuals within the workplace coordinate their actions with each other,
where they have differential access to both the activity and the sources of
information with which it is concerned, forms the focus of this chapter. Drawing
on data gathered within the financial section of an International News Agency in
London, we examine the ways in which journalists voice or animate texts and
thereby provide colleagues with various characterizations of the news on which
they are working. In particular, the analysis addresses the ways in which they
establish appropriate forms of co-participation for the selective rendition of
particular stories, and how these (re)tellings are relevant to the management
and distribution of news within the agency and, in consequence, to its clients .
The Setting
A number of international news agencies provide real time, screen-based
information to the financial sector as well as to other customers, including
television companies and newspapers. These include Blomberg, Nightrider, and
Reuters. Reuters is by the far the largest concern and has the most customers. It
has offices in most major cities throughout the world and coordinating centers in
London, Tokyo, and NewYork. In London, the Financial News Section of Reuters
is divided into four desks, each with its own editor, journalists, and subeditor(s).
These desks are Money and Capital, Equities, Oil and Minerals, and
Commodities. The desks are positioned near each other in a large open plan office.
The desks receive stories from the various offices throughout the world. The
stories are coded for a particular desk and appear within a "basket" on the editorial
screen. Journalists take stories from the basket, check their address and topic
coding, and edit the contents of the headline (the Header) and the story. The
corrected story is then transmitted or, in the case of longer pieces (say a couple of
pages or more), passed to the subeditor for final checking.
Although individual journalists on the various desks have a fair amount of
discretion in coding, editing, and prioritizing stories, and largely work on
particular items individually, they are sensitive to the interests of colleagues both
on their own desk and those working in other areas. For example, a news story
addressed to Money and Capital might also be relevant to the customers served by
66
the journalists in Equities . In many cases, different desks will transmit different
versions of the same story, even though the original text was only received by
one desk. At the morning editorial meetings, which review the previous day's
work and discuss the more important stories of the upcoming day, it is not
unusual to hear critical remarks concerning the failure of certain desks to pick up
on, or inform, others of potentially relevant stories . A case in point was a story
about a press release from Kleeneze concerning innovations in battery technology.
It was felt that the story should have been picked up by the Oil and Minerals desk
because the news had potentially important implications for lead consumption and
sales . It is important, therefore, that journalists remain sensitive to the interests
of colleagues working on other desks. They have to put their mind to the concerns
and relevancies of others (cf. Sacks, 1992) and to inform colleagues of incoming
news that might be relevant to their particular customers . It is also important,
however, that, given the enormous traffic in news stories handled by the various
desks, journalists do not bombard their colleagues with the details of all stories
they receive that might have some implications for customers handled by other
desks. Journalists need to discriminate among news stories with respect not only
to their own readers but also to the readers served by their colleagues. They need
to inform colleagues of potentially relevant stories but not overload them with the
information that they are receiving.
Journalists need to remain sensitive to the relevance of their own stories for
colleagues on other desks and to listen for potentially interesting items being
handled elsewhere; but they also have to work closely with colleagues in their
own area. For example, in passing stories on to subeditors, they might need to
provide some instructions concerning the ways in which the story should be
handled, or, in editing a particular item, it might be important to inform
colleagues of how the news is potentially relevant to other stories in the same
basket. Or, for example , decisions to make major changes to a story, its priority,
or topic coding may often be done in consultation with colleagues . A case in
point is spiking, a term still used in the electronic office. Journalists often talk
through stories with colleagues on the same desk before deciding to veto an item.
There is a strong orientation to keep colleagues informed, however indirectly, of
the stories on which you are working and to .let others know of the more
consequential decisions that you might be taking. In this way, while dealing with
a substantial amount of stories during the day, colleagues keep tabs on the more
significant items that are being handled by colleagues both on their desk and
within the news room. Thus Reuters provides a news service that is timely,
coherent, and reliable. Given the substantial amount of news pouring into London
during the day and being handled by the four desks, this is no mean achievement.
On the one hand, journalists work individually, handling and editing stories
received by the desk. On the other hand, they have to remain sensitive to the
interests of colleagues both on their own desk and on other desks. They need to
provide their colleagues with a sense of the work in which they are engaged and in
particular any potentially relevant stories ; but this needs to be accomplished
Animated Texts
67
unobtrusively, without interupting the activities of others. The situation is made

more complex still by the fact that the various news stories are not visible to all
within the room; indeed, they may not even be in the basket of stories that the
particular desk has received . In consequence, journalists have to make available the
details of stories that are displayed on screen without necessarily demanding that
colleagues read the story or even take any notice at all. And this has to be
accomplished without overwhelming colleagues with information that may be
utterly irrelevant to their readers or may have been dealt with under the guise of
another story. Textually embedded stories, located on an individual's work station,
have to be momentarily rendered visible, and this has to be achieved accurately and
unobtrusively.
Reading Aloud
It seems to be a feature of many co-present working environments in which there
is a sharp division of labor, yet a strong interdependence of tasks, that personnel
take it upon themselves to openly voice bits and pieces of information that may
be relevant to various activities within the domain . Although calling out and
shouting, perhaps best exemplified in the trading rooms of financial institutions
(cf. Heath et aI., 1995), appear, at least initially, to be addressed to all those
within audible range of the speaker, it soon becomes clear that such objects are
recipient selective or at least designed to be heard by particular categories of
personnel within the ecological domain . All the same, such objects are rarely
addressed to particular personnel; they do not demand a response or even that the
object is acknowledged; rather the information is there if it is needed or relevant.
By openly voicing, calling out various pieces and bits of information, speakers
are able to respect the current commitments of personnel within the domain while
providing information that may have some significance. Indeed, in contrast to the
impression of visitors to such environments, often overwhelmed by the shouting
and noise, it is in part the very fact that information is voiced in this way, that
allows personnel to retain some semblance of concentration and distance from
each other.
At Reuters, especially when major news stories are breaking, we find the sorts
of open voicing and calling out that arise in related environments in other
organizations. For the most part, however, information needs to be made available
among journalists working in a particular area or those based at adjoining desks.
Much of this information arrives within the agency, not through the telephone, as
it does in dealing rooms, control centers, and dispatch services, but in particular
baskets that are accessed on screen through the Reuters editorial system . For the
journalists, therefore, it is largely textually embodied stories that may be of
relevance to the work of colleagues . So while journalists are editing particular
68
stories, it is not unusual to find bits and pieces of news information being made
available to colleagues within the local milieu.
The fragments that follow are simplified examples of the original transcripts.
Where necessary, features of the original orthography are included, using a system
developed by Gail Jefferson. Colons indicate that the preceding sound is stretched,
underlines that the word or part of the word is emphasized, and numbers in single
brackets represent pauses or silences in tenths of a second. A full stop with
brackets represents a mini-pause, roughly a tenth or two tenths of a second.
Details of the orthograghy can be found in Atkinson and Heritage (1984)
and numerous other monograghs and collections on conversation analysis.
In the first instance , Peter makes a joke of a story he is editing on screen:
Fragment 1, Transcript 1
Peter :
Peter :
Peter :
Bank of (.) I s:ra : el interest :t ra(i)te droQ : .

(0 .3)
Down , down, down.
(0 .4)
Didn't it do this last week.
(13 .0)
In making his comments on the interest rate fall, Peter adopts a pronounced
Jewish accent. The remarks are not addressed to colleagues on his own desk
(Money and Capital) or to those on the adjoining desk (Equities). While talking
aloud, Peter looks at his monitor and continues to edit the story.
Peter's talk animates the story. It voices the text, or at least elements of the
story, and gives it the character of a jo ke. Peter' s remarks are loud enough to be
audible to colleagues on the surrounding desks, yet they are not addressed to any
particular colleague. The story is rendered visible, but does not demand or even
implicate a response. The way in which it is articulated avoids generating
sequential import; that is, no particular person is placed under obligation to
respond to the utterance. Even the question, "Didn't it do this last week,"
elaborates the joke and is not treated as sequentially implicative. The gist of the
story is voiced, but it does not place anyone within its audible range under an
obligation to respond.
Twelve seconds or so later, Alex, who is sitting some six feet away at the
Equities Desk momentarily changes his orientation . Peter utters "ar :::::" and after
one second delivers the following utterances. In the illustrations, Peter is second
and Alex third from the right:
Animated Texts
69
Fragment 1, Illustration 1 & 2, Transcript 2
(13 .0)
P:
er : : : (1 .0) Bank of Israel er.
P:
cut its er daily (0.4) the rate on its daily money

tender, (0 .2) to QQffiffiercial banks .
A:
Yeah (.) Got that now . ( .) Thanks Peter
P:
a .kay?
(3 .2)
(0 .6)
(0 . 6)
Peter's talk is now addressed to Alex. It is not simply a further gloss of the story
on which Peter is working, but rather a specific quote from the text itself. In its
delivery, Peter displays - through his orientation toward the screen, the pace of
the talk, and the self repair - that the information he is providing to Alex is an
extract from the news story he is currently editing.
The delivery sharply contrasts with the earlier version. It is not rendered as a
joke or as a precis but rather as part of the original, authentic story. The way in
which the talk is produced, coupled with the accompanying visual conduct,
provides colleagues with the resources to differentiate the status of the description
and in particular its relationship to the textual version of the story.
The delivery of the quote is occasioned by Alex's (re)orientation to Peter. The
(re)orientation is treated by Peter as inviting further information concerning the
story, as demanding an exposition of the story rather than continuing the joke.
Although the original joke is not specifically addressed, the informing is designed
for Alex and shaped to provide accurate and authentic information concerning the
recent shift in Israeli interest rates .
Even following Alex's acknowledgment of the story, Peter remains sensitive to
his colleagues' potential interest in and use of the item. In particular, some
seconds later, he reads aloud the sentence that describes the actual fall in interest
rates . By pausing in the delivery of the sentence, Peter momentarily renders the
description problematic and, on completing the sentence, goes on to make a
correction to the story. The correction is accomplished by the speaker realigning
70
his position to the text, from narrator to commentator. Peter differentiates his
version from the original text and publicizes , at least across the two desks, the
editorial correction.
Fragment 1, Transc ript 3
Peter :
Peter :
Peter :
Alex :
Half a percent, (1 .2) to eleven percent .

(0.2)
I think they mean a half a percentage point
(15 :04)
Service Jerusalem (0.5) with a .9,rQI2 copy to Nicosia,
right?
(0. 7)
Yes
Finally, Peter marks the finish of business, handling the Israeli Interest Rate
story, by checking with Alex concerning who should receive copies of the
corrected version.
What began as a joke, therefore, turns out to have some serious import for
news production . The Israeli Interest Rate story gets publicly corrected and
distributed to more than one desk and subsequently to the customers of both
Money and Capital and Equities . It also is featured in and referred to in other
stories that are handled by the two desks on that day. The story achieves its wider
circulation by virtue of Peter' s joke. The joke is delicately designed to establish
these possibilities. It does not demand that others abandon the activities in which
they are engaged or even take up the story. Rather Peter' s joke renders visible the
gist of the story that he is currently editing . It momentarily displays the activity
in which he is engaged. It provides colleagues with news concerning the Israeli
Interset Rates but does not demand a response. The talk is produced as if devoid of
sequential relevance and yet invites others to consider the import of the story with
respect to their own activities and responsibilities.
In gaining some indication that a colleague is interested in hearing more of the
story, Peter transforms the way in which he presents the text to the others. The
joke is abandoned, and the speaker provides an authentic rendition of the text,
(re)presenting the change in interest rates. The speaker therefore differentiates the
informing by virtue of the ways in which he presents the text, although in both
cases, it is as if he is simply reading aloud the story on the screen. In the final
part of the informing, the speaker again alters his standpoint vis-a-vis the text ,
visibly locating an error in the original copy. In rendering his activity visible, the
speaker exploits, through the ways in which he talks through the story,
differential standpoints with respect to the text itself. In this way, he ongoingly
tailors the sequential import and sense of the story for those within the local
milieu.
Animated Texts
71
A postrophic Readings
In the first example, we find the speaker reading aloud and then, in the light of an
almost minimal reaction by a colleague, delivering a detailed exposition of a
story. The exposition contrasts with the humor of the reading; the exaggerated
Jewish accent is abandoned as Peter provides his colleague with details of the
story and in particular the specifics that may be of particular relevance to the
Equities Desk. Peter therefore does not simply make visible, through reading and
explicating the story, the activity in which he is engaged but rather, through the
talk, differentiates elements of the text. The text, at least as far as the recipient is
concerned, is constituted through the talk, and the talk provides the resources
through which colleagues can differentiate the character and status of the story . As
the following fragment suggests, the ways in which the story is characterized,
even when the characterization is itself a quote or rendition from the text, may
transform, not only through successive utterances, but within the developing
course of a single turn at talk. Differentiating and transforming the text in this
way is embedded in and coordinated with the actions of the potential recipient(s)
and the ways in which he ongoingly participates in the text's rendition.
In the following instance, Peter returns to the desk after his lunch break. A
fellow journalist on the Capital and Money Desk begins to tell Peter what he is
working on:
Fragment 2, Transcript
Jan :
((Peter sits down))

I'm looking at this ~eria Crisi2. (.) I don 't
think it ( .) tells us anything we don't know
already .
(4 .2)
Jan:
Peter:
Jan :
There's people saying i(t) . it's unclear why the

Government has opted for new elections.
Well then ill.ke it . It's c1:ill2 .
erm (0.8) I fear the way things are going,
said a cigarette vendor.
Isa:
9h hah .
(.)
Peter:
A cigarette vendor?
Peter :
Jan :
He's one of the er majQ analysts we spoke to?

He's one of our key sources.
Jan :
No I don't think we need this.
(0 .2)
72
Jan announces that he is working on the Nigerian Crisis and raises the possibility
that the story might be dropped. He continues by giving Peter a flavor of the
story, an utterance that serves to project further information. Peter, however, takes
up the possibility raised in the introduction (namely, spiking the story) and,
remaining oriented toward his screen, starts typing right on the completion of his
utterance.
Despite the response, Jan attempts to encourage Peter to listen to further details
concerning the story. He utters "erm," projects continuation, but withholds the
projected turn. Roughly one third of a second into the pause, Peter lifts his hands
from the keyboard and turns from his monitor toward Jan.
Fragment 2, Illustrations 1 and 2
v
v
=errn- - - - - - - -I .fear the way things
(0 .8)
In the light of Peter's commitment to listen to the story, Jan transforms the ways
in which he is voicing the text. Rather than describing what the story says, Jan
quotes a quote from the text. The exposition is hearable as a quote of a quote by
virtue of Jan's visual and vocal conduct. By preserving his visual orientation
toward the monitor, even after Pete has turned toward him, Jan displays that he is
reading the story to his colleague. Moreover, through the character of the text that
is read aloud and, in particular, the use of the first person pronoun, the present
tense, and an expression such as "I fear," Jan displays to the recipient that the
quote is itself a quote in the text. Jan seemingly produces an authentic version of
the text, which is itself a quote, a quote that is presented as a simple, unsoiled,
voicing of the voice quoted in the text itself.
The ascription of the quote to a cigarette vendor delightfully transforms the
utterance into a joke. It not only underscores retrospectively the absurdity of the
quote but also reconfigures the speaker's position vis-a-vis the talk he has
produced; it even renders questionable the authenticity of the quote, which it now
turns out was delivered dead pan. The joke, of course, turns on the idea that a
Animated Texts
73
quote from a cigarette vendor would feature in a Reuter's story . Through the
ascription, the speaker's voice is reconstituted retrospectively and prospectively
reconfigures the nature of the activity and its sequential relevance.
It is no longer relevant, at least in next turn(s), for the co-participant(s) to
address the import of the story with respect to its news worthiness, but rather to
recognize and respond to the joke. Isa giggles and Peter picks up on and escalates
the joke. Rather than deal with the organizational relevancies of the story, Peter
delightfully juxtaposes cigarette vendor with major analyst, with Jan in turn
producing keysources.
Although Jan's remarks may appear more concerned with telling a funny story
than dealing with the news, like Fragment 1, the exchange has some relevance for
the work in which the participants are engaged. It not only informs Peter what
story Jan is working on, but also allows Peter to know what story he should take
from the basket next. Jan's remarks also provide Peter with a characterization of a
story, a characterization that suggests that the story might be worth spiking, and
invites Peter's response. Although Jan may be more concerned than Peter as to
whether the story is worth salvaging, he establishes Peter's support in spiking the
story if he so wishes. By the completion of the episode, not only is the story
spiked, but also all those working on the desk know that it has been spiked and
why. In one sense, therefore, no matter how jocular or trivial Jan's remarks might
seem, they provide the foundation for a collaborative decision not to transmit a
news story concerning the ongoing crisis in Nigeria. This might seem
unimportant to us now, but for those in financial services and dealing rooms in
the City of London and elsewhere at that moment, the decision to spike the story
may not necessarily have been inconsequential. It is not surprising, therefore, that
you might seek the thoughts of your colleagues, however indirectly, before taking
such a decision.
The articulation of the story is accomplished progressively , in the light of both
the speaker's ability to establish particular forms of co-participation and the
recipient's willingness to cooperate as an interested listener. An interesting feature
of the story's articulation is the speaker's shifting alignment to the text displayed
within the talk. The different standpoints that the speaker adopts in relation to the
story and particular components within that story, including speakers and their
utterances, are contingent on and accomplished through the (co-)participation of a
colleague(s). Moreover, in the light of particular forms of co-participation, the
speaker not only can articulate particular voices and shift these voices
retrospectively but also, in so doing, can transform, within the articulation of a
single utterance, the activity in which he is engaged. In consequence, the
sequential implicativeness of a particular action and the trajectory that emerges
therein transforms as the speaker shifts the way he or she voices the text. The
active listener so richly described by Bakhtin (1976) not only responds to and
develops the utterances of his interlocutor but also, through the ways in which he
behaves both prior to and during the course of the talk, permeates the very things
that are said and done.
74
One element of this interactional organization is the way in which the ongoing
behavior of the co-participant informs, and is informed by, the varying
standpoints the speaker adopts with respect to the text in which the talk is
embedded.
Collaborative Viewings
In some cases, journalists encourage colleagues not only to listen to a story on
which they are working but also to actually look at the text with them. These
collaborative viewings seem to be a recurrent feature of working together on the
news desks, but they may be less concerned with reading than with establishing
an appropriate orientation for an extensive exposition . Consider the following
instance, drawn from the Equities Desk, in which Paul tells Nick about the
layoffs announced by IBM a few minutes earlier. ("*hhhh" represents an inbreath;
the brackets indicate that the utterances or parts of the utterances are produced in
overlap.)
Paul :
Nick :
*hhhh Well that certainly is definition of

downsizing.
(0 .7)
What thirty five thousand?
( .)
Paul :
Nick :
Yeap
(0.3)
How many. Out of . how
I Fifty
Nick:
Viv :
thousand in a (v) : ( .) left or

committed to leaving in ninety three:: an through
actions taken today : : : ( .J another thirty five
thousand ( .) are leaving I . B.M. :
Bloody Hell
What is that in terms of
.
Paul :
It
Paul :
Viv:
used to be: three hundred and fifty thousand (.)

it may:
It may not be such an incident
Animated Texts
75
Paul is looking at his editorial monitor. A moment or so before beginning the

preface, he turns toward Nick and then Nick's focal domain (his colleague's
editorial monitor). As he turns, Paul produces the loud inbreath and (re)turns his
gaze toward his own editorial monitor. As Nick turns, he begins to speak. By the
time his gaze arrives at his own monitor, Nick turns toward Paul.
The preface not only serves to establish a recipient for the informing (What
thirty five thousand?) but, in the way that it is built, also encourages the coparticipant to align toward the teller. Moreover, Nick's realignment of gaze toward
Peter is itself perhaps designed to provide the relevant response, an appropriate
alignment for Peter to deliver the story. After receiving no immediate response
from Paul, Nick produces the query that elicits confirmation and then,
momentarily glancing toward Paul's screen, attempts to elicit further information .
So Paul establishes a recipient who is not only prepared to listen to the details
of IBM's decline but is also visually oriented toward the speaker. Paul exploits
the recipient's commitment and delivers an extensive quote from the story. Even
so, as the quote emerges, the perturbation "a(v): (.) left" successfully encourages
the recipient to turn from the speaker and look at the editorial monitor. The
speaker therefore secures the cooperation of a colleague to listen to the story and
to view the text with the speaker as it is read aloud. In the following pictures,
Paul is on the left, and Nick is second from the left.
The recipient's conduct provides a foundation for an exposition of the story and in
particular a lengthy quote from the text itself. It appears to be the case, at least for
a reading that will necessitate a number of phrases and/or clauses, that to have a
recipient orient toward the text with you provides a more suitable environment
than looking at the speaker (and in particular the speaker's face). Indeed, for
relatively lengthy readings, it appears that tellers go to some trouble to have the
recipient orient toward the text rather than the speaker.
The text is in some sense the source and the means of the activity, and
alignment toward the text can in consequence demonstrate participation and
involvement in the activity . It would also seem, however, that an orientation
toward the speaker rather than the text during the delivery of such readings can
prove problematic, a fragile state of affairs, almost in tension . Indeed, looking at
76
the speaker rather than the text can serve to display the temporariness of the
recipient's alignment and commitment to the exposition . Such tensions may be
related to the ways in which gaze toward another can serve to engender action, to
elicit, for instance, the gaze of the person at whom one is looking and to establish
mutual gaze (cf. Kendon, 1991, Goodwin, 1981, Heath, 1986). By looking at the
speaker during the delivery of a quote, you encourage a mutual, face-to-face
alignment, an alignment that itself can serve to bring the quote to a premature
end. In encouraging the recipient to look at the text with you, even though in one
sense it is futile because he is too far away to read the text for himself, you not
only secure commitment to the telling of the story but also avoid placing yourself
under undue pressure to address the remarks visually toward the co-participant. If
the speaker were to return the gaze of the recipient, it might well serve to
undermine the possibility of delivering a quote from the text. Mutual alignment
toward the text provides an appropriate interactional environment for one
participant to read aloud to another and thereby tell the story.
Toward the completion of the quote, the recipient turns back to his own
monitor. In response, he simply mumbles, "*Bloody Hell," and it is left to Viv
(the editor on the desk) to elicit further information about the story. Indeed, she
attempts to elicit just the information that Nick was asking for earlier: namely,
what proportion of ruM's work force is being sacked. As the discussion
continues, Viv later announces that they will write a short feature on the ruM
story .
So although the original telling may be making no more than a simple moral
point concerning the behavior of ruM and the stock market, the voicing of the
text brings the story to the attention of others on the desk and leads to a feature
that may not otherwise have been written. The moment of small talk turns out to
have some relevance for the organization of the work in which the participants
are engaged and in particular in collaboratively deciding what to do with a
particular story. Many of these tellings have this sort of character. In various
ways, they render some part of the participants' current activities visible to others
within the domain and in particular provide others with a sense of the stories on
which they are working. These tellings may simply lead to a collaborative
decision to spike a particular story, but they may also generate discussion
concerning the relevance of the story for customers who might not otherwise
receive the news. So although these tellings may in some way appear trivial and
irrelevant to the various responsibilities that journalists have within the news
room, they can have a profound impact on who receives what news and the
consequent behavior of the markets.
Animated Texts
77
Forestalling a Reading
Occasionally speakers fail to establish an appropriate orientation from a colleague
for the exposition of a story. Consider the following example, in which Barry
talks about a story he is working on, a severe typhoon that is gathering in
Argentina.
Barry :
I 've got a bit IDQ[e: on the : : er:rn,
Barry :
This
Barry :
Celia :
Carol :
the Typhoon, it 's going to (con) su : :per typhoon ,

Will it ?
Barry:
Urn yeh.
Barry :
a hundred and forty knot winds
(1.0)
( .)
Oh
~.
(.)
(9 .0)
((26 .00) )
Barry and Celia are alongside each other on the Commodities Desk. While Celia
is typing , Barry begins to talk. His initial utterance, "I' ve got a little more on the
typhoon," suggests more is to follow and is designed to secure some form of
commitment from Celia to listen to the story (cf. Sacks , 1992) . Although Barry
is unable to see exactly what Celia is working on, he positions his initiating
utterance to maximize the possibility of gaining some commitment. It is
juxtaposed with Celia momentarily removing her hands from the keyboard .
Unfortunately, as Barry begins to speak, Celia begins once again to type.
Barry does not abandon his attempt to tell Celia the news . Rather, he attempts
to encourage Celia to temporarily suspend the activity in which she is engaged
and orient toward the story on which he is working . He leans sideways toward
Celia while simultaneously exaggerating his orientation toward his own monitor.
In response, Celia reorients posturally towards Barry's monitor but steadfastly
continues with the activity in which she is engaged. Celia's divided bodily
orientation attempts to reconcile Barry's eagerness to tell the story with her own
interest in continuing the activity in which she is engaged.
Despite Celia's attempt to divide her orientation and encourage the speaker to
tell the story while she continues to work on some other text, Barry attempts to
secure more commitment. Prior to completing the preface to the story, Barry
stretches the word "the::," produces "er.m," and pauses.
78
Fragment 4, Transcript 2, Illustrations 1 & 2
B. reorientes
v
--I've 9Qt a bit more : on the : : er :m,----- - --this-the
C. ceases
and again
starts typing
C. reorientes
Neither the sound stretch nor the hesitation serves to encourage Celia to provide a
more wholehearted commitment to the story . She continues to type and remains
oriented toward her own monitor and text. The speaker pauses, withholding the
projected item and upcoming story. Roughly half a second into the pause, Barry
momentarily glances at Celia's screen. The glance may be more concerned with
encouraging her to orient toward him rather than with assessing the state of her
current activity. As Barry returns his gaze to his own monitor, Celia produces a
series of staggered taps on the return key that appear to project an upcoming
boundary in her current activity.
Barry exploits the opportunity and begins to speak precisely on the anticipated
break in Celia 's activity .
Rather than continue with the projected news, however, Barry once again
attempts to encourage Celia to upgrade her commitment to the story. He
temporarily delays telling the tale, thrusts his head toward his monitor, and utters
"this." The demonstrative pronoun, coupled with his visual actions, appears to be
designed to show simultaneously (the source of the) upcoming story, while
encouraging Celia to look at the story and thereby to temporarily suspend the
activity in which she is engaged.
If his actions are an attempt to draw Celia's attention to the text and telling and
to elicit some greater commitment from Celia, they fail. She once more begins to
type, and Barry abandons any further attempt to build the story in the light of a
mutual orientation toward the text.
He returns to the prior activity and completes the preceding utterance, replacing
"this" with "the typhoon ." He does not follow the initiating utterance with (the
Animated Texts
79
projected) details concerning the severity of the typhoon. Rather he produces a

description of the typhoon that attempts to upgrade the excitement and appeal of
the story with which he is dealing: "it's going to (con) su: :perA typhoon." The
actual movement into telling the story and in particular perhaps a quote from the
actual text are forestalled in the light of the speaker's failure to establish an
appropriate alignment from the co-participant. Rather than providing details of the
typhoon at that moment, Barry produces a description of the event that potentially
foreshadows more to follow if the recipient(s) were to display more commitment.
In response, it seems that neither Celia nor Carol, who is sitting opposite,
provides Barry with the encouragement he needs to deliver a full-blown exposition
of the story. They both continue to work on their own stories as they utter their
responses . Sadly, despite Barry's successive attempts to build a suitable
environment for the delivery of the story, he largely abandons the activity and
sharing his excitement concerning the typhoon that is brewing in Argentina.
Some moments later, as he continues to work on the story, he does quietly utter ,
almost under his breadth, "a hundred and forty knot winds." Although this may
well be the quote for which he was attempting to establish a suitable audience, by
the time it is delivered, it fails to secure even an acknowledgment.
As in the previous instances, we see the way in which a speaker attempts to
coordinate the production of a story with the conduct of the (potential) recipient.
Tellers differentiate components of the story with respect to the co-participation of
the recipient and systematically attempt to secure particular forms of alignment
for specific elements of the story. In the case at hand, although the beginning of
the story preface is coordinated with a potential break in the recipient's activity,
the speaker does not, at that moment, demand an immediate commitment to the
telling and the text itself. However, even prior to the completion of the preface, it
is apparent that the successful articulation of the story is contingent on the
recipient providing heightened participation in the telling , most likely a
temporary suspension of the activity in which she is engaged. The speaker's
successive failure to establish the relevant (sought for) co-participation at specific
junctures within the developing course of the activity has him transforming and
then largely abandoning the telling.
Within this step-by-step negotiation, it is interesting to note that the teller
systematically attempts to coordinate specific elements of his talk with his
colleague's activity, even though Celia's activity is only partially visible or
available to Barry. Indeed, while Barry attempts to selectively deliver elements of
the text on which he is working, he coordinates the delivery of the story with the
inferred stages or boundaries of the activity in which Celia is engaged. So, for
example, Barry positions the onset of the initiating utterance with respect to a
upcoming boundary within Carol's activity, a boundary that is projected by virtue
of a sequence of key strokes. And, a little later, he produces the key word, the
topicalizer, "the typhoon," with respect to a slight reorientation by Celia on the
keyboard and the screen. So although Celia's activities are largely invisible,
embedded in the text on which she is working, he attempts to systematically
80
coordinate the production of the telling with whatever he can retrieve from the
visible elements of his colleague's conduct.
Although Barry fails to establish an appropriate audience for an exposition of
the story, the item he mentions does not pass unnoticed. A minute or so later,
Carol seeks further information about the item from Barry, clarifying how it fits
with some weather news within another story that has entered the basket. Some
fifteen minutes later, as Barry is leaving the desk to go home, Carol asks where
he put the typhoon story and to whom it was sent. So although these tellings in
themselves may not always have much immediate sequential significance, they
can provide resources for the production of news sometime later. In particular,
they not only display who is working with what story, but also provide a sense of
the gist of the story, which can turn out to be relevant for other sorts of items
with which colleagues are having to deal. Momentary exchanges or tellings, or
simply catching someone reading aloud parts of a story on which he is working,
can be retrospectively found to have some import for an activity in which you
are engaged some time later.
Discussion: Texts in Interaction

Personnel in the news room continually inform each other of the stories on which
they are working and collaboratively make decisions on how to handle particular
items and on who might have an interest in particular pieces of news . Relatively
few of these informings involve the immediate off-loading of a particular item, a
blunt and unprepared delivery of a story or a quote. Indeed, if this were the case, it
would be difficult, if not impossible, for journalists to fulfill their various
responsibilities. They would be continually bombarded with and subjected to the
demands of colleagues as they are informed of news without regard to their
immediate concerns and activities. Rather we find that journalists are sensitive to
the current commitments of their colleagues and attempt delicately and
interaction ally to build a suitable environment for the delivery of news in close
cooperation with colleagues . Even the actual tellings are undemanding. They have
a lightness of tone. They are presented as entertainment and gossip, as jokes and
quips. Journalists respect the integrity of activities in which others may be
engaged, their current concerns and involvements, while delicately seeking
particular forms of cooperation for the delivery and presentation of some
potentially relevant piece of news .
The delivery or presentation of some news involves not only securing some
commitment to listen to a story but also differentiating particular forms of (co-)
participation within the accomplishment of the informing . We have seen the ways
in which tellers progressively establish the particular form of co-participation they
require within and across utterances. So, for example, a teller might seek to
Animated Texts
81
establish recipient alignment toward the text for the delivery of a quote, whereas a
precis of the story might be delivered while the co-participant is looking at his
own monitor and editing a separate story . Moreover, the data suggest that the
teller, in failing to secure relevant co-participation for the accomplishment of a
particular type of action such as a reading, may transform the projected activity in
which he or she is engaged and deliver the news in a very different fashion than
was suggested in the preceding talk. Journalists not only render particular
activities in which they are engaged visible to others within the local milieu, but
also develop selective characterizations and portrayals of stories and differentiate
the various forms of co-participation they require for different actions within those
renditions. Quotes, precis, readings, and summaries are differentiated in the talk
itself (and the characterization that is being developed) and systematically
accomplished with respect to different forms of co-participation from the
recipient(s). The accomplishment of the tellings, the step-by-step production of (a
characterization of) a story, are produced with respect to the current conduct of the
co-participant and in particular his or her orientation to different components of
the characterization during its articulation . Mutatis mutandis, the characterization
itself and the ways in which the textual story is rendered visible, is contingent on
the co-participation of the recipient during the course of its production .
The relationship between the informings and the text is both curious and
complex. In the production of an informing, tellers differentiate the status of
different components of the characterization with respect to the original text. The
text itself, the existence of an authorized and written account, is exploited in the
telling for a variety of purposes. It can be used to produce and present the factual
version of some set of events to enable others to build or transmit stories that
will have a significant impact on the behavior of particular markets. It can be used
to make political comments or to ridicule the journalism of colleagues based in a
distant office. Within each fragment, we find the teller systematically displaying
(and exploiting) the relationship between the informing and the text and
demarcating his own standpoint and "voice" from both the original author and
even sources within the text itself. So, for example, we can observe the ways in
which the teller can display that he is rewriting that story within the course of its
telling and retrospectively recast the authority of the text. Or, for example, we can
see how tellers prospectively establish a quote of a quote and display their own
alignment toward the relevance of the story for news production . The text is
selectively rendered visible. It is revealed within the talk and through the ways in
which the teller animates or embodies the text. The developing rendition , the
ways in which the talk embodies the text, is dependent on and embedded in the
emergent interaction with others within the local milieu and, in particular , the
teller's ability to establish and sustain particular forms of co-participation during
the production of a telling.
The text, therefore, or at least the text displayed on the monitor, does not so
much mediate the interaction but rather is ongoingly constituted in and through
the interaction . The characterization that is developed, the animated versions of
82
particular elements, and the quotes and the summaries do not simply transmit or
present the text but rather constitute the text, the story, for the other(s). The coparticipant's access to the story, even when he looks at the screen with the
speaker, is constrained by both what he can see and his understanding of the
contents of the piece. His colleagues' talk constitutes the text and the ways it is
seen on this occasion, its intelligibility, and its sense here and now. In
constituting the text, personnel provide a sense of their current work and the
activities in which they are engaged . Access to the story is asymmetrical. It is not
simply that the co-participants may not necessarily be able to see the same story;
rather, what they know of the text, even during the course of a collaborative
reading, is permeated through and embedded in the ways in which the teller is
concurrently characterizing the text. The reading is interaction ally constituted, not
only in the traditional sense through the author's relationship with the reader (Iser,
1982), but also through the interaction between teller and recipient. The talk
elaborates the text, and text elaborates the talk within the developing production
of the interaction .
The observations discussed here provide further support of the critique of the
conduit metaphor of communication, a critique developed by Bakhtin (1976) and
others such as Wertsch (1991). Moreover, the observations underscore the dialogic
nature of talk, even where the talk itself is reporting reported speech. And, we
might, following Bakhtin (1976), consider the ways in which talk reproduces and
relies on a particular textual genre (news reporting) that may be theoretically
distinct from the current context and yet forms an integral part of retelling and
editing stories within the newsroom. But the character of the dialogicity and
textual rendition discussed here goes beyond the idea of genres characterized so
profoundly by Bakhtin and developed in diverse ways by Todorov (1990), Lodge
(1990), and Goffman (1974) in his frame analysis of talk. In particular, in the
materials at hand, we find that the delivery of a piece of news to a colleague , itself
a selective rendition of a textual report, is produced with respect to the shifting
alignment and participation of the (potential) recipient. How the text is
(re)produced is thoroughly contingent on the interaction as it emerges between the
participan ts. The relationship between the speaker and the author, the report and
the original text, and the distinction between de dicto and de re (cf. Coulmas,
1986) is thoroughly contingent on and embedded in the emergent interaction
between the participants, even during the shifting course of the utterance and
textual reading.
As in other organizational environments , personnel in the newsroom at Reuters
have asymmetrical access to each other's activities. Aspects of colleagues'
activity, even when they are discussing a particular story with a colleague and
appear to be in mutually focused interaction, may be unavailable to the coparticipant. Moreover, in the newsroom as in other working environments,
individuals may simultaneously participate in multiple overlapping and
interdependent activities that demand a complex and continually shifting range of
configurations between those within perceptual range of the event (cf. Goffman,
Animated Texts
83
1981). For, as we have seen, while simultaneously engaged in reading or editing a

story, a journalist may render certain aspects of the activity potentially visible to
others within the local milieu, where the visibility of the activity itself is
embedded in the participants' concurrent interaction and the teller's ability to
encourage the co-participant to align toward the emergent activity. The individual
activities that are being accomplished with and within the interaction are more or
less available to each of the participants and necessitate differing standpoints and
forms of participation at different moments within their accomplishment.
In consequence, during these moments of exchange and discussion , an
individual's actions are coordinated with what is available and made available
within the other's actions and activities . For example, as we have seen, the onset
of an utterance may be coordinated with no more than the anticipated completion
of a series of keystrokes or the sudden movement of a mouse to scroll through
some text. On the other hand, the selection and articulation of a quote may be
delicately coordinated with a recipient's visual orientation toward the text that the
speaker is editing . Either way, the participants ' access to each other's activities,
the text in which at least one of their activities is embedded (and embedding), and
the system through which the original story is being manipulated are differentially
available to the co-participants. These emergent asymmetries both inform and are
constituted through the ways in which an activity emerges and through the
interaction of the participants in the developing course of task production .
The analysis of tool-based action and interaction both in the cognitive and the
social sciences has largely delineated the ways in which technologically mediated
activities involve specialized individual competences that may be systematically
interleaved with the skills possessed by other members of an organization. Even
the burgeoning body of research in computer-supported cooperative work has
preserved a sharp distinction between the individual and the collaborative and has
characterized the production of tasks in terms of successively interrelated,
individual contributions. Although such an analytic orientation may prove useful
for the design of advanced technologies, we can begin to see the shortcomings of
such an approach. In the materials at hand, it is clear that it is difficult to draw a
sharp distinction between individual and collaborative activities or even between
one person 's task and another's. Indeed, in Reuters , as in other working
environments, individuals simultaneously participate in and contribute to
multiple, interrelated activities. An important and unexplicated element of this
interactional accomplishment is the way in which participants orient to and deal
with their "limited" access to each other's activities and the documents in which
they are embedded.
84
Acknowledgments
We would like to express our gratitude to journalists at Reuters who so
generously cooperated with the research. We would very much like to thank
Lauren Resnick, Jon Hindmarsh, Charles Goodwin, and Paul Luff for comments
on an earlier version of this paper. We would also like to thank Bernard Conein ,
David Greatbatch, Suzanne Guenthner, Isaac Joseph, Hubert Knoblauch, and Per
Linell . We would also like to thank Lucy Suchman for her support and
encouragement for the project of which this study forms one part. The study is
supported by the Rank Xerox Research Laboratories Cambridge EuroPARC, the
European Commission RACE Programme Project MITS (Metaphors for
Telecommunication Services), and the ACTS project MEMO .
Animated Texts
85
References
Atkinson, J. M., & Heritage, J. (Eds.) (1984) . The structures of social action.
Bakhtin, M. M. (1976) . Speech genres and other late essays . (V. W. McGee, trans.)
Austin : University of Texas Press.
Brown, J. S., & Duguid, P. (1994) . Borderline issues : Social and material aspect s of
design . Human-Computer Interaction, 9(1), 3-36.
Coulmas, F. (1986) . Direct and indirect speech. New York: Mouton de Gruyter
Drew, P., & Heritage, J. (Eds.) (1992) . Talk at work: Interaction in institutional
settings. Cambridge, England: Cambridge University Press .
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: PrenticeHall.
Goffman, E. (1974) . Frame analysis: an essay on the organisation of experience. New
York: Harper Row.
Goffman, E. (1981) . Forms of talk. Blackwell: Oxford .
Goodwin , c., (1981) . Conversational organisation: Interaction between a speaker and
hearer. London : Academic Press.
Goodwin , C., & Goodwin, M. H. (1996) . Formulating planes: Seeing as a situated
activity . In Y. Engestrom & D. Middleton (Eds.), Cognition and communication at
work: Distributed cognition in the workplace. Cambridge, England : Cambridge
University Press
Greatbatch, D., Luff, P. K., Heath, C. C., & Campion, P. (1993) . On the use of paper
and screen in the general practice consultation . Interacting with Computers, 20,
216-234 .
Harper, R., & Hughes, J. (1993) . What a f-ing system! Send 'em all to the same place
and then expect us to stop 'em hitting: Making technology work in air traffic
control. In G. Button (Ed.), Technology and the working order (pp. 127-144) .
London : Routledge.
Heath, C. C. (1986) . Body movement and speech in medical interactions. Cambridge,
England: Cambridge University Press. (Editions de la Maison des Science de
l'Homme, Paris).
Heath, C. C., Jirotka, M., Luff, P., & Hindmarsh, J. (1995) . Unpacking collaboration :
Interactional organisation in a city trading room . Journal of Computer Supported
Cooperative Work, 3(1), 147-175.
Heath, C. C., & Luff. P. (1992) . Collaboration and control: Crisis management and
multimedia technology in London underground line control rooms . Journal of
Computer Supported Cooperative Work, 1(1-2), 69-94.
Hutchins, E. L. (1990) . The technology of team navigation. In J. Galegher, R. E.
Kraut, & C. Egido (Eds.), Intellectual teamwork: The social and technological
foundations of cooperative work. Hillsdale , NJ: Lawrence Erlbaum Associates.
Iser, W. (1985) . The act of reading. London : Routledge.
Kendon, A. (1991) . Conducting interaction: Patterns of behaviour in focussed
encounters. Cambridge, England : Cambridge University Press.
Lodge, D. (1990) . After Bahktin. London : Routledge.
Markova, I., & Foppa, K. (1990) . The dynamics of dialogue. London : Harvester
Wheatsheaf.
86
Sacks, H. (1992). Lectures in conversation:Volumes I and II. Oxford: Blackwell.

Suchman, L. (1987) . Plans and situated actions: The problem of human machin e
interaction. Cambridge, England : Cambridge University Press.
Suchm an, L. (1993). Technologies of accountability: On lizards and aeroplanes. In G .
Button (Ed.), Technology in working order (pp. 113-126). London: Routled ge.
Todor ov, T. (1990 ). Genres in discourse. (C. Porter, trans.) Cambridge, England :
Cambrid ge University Press.
Vygotsky, L. S. (1978). Mind in society: The development of higher ps ych ol ogical
process es. Boston, MA: Harvard University Press .
Wertsch , J. V. (199 1). Voices of the mind: A sociocultural approach to mediated
action. London : Harve ster Wheatsheaf.
Whal en, J. (1995). Technology and the coordination of human activity: Computeraided dispatch in public safety communications. In P. ten Have & G. Psathas (Eds.),
Situated order: Studies in the social organisation of talk and embodied activities .
Washington, DC: Univer sity Press of America .
Winograd, T., & Flores, F. (1986) . Understanding computers and cog nition: A new
f oundationfor design. Norwood, NJ: Addison-Wesley.
Chapter 4
To Resolve a Technical Problem Through

Conversation
Alain Trognon and Corinne Grusenmeyer
Research Group on Communications I Nancy II University
and French National Research and Safety Institute I Nancy-Vandoeuvre, France
Abstract
In everyday life, as in work situations, problem solving is a situated,

accomplished process . However, few studies aim at describing this process, taking
its conversational developments and its logical accomplishment into account.
This chapter proposes an analysis of such a process, using a theory called
interlocutory logic. First, reasons explaining, in our point of view, the relatively
few studies in this area are presented . Second, some elements of interlocutory
logic are introduced . Third, an application of this theory is shown . It concerns a
conversation of two experienced operators during shift changeover in a paper
production company . The analysis, in taking into account the conversational
process progressively, shows that these verbal exchanges permit a logical
resolution of the problem and that rationality emerges from conversation. These
results highlight that in situ accomplishment of rationality is a global interactive
process, including semantic, conversational, and logical aspects .
In everyday life, the normal resolution of a problem is a process, a situated

accomplishment, and an activity oriented toward and coming from a course of
action, rather than an intellectualized pursuit. This claim is not original. It is
found in numerous works carried out within the framework of ethnomethodology
and conversation analysis. It is also applicable to interaction in work situations
88
(cf., for example, the works of Hastie & Penningt on, 1991; Hughes, 1990;
Resnick, 1991; Suchman, 1987).
However, how should this process be described with respect to its
conversational developments, as well as in terms of its logical accomplishment?
There have been relatively few studies carried out in this area, even though, in
principle, there is no reason to consider that the fundamental principle of
ethnomethodology, as stated by Garfinkel, should not be applicable to this sphere.
Every topic of logic, order, reason, meaning, and method is to be discovered and is
discoverable, and is respecified and respecifiable only as locally produced,
naturally accountable phenomena of order. These phen omena are immortal,
ordinary society's commonplace , vulgar, familiar , unavoid able, irremediable and
uninteresting "work of the streets." (1990, p. 77)
The Works Carried Out on Conversational Intelligence

Why are studies of the problem of conversational intelligence less numerous?
Two reasons emerge as likely explanations of this observation. The first reason
can be found in the epistemology of ethnomethodology and the second in the
epistemology of work psychology.
The Epistemology of Ethnomethodology

Ethnomethodology amalgamates three theses. The first favors the inductive
approach. The second stipulates that the categories inductively highlighted must
be those recognized within the process by the conversationalists during the course
of their interaction:
The movement arose in reac tion . . . and the arbitrary imposition on the data of
supposedly objective categories. . . . In contrast, it was argued cogently, the
proper object of sociological study is the set of techniques that the members of a
society themselves utilise to interpret and act within their own social world .. . .
Out of this background comes a healthy suspicion of premature theorising and 00hoc analyti cal categories: As far as possible, the categories of analysis should be
those that participants themselves can be shown to utilise in making sense of
interaction ; unmotiv ated theoretic al constructs and unsubstantiated intuitions are
all to be avoided. In practice this results in a strict and parsimonious structuralis m
and a theoretical asce ticism . The emphasis is on the data and the p att ern s
recurrentl y displayed therein. (Levinson, 1983, p. 295)
The preceding thesis reflects the concern for theoretical prudence and the
prevalence given to the interpretations effectively made by the speakers. But
To Resolve a Technical Problem Through Conversation
89
sometimes ethnomethodologists go beyond these theses and subcribe to a third

thesis that prohibits introduction of analysis categories external to the situation
under study.
In our point of view, therefore, the conjunction of these three theses primarily
explains the few works devoted to studying cognitive accomplishments at a
logical level. On the whole, we subscribe to ethnomethodology, which, in effect,
saves us from a form of sociological or psychological reification, in keeping with
the arguments put forward by Garfinkel in his critique of mainstream academic
sociology. However, we reject the third thesis because, in principle at least, it is
possible that the accomplishments form the identified logical or grammatical
properties expressed and even formalized elsewhere : namely, in other spheres of
research and clearly outside the present interaction occurring between the speakers .
We even consider that conversation is a device invented by the human species to
make do with the logical impasse of human communication abstractly identified
within the framework of formal pragmatism (Brassac & Trognon, 1993; Trognon,
1991, 1992; Trognon & Brassac, 1992), and that the logical properties of
conversational sequencing of illocutionary acts, at the heart of this device, must
be taken into consideration if an empirically adequate description in harmony with
the rules of ethnomethodological description is wanted (Trognon, 1994; Trognon
& Brassac, 1992).
The Epistemology of Work Psychology

The second main reason for the lack of studies in this area lies in the situated
action/planned action debate, which somewhat opposes ethnomethodology and
work psychology, if the numerous works of this latter discipline on planning are
to be believed (cf. notably the works by Hoc, 1987) . The real or imaginary
existence of the specific and antagonistic objectives of both disciplines, thus ,
often leads work psychologists primarily to carry out an analysis of the content of
exchanges by means of categorizations, thereby , reducing the fundamental
character of the on-going interac tional process of exchanges. A recent tendency
aimed at taking into account the language process has nevertheless emerged over
the last few years (Karsenty & Falzon, 1993; Lacoste, 1993; Lacoste & Rogard,
1988) .
The Concept of Conversational Intelligence

It is clear that rationality is, above all, accomplished originally and naturally in
conversation (Levinson, 1983). Some of the greatest developmental psychologists
insisted on the preceding point of view a very long time ago . For example, in
1965 Piaget wrote that:
90
Cooperation simultaneously opposes both spontaneity and constraint. It

progressively reduces the autistic or egocentric thought processes. . .. Discussion
brings about internal reflection. Mutual control provokes the need for proof and
objectivity. Thought exchange supposes the principles of contradiction and
identity designed as discourse regulators, etc. As regards constraint, cooperation
ruins it in so far as the differentiation between the individuals and open discussion
develop . (p. 92)
Nevertheless, it is often maintained that we would naturally be inclined to confirm

our initial hypotheses by citing examples that illustrate them rather than by
proving these hypotheses by searching for counter examples that falsify them
(Wason & Johnson-Laird, 1972). Nisbett and Ross (1980) even argue that we
keep reinforcing our beliefs, regardless of the arguments to which they are
exposed, and whether the arguments support or refute our beliefs .
However, we are not so dogmatic, because we are able to assess the veridicality
of our reasoning (Evans, 1984, 1989; Johnson-Laird & Byrne, 1991) and, in our
discussions, we try to bring out the falsity of statements that contradict our
beliefs . Whatever it may be, the conversation produces, by necessity, a positional
differentiation that leads to a potential sphere of contradictions. This probably
explains why conversation is a natural matrix of rationality .
It is still necessary to define clearly the concept of rationality. Practical
rationality, the natural logic that governs our ordinary dealings with people, is
often opposed to abstract rationality, such as that expressed in logical systems.
Prudence would advise that we focus on the former. However, it is much more
decisive, both in the theory of conversation and in the methods brought to bear on
conversation (e.g., the expert systems method), and, consequently, in the theory
of distributed cognition as well, to demonstrate that conversation can be carried
out under certain conditions of formal rationality. It is, therefore, to this
interpretation of the notion of rationality that we refer in this chapter.
Thus, if formal rationality exists in conversation, it must occur according to the
principles that govern conversational activity. Take, for example, a conversation
that serves to resolve either an arithmetical problem or, as described shortly, a
problem concerning the operation of a machine. First, these problems will be
elaborated simultaneously with the problems of conversational organization that
their treatment will undoubtly raise. Second, the problem solving will be
elaborated and treated as conversational problems are themselves elaborated and
treated: step by step, sequentially, in collaboration and, rather tacitly, through
successive accomplishments rather than in a declarative form (Ghiglione &
Trognon, 1993; Heritage, 1990; Larrue & Trognon, 1993; Trognon & Retornaz,
1989), and when the problem arises.
Clearly speakers do not always try to integrate or resolve their contradictions as
they arise during conversational treatment of a problem. They, nevertheless, do it
much more if they believe it is objectively in their own interest to collaborate or
if they are organizationally obliged to do so. In other words, they try to integrate
91
and resolve when they have a common goal, whether this goal comes from
themselves or is externally imposed on them, as in the sequence examined here.
The Objectives of the Chapter
The three theses mentioned earlier - that rationality is accomplished and exercised
conversationally, that this emerging rationality is no different in nature from that
explaining logic systems, and that it is organically linked to the processes that
govern conversational organization - merit an epistemological, theoretical, and
methodological analysis beyond the scope of this article. This is why we tackle
them here in a practical manner, by analyzing a conversation.
We begin with an example of an interactive process of problem resolution .
After quickly describing the context, we provide the conversation that permits
resolution of the problem. Next, we attempt to show how this conversational
process progressively brings about the logical solution to the problem, which
will amount to highlighting the idea that rationality as rationality itself emerges
from the conversation .
The Interaction of Problem Resolution

Description of the Problem
The transcript presented in the next section is part of a conversation recorded

during a shift changeover in a paper production company (Grusenmeyer, 1995a).
Information about the nature of the encounter is important because this work
phase, which aims to avoid any interruption in production due to the succession
of teams, supposes operators' implicit or explicit cooperation through certain
work activities , particularly by verbal exchanges (regarding this subject, see
Davillerd & Grusenmeyer, 1993; Grusenmeyer, 1991, 1995a). The objective of
this work phase is essential if the operators setting up a shared functional
representation of the situation are to be able to follow up on the actions initiated.
(Grusenmeyer, 1992, 1995a, 1995b).
This study involves a natural work situation that was not set up for the
objectives of the study. The interaction involves two experienced operators, one of
whom was about to take over the work station and one of whom was about to
leave the work station.
The manufacturing process concerned took place mainly at the paper machine
(see Figure 1). It concerns a semicontinuous, partially automated process, with a
response delay that is relatively long. It is composed of different successive
operations, the main one being the manufacturing of sheets of paper once the
92
paper pulp has been prepared and colored. This manufacturing work covers the
different operations carried out continuously on the sheet of paper: The formation
of the damp sheet on the fabrication table and its progressive drying out in the
damp part of the machine; drying out the sheet; and rolling the sheet of paper.
The situation was as follows . The work on the paper currently being
manufactured at this change of shift had begun four days earlier. This implies that
each of the two operators had already been involved in this manufacturing process .
The processing of the paper had already caused a number of difficulties, and the
process had therefore been subjected to adjustments by the operators since its start
up. Thus, a part of the manufactured product had to be discarded on two occasions:
the previous day, because of pollution coming from the machine circuits, and
during the night shift, because of pollution from the recycled paper pulp. The
latter pollution incident had resulted in the paper pulp circuit being stopped,
cleaning of the corresponding mixing vats, and then progressive restarting of the
recycled paper pulp. Several paper breaks had occurred at the first press since the
start up of this manufacturing.
Damp sheet making on the

manufacturing table
Progressive wringing of the sheet
~""
INPUT:
coloring
pulp composition
pulp concentration
type of pulp
characteristics of pulp
refined
recycled pulp rate
products added
continuously
CONTROllED PARAMETERS:
pulp flow
speed of jet of pulp
speed of manufacturing linen
speed of press
pressure of press
pulp level at the machine head
pulp gate opening
wringing
format of the sheet
recycled pulp ...
OU1PUT:
width of the sheet
thickness
weight of the sheet
tint of the sheet
porosity
hot pressing
ashes rate
acidity/basicity rate
(pH)
paper fibers
arranging ...
Fig. 4.1. Damp part of the machine.
The operators of the machine, who constantly check the dampening component,
which is the most delicate section, diagnosed these malfunctions in the following
93
way. The paper currently undergoing manufacture was of relatively low gram
weight (70 gr.). Consequently, projections of paper pulp had settled at the edge
sprays, otherwise known as pissettes by the operators. The edge sprays are two
jets of water that pass across the sheet in the dampening part of the machine , thus
allowing its width to be demarcated so that it can be cut on either side. The
projections of paper pulp were caused by the low gram weight and a jet of water
that was not completely steady. The pulp had then accumulated at edge spray
level, thereby forming a pile that had come off and then fallen on the sheet being
formed: hence the appearance of paper breaks at the first press. Furthermore, two
paper breaks had occurred at the stack during the morning (this resulted from
another malfunction that is not related to the one discussed here).
The Extract
The following transcript concerns the malfunction caused by the paper pulp
projections. The purpose of the conversation is to inform the on-coming operator
about this. The information concentrates on the setting of the machine: in other
words, the operating modes. The mutual recuperation (and the cognitions) of the
operators is aimed at adapting the operation of the machine to the constraints
imposed on it by the product currently being processed. A is the out-going
operator (leaving the work station), and B is the on-coming operator (the
replacement).
[(...) indicates that some of the text has not been reproduced; ... indicates a
pause]
Al :
B1:
A2:
B2:
A3 :
B3 :
A4:
B4:
AS:
( . . . ) And the pissettes, they seem to be working better .

And that one behind, it still lifts the sheet a bit. Have
you noticed?
Unun, yes , maybe. But I didn't get as much pulp as
yesterday , I didn 't have any pulp after, huh.
I got s ome.
You got some? Me, I didn't get it, huh .
And I closed it off a bit because I found it was
deflecting the sheet a bit and that made it, umm, squirt
out.
oh, right, me what I, I, reopened it, the front one, this
morning a tiny bit, because, you know, you saw today I
broke it (the out-going operator shows the notebook for
morning, right hand side), OK, I pulled the tail end,
umm . .. three times
Yes .
But the band . . . it wasn't cut ... I haven't cleaned the
pissettes umm, I haven't even lifted the pulp by hand . ..
nothing and there's none on them, there are just a few
fibers, that's all OK.
94
B5:
A6:

That's because they 're well set.
That's as good , isn't i t ?
The Interlocutory Resolution of the Problem

Elements of Interlocutory Logic
We use a theory that we have termed interlocutory logic (Ghiglione & Trognon,
1993; Trognon , 1993; Trognon & Brassac, 1992; Trognon and Larrue, 1994) to
help understand the process by which the interlocutors of this conversation
construct a theory of relative recuperation of this machine malfunction and
consequently its condition.
This interlocutory logic is an interpretation (as in the theory of models) of the
general semantic of Vanderveken (1990), which combines an illocutionary logic
(Searle, 1985; Searle & Vanderveken, 1985) and a formal logic. Illocutionary
logic is a logic of the relationship between illocutionary components of the
illocutionary forces. It is a theory of non-defective illocutions: An illocution
succeeds without failure if and only if all the conditions of success and of
nonfailure are fulfilled. Formal logic involves the relationships between
propositional contents of illocutions. Thus, the general semantic is a theory of
success without failures (because it integrates illocutionary logic) and of the
fulfillment of illocutionary acts. An illocution is fulfilled if and only if its
propositional content is true according to its direction of fit. There are four
directions of fit: Word-to-world (for assertions), world-to-words (for directives and
commissives), words-to-world and world-to-words (double direction of fit; for
declaratives), and the empty direction of fit (for expressives). Thus, for example,
an order is fulfilled if it has been obeyed, if its propositional content becomes true
in the world, because it has been given: in other words, if the world becomes as
the words and because of them.
More precisely, we would require conversational interpretation of the following
two theorems (A and B) of the general semantic.
Fulfillment of an IIIocution Entails Success Without Failure

(Theorem A)
In illocutionary logic, this theorem means that conversational value of an
illocution is established through the accomplishment of the illocution that
follows it. Thus , accomplishment (in the sense of conversational analysis) is
always the fulfillment (in the sense of the general semantic) of a preceeding
illocution. Through this, we again find the theses that ethnomethodology has
95
advanced in conversational sequentiality and its role in determining the actions

being accomplished in conversation (Trognon, 1991, 1994; Trognon & Brassac,
1992).
Yet, for reasons too complex to explain in detail here (involving the theory of
illocutionary engagement), the conversational interpretation of the previous
theorem makes a rule of it by default (Trognon, 1993; Trognon & Brassac, 1992)
in such a way that the conversational interpretation (by default) of an illocution
does not suppose only two illocutions (initial illocution and the conversational
accomplishment that follows it), but at least three. When everything goes well,
this sequence of three illocutions allows the speakers to face the problem of
mutual knowledge, that is, to continue the exchange (Brassac & Trognon, 1993).
This type of three-way exchange is thus indicative of a pooling between the two
interlocutors.
The inter-understanding of the Illocutions, both directive and assertive, seems to
adhere to the process just described (Trognon, 1993; Trognon & Brassac, 1992).
In our opinion (Ghiglione & Trognon, 1993), the restoring exchanges (as per
Goffrnan, 1969) and the interactionally complete exchanges (Moeschler, 1985;
Roulet et al., 1991) are the conversational translations of this type of process.
The Fact that an IIIocution Is Unsuccessful or Defective Implies

Its Non-Fulfillment (Theorem B)
This rule reciprocates the previous rule. Its antecedent corresponds to a failed or
defective illocution. Consequently, a failed or defective illocution leads to nonfulfillment of this illocution. An illocution is non-defectively successful (Searle
& Vanderveken, 1985; Vanderveken, 1990) if and only if all its conditions of
success (e.g., point, accomplishment mode, and propositional content conditions)
and non-defectiveness (e.g., preparatory conditions and sincerity conditions) have
been satisfied. For example, with a touch of simplification, a speaker succeeds in
issuing a non-defective assertion to the hearer if and only if:
- The point of the speaker's utterance is to say how things are (point);
- The speaker presupposes that he or she has reasons (or grounds or evidence) that
count in favor of or support the truth of the propositional content (preparatory
conditions), and this presupposition in fact obtains.
- The speaker expresses and actually has a belief that (p), the propositional
content (sincerity conditions) .
Consequently, the non-fulfillment of these latter conditions leads to the negation
of the success and the non-defectiveness of the illocution and, thus, its nonsuccess or defectiveness (Ghiglione & Trognon, 1993; Trognon, 1990, 1993;
Trognon & Larrue, 1994, for a more precise presentation and illustrations). The
non-fulfillment of a success and non-defective condition by a second speaker of an
96
illocution given by a first will be translated conversationally (Ghiglione &

Trognon, 1993; Trognon, 1993) by the opening of a lateral sequence (Schegloff,
1992), otherwise called a subordinate exchange (Moeschler, 1985; Roulet et al.,
1991).
Treatment of the Problem

The previous theoretical conditions were essential , because, through the
conversational processes, they describe that the operators were going to rationally
solve the problem with which they were confronted. The problem they faced was
to explain to each other why the machine was functioning correctly when
Operator B was coming. Their discussion began with a disagreement. They then
solved it by presenting each other with the experiments that each one had carried
out, the results of which were complementary. In this way, they finally reached
agreement and ended their conversation.
The Initial Disagreement

The structure of the initial disagreement gradually develops during the turns
(AI,..., A3). At least three types of processes take place more or less at the same
time . An interlocutory process first establishes a debate situation by the two
initial illocutions (Trognon & Larrue, 1994). A semantic process allocates
content to the propositions to be discussed. Finally, a logical process logically
defines the propositions on which the speakers will work.
The Semantic Process

While developing their debate, the interlocutors are also organizing its semantic
content. Logically speaking, they instantiate (i.e., they give constants to) the
variables that come into the propositions to be used . "Thepissettes, they seem to
be working better" from the third speaking turn onward, thus becomes "1 didn't
get any pulp," as if the interlocutors attributed a more concrete content or
established the empirical criteria that each person can experience (and has even
experienced) instead of a much more abstract assessment: "The pissettes, they
seem to be working better." Moreover, apart from the rephrasing concerning the
characteristics or the consequences of the pulp projections (e.g., breaking, cutting
of the tape), the participants then stick to this interpretation during the entire
dialogue.
Because the rearpissette "lifts the sheet" is an accepted expression from the
start, it will be represented by the constant (q). Because "The pissettes, they seem
to be working better" gradually becomes a constant during the three speaking
97
turns, we begin by showing this expression as a variable (x) until the fixing of its
semantic interpretation . From this point on, it will be represented by the constant
(p) .
Th e Inte rlocuto ry Process
This process is governed by the logical rule presented earlier, which states that the
non-fulfillment of a preparatory condition for an illocution leads to the nonfulfillment of this illocution . By stating an objection (B 1), which is expressed as
is usual during a conversation ( "And (.oo) Have you noticed?"), B denies a
preparatory condition that he attributes at the same time to the assertion from A I .
It would seem as though, in B's opinion, that A did not have good reasons for
stating AI . Let x be the proposition , "The pissettes. they seem to be working
better," and q the proposition, "it still lifts the sheet." Following x, the q
statement establishes two things. First, it establishes between the speakers, as a
product of their discursive cooperation, an interlocutory couple that introduces a
debate (Trognon & Larrue, 1994) and is a product of their discursive cooperation.
We express this in the formula (x -> not q), where x is statement Al and (not q)
is the preparatory condition, introduced by B, for this statement. Subsequently , q
expresses the position of B in the face of the previous formula; the effect of this
position is that the assertion from Al is rendered defective and thus false. This
exchange can be represented in the following way:
A1:
B1 :
A1 :
B1:
( .. . ) And the pissettes seem t o be working better .

And the on e behind, it s t i l l lifts the sheet a b it . Have
y ou noticed?
x
q
Table 4 . 1 . Representation of the debate situation introduced by the two initial

ill ocutions .
The first column represents what A, the out-going oper ator , and B, the on -coming
operator, explicitly say . The third column is what B attrib utes to A when he says B I
after AI . The second column is what B must implicitly conclude if he calculates
together premises that he attributes to A and what he himself says . This co nclusion
means that, according to B, x is a defective assertion and thus a false propo sition.
What A and B say
What B concludes
What B attributes to A
AI : x
Al B: (x --> not q)
BI: q
BI ' : not x
98
Because the assertion from B1 amounts to attributing a preparatory condition to

the assertion by A1, note that the point of view of A in terms of B (i.e., the point
of view that B ascribes to A by asserting q) is (x -> not q). This formula shows
the retrospective effect of the q assertion. It thus plays the role of a logical
introduction rule. Now, by stating q, given that (x -> not q), B is committed to
the inference (not x). This conclusion, obtained by [((x -> not q) and q) -> not x],
plus the fact that B urges A to make it, is the proactive effect of the q assertion,
so that the (x, q) sequence can be represented as shown in Table 1.
The Logical Process
By linking up with A2, A now accepts that the rear edge spray is lifting up the
sheet. He accepts the q proposition. At least we can note this for the sake of
simplicity, because he does not simply accept that the rear edge spray lifts up the
sheet: He states that he noticed it (" haveyou noticed?/umm, yes"), which actually
amounts to accepting the willful content of this perception but in a possible
world ( "maybe H). Furthermore, by following up with "But I didn't get as much
pulp as yesterday I did nothaveanypulp afterhuh," A clarifies, on the one hand,
his initial formulation ("they seem to be working better" is instantiated by "not
having any pulp"), that is, (x =p), and, on the other hand, he attributes to B the
idea that pulp projections are the result of the rear edge spray lifting up the sheet,
that is, (q -> not p), thanks to the conjunction but. The assertion by A2 has the
retrospective effect of attributing the theory that (q -c-not p) to B. The proactive
effect of the same statement is the inference [not (q-c-not p)] to which A is
committed and which B is invited to do. Indeed, it could not be (not q), because A
has just accepted q. Hence, employing the same type of representation as
previously, we show their continuing exchange:
Al :
Bl :
A2 :
Al:
Bl :
A2 :
( . .. ) And the pissettes, they seem to be working better .

And the one behind, it still lifts the sheet a bit . Have
you noticed?
Umm yes, maybe. But I didn't get as much pulp as
yesterday, I didn't have any pulp after, huh .
x
q
q but (not p)
99
Table 4.2. Representation of the logical process emerging from the three initial
illocutions.
The columns repre sent the same things as in Table I (A is the out-going operator and
B the on-coming operator). But we have added A's reply . In saying "yes, maybe , " A
agrees q. He says also that he "didn't have any pulp after, " that it is proposition p .
Thi s propositio n is a more preci se reformulation of x. A implicitl y attributes to B (q>not p) a proposition shown in the third column . Because BI A, q and p, A concludes
implicitly A2 (second column).

What A and B say
What B and A conclude
What B attributes to A and

What A attributes to B
AI: x
Al B: (x --> not q)
BI : Q
BI ': not x
A2: Q
B I A: (q --> notp)
A2: p
A2' : not
(0
--> not
0)
In the model shown in Table 2, the first column shows the illocutions
successively expressed in the convers ation; the second column shows the
"conclusions" drawn from these illocutions; and the third indicates the allocations
that the interlocutors make during their illocutions. The third column is
particularly interesting. First, let us consider that Al B is the formula that B
attributes to A retrospectively at the second speaking turn. When A clarifies his
initial formulation at the third speaking turn (x=p), Al B becomes something of
the following type: (p-c-not q). This instantiated formula can now be compared
with the formula that A retrospectively attributes to B at the third speaking turn:
(q-c-not p). The latter is the contraposition of the former, which means that the
interlocutors allocate equivalen t formulae to each other. In some ways, therefore,
the interlocu tors share the same formula: namely, (q -c-not p). It defines the
proposition being debated : Both admit q (that the rear edge cutter is lifting up the
sheet ), but A insists that there is no projection of pu lp (q and p), whereas B
insists that there is (q and not p). This conforms exactly to that which is
accomplished by the participants in the (B2, A3) pair:
B2 :
A3 :
I got s ome .
You got s ome? Me, I didn 't get it , huh .
100
The process of enacting the debate in an interlocutory way is shown in Table 3

(we have added a middle column showing the thesis that the interlocutors have in
common. which constitutes the topic of their debate; this changes slightly from
the format of the previous tables):
Table 4.3 . Representation of the debate betwee n A, the out-going opera tor, and B .
the on-coming operator.
Table 3 represe nts the debate between A and B and includes Table 2 and B2 and A3 .
which are the propos itions about which A and B contradict themselves . Table 3 differs
from Tables 1 and 2 because the first column represents the sequence of illocutions
successively uttered by the interloc utors and the first line represents the interlocutors.
Thus we see that A and B debate (q- >not p) and that B agrees with not p and A with p .
Their contradiction is situated between p and not p.
A
Illocutions
What
A
says
Al
What A
concludes
AandB
WhatB
attributes
toA
What is
common
toA
andB
B
What A
attributes
toB
WhatB
conclud es
What B
says
x--> not q
q
BI
not x
A2
A2
q -->
notp
not (q -->
not p)
q -->
not p
notp
B2
A3
Th e Resolu tion of the Disagreement

The contenders are now going to inform themselves and present the experiments
that they carried out. the results of which will allow them to reach an agreement.
The (B3..... AS) sequence carries out this process. By exchanging information. the
par ticipan ts achieve more than just logically and experimentally solving a logical
problem; they also create a common his tory of their relationship centered around
101
the malfunctioning of the machine; it is a history of the production process,

which reaches beyond their individual participation in the process . The
grammatical structures used in the conversation (e.g., tenses of verbs, personal
pronouns) particularly emphasize this activity.
B3 :
A4 :
B4 :
AS :
And I closed it off a bit because I found it was

de f l e c t i ng the sheet a bit and that made it , umm, squirt
out .
Oh , right, me what I, I , reopened it, the front one , this
morning a tiny b it , because , you know , you saw today I
broke it (the out-going operator shows the notebook for
morning, right hand side), OK, I pulled the t a il end ,
umm . . . three times .
Yes.
But the band . .. it was n't cu t . .. I h av en ' t cleaned t h e
p issett es umm, I haven't even lifted the pulp by hand .. .
nothing and t here's none on them, there are just a f ew
fibers , that's a ll OK.
Tab le 4. 4 . Representation of the on-coming operator' s (B) reasoning b y

accomplishing B2-B3.
Table 4 represents an analy sis of B2-B3, the discourse allowing B to tell A his
experimentatio n of the proposition that they had in common . The last column is the
ju stification of B' s actions (with its logical meaning) . The second column is what he
does (with its logical meaning ) in order to test his hypothesis . The first column is
what he obtains (with its logical meaning). In presenting this experiment, B argues
for not p.
I g ot some
and I cl osed it o f f a
b it
because I found it
was deflecting the
sheet a bi t and that
made it squirt out
Result of B' s action

(not p)
B' s action
(not q)
B ' s hypothesis
(q --> not p)
and by invited inference
(not q --> p)
Testing B' s hypothe sis
Reason for B' s action
102
From this standpoint, the B2-B3 interventio n is decisive as B establishes the fact
that preventing the rear edge cutter from lifting (or letting it lift somewhat less)
does not mean the end of paper pulp projection. Indeed, "I got some. ... And [
closed it offa bit because [ found it was defl ecting the sheet a bit and that made it,
umm, squirt out" could be interpreted as "I got some" despite the fact that,
although "I closed it offa bit" because "[ f ound it was deflecting the sheet a bit"
which caused pulp projection. It can also be seen that, by accomplis hing B2-B3,
B, this time fully and clearly, accepts the (q-c-not p) proposition to which he was
committed in B1 (by contraposition), which A had attributed to him in A2 while
contradicting it. The reasoning made by B is a truly experimental form of
reasoning that could be presen ted as in Table 4.4 .
Intuitively:
- (If the rear edge spray separa tes the sheet, there will be pulp projection s);
- By invited inference (if the rear edge spray does not separate the sheet, there are
then no projections);
- Therefore, I reduce the rear edge cutter, and I expect the projections to stop;
- However, they do not;
- Consequently (implicitly), it is untrue that the separating of the sheets by the
rear edge spray is a necessary condition for the projection of pulp .
At this moment in the exchange, the conditions experienced by A and B, the
"common world" in front of them, are shown in Table 4.5:
Table 4.5. Representation of A's and B' s common world after B' s reasoning.
The last column represents the state of affairs (the world) discussed by A, the outgoing operator, and B, the on-coming operator. So the last column is a set of truth
values resulting from the composition of the truth values of two propositions : q ( "that
one behind, it still lifts the sheet") and not p ( "getting some p ulp "). For instance, in
saying B2, B says that it is true for him that not p is true and q is true.
B ' s observation
not p
1
A's observation
B ' s obser vation
B' s observ ation
True for B
(cf. B2)
True for A
(cf. A2 &A3)
True for B
(cf. B3)
False for B
(c f, B3)
In short, A and B tested the configuration (q and not p) together (see Table 4.6).
103
Table 4 .6. The truth table emerging from A's and B's interaction after B's
reasoning .
This table is simply Table 5 interpreted as a truth table of propositional logic. This
truth table is the truth table of a logical disjunction. Thereby the state of affairs
discussed by A and B is the disjunction of q ("that one behind, it still lifts the sheet")
and not p ( "getting some pulp ") .
q
not p
(q or not p)
From A2 onward, A establishes that an absence of pulp projections corresponds

perfectly to the fact that the rear edge spray lifts up the sheet: (p and q) = not (qnot p); it is not true that the rear edge spray lifting up the sheet leads to pulp
projections. Furthermore, in B2 and then in B3, B himself establishes that the
pulp projections persist even when the rear edge spray lifts up the sheet less: (not
p and not q) = not (not p-c-q) = not (not q->p); it is not true that the rear edge
spray lifting up the sheet less leads to an end of the projections. A and B have
thus established jointly that (p and q) and (not p and not q) can co-exist perfectly
well. Therefore, not p and q are independent events, and not p is neither a
sufficient nor a necessary condition of q. In other words, A and B have
demonstrated together that the setting of the rear edge spray is not linked to the
production of pulp projections.
What then causes these projections? The solution comes from A in AS: It
concerns the setting of the edge spray located at the front of the machine. His
reasoning is more or less the following (see Table 4.7).
Intuitively:
- (If the front edge spray is closed, then there are projections);
- By invited inference (if the front edge spray is open, then there are no more
projections );
- Consequently, I open the front edge spray, and I expect the projections to stop;
- This is exactly what I observe.
104
Table 4.7. Representation of the out-going operator's (A) reasoning by

accomplishing A4-A5.
Table 7 represents an analysis of A5, the discourse used by A to tell B his own
experiment. The last column is the justification of A's action. The second is what he
does and the first the result he obtains from a logical point of view.
I pulled the tail

end, umm three times
.. . . bu t .. .. I
haven't cleaned the
pissett es
I , reopened it ,
the front one
because ... . I broke
Result of A's action
A's action
(p)
(not r)
A's hypothesis
(r --> not p)
and by invited inference
(not r --> p)
Testing A's hypothesis
Reason for A's action
We have just structured this reasoning in the same way as that of B. It is, in fact,
more complex, because it seems to incorporate double reasoning: The first is that
previously stated; the second can be paraphrased as "It is even more true that I
didn't have any pulp thanks to the setting I made myself, that I was driven to pull
the tail end three times for a completely different reason (?) but the band wasn't
cut." "To pull the tail end," is re-starti ng the manufacturing process with a narrow
strip of paper on which the justification is gradually increased until the desired
width is achieved. In this way, if the edge sprays had been blocked up with pulp,
the tail end would have automatically broken . The second form of reasoning
reinforces the first, whether it be independent from the first (A had already carried
out the setting, and, for a completely different reason, he had had to re-start the
process), or whether it be integrated into the first one (A had to re-start the process
because he had not previously set the edge spray). Moreover, A draws a single
conclusion from the two forms of reasoning.
The Final Agreement and Closure of the Conv ersation

B cannot but be in agreement. This is what happens in B5.The range of this
agreement is dual, proximal, and distal. From a proximal standpoint, B concludes
the analyses that the interlocutors have just done. B5 stands for a conclusion . He
presents the good operation of the machine as a product of their cooperation.
From a technical point of view, A's action made the problem disappear. However,
it is still true that the understanding of the machine malfunction is a result of the
respective analyses carried out by the two operators: Thanks to B, A knows (and
105
we know!) that the rear edge spray was not the problem; and thanks to A, B
knows (and we know!) that the malfunction was due to the front edge spray. The
different types of behavior adopted by the interlocutors toward theses sustained by
B and A should be noted. The first is logically, experimentally, and exhaustively
assessed. The simple confirmation of the second is sufficient for the interlocutors.
On the other hand, B5, referring to the two edge sprays, now explains the initial
proposition made by A at the same time as B5 justifies the illocution that induced
it as the propositional content, to which he was opposed in B1. The pissettes
seem to be working better [= there is no more pulp projection p] because they
are properly set; you say p because they are properly set. As an explanation of
AI , B5 satisfies this assertion. B changes his attitude in terms of this illocution
by justifying this illocution , which he had initially considered defective. From
this point on, the disagreement between the interlocutors is resolved.
Thus, A himself can only confirm, which is what he does in A6 by asking a
question "implying" a positive answer and by passing the microphone to B:
A7:
Ok B, I'm going to give you that (removes the transceiver

microphone)
?? (inaudible)
(gives the microphone to the on-coming operator) . So you
let that hang down because it is the emitter, right .
B6:
A8:
This sequence is only apparently external to the treatment of the problem,

because, on ending the interaction, it shows the mutual satisfaction of the
speakers. We are therefore experiencing a restoring exchange, according to
Goffman (1969), or an interactionally complete exchange, according to Roulet et
al. (1991), translating the rational resolution of the manufacturing problem,
whose model is as shown in Table 4.8 and Figure 4.2:
Table 4.8. The different phases of the conversational process between A and B
(respectively, the out-going and oncoming operators).
Table 8 represents the different phases of the conversational process between the
interlocutors . Although they agree on a macroscopic level, this process of agreement
goes through a sub-process of discussion during which they first disagree, then
contradict themselves , and finally elaborate their problem and then resolve it.
Bl
(BI, ..., A3)
(B3, ..., A5)
Development of
Disagreement
Treatment of
Disagreement
Agreement between the Operators
B5, A6-A7
106

I
Al
Bl
of cause ]but
A2
B2
A3
B3
A4-AS
L-EJI------------
B5
A6-A7
Fig. 4.2. Representation of the discursive structure of the conversation.

Figure 2 shows how the different sub-processes are organized in the whole
conversation. It is built according to the theory of the discursive structures of
conversations proposed by Rouletet a1. (1991) and Moeschler (1985) . E means an
exchange between two interlocutors . I means an intervention of an interlocutor. We
see that agreement between A and B is represented by the more global E situated on the
left of the table . We also see that the disagreement and its resolution are represented in
subordinated structures. As explained in the text, a linear linking corresponds to the
satisfaction of an illocution, and the exchanges focused on conditions of success and
non-defectivity of an illocution are translated into hierarchical structures . See
Ghiglione and Trognon (1993) , Larrue and Trognon (1993) and Trognon and Larrue
(1994) for technical explanations and examples .
Thus, this conversation, which integrates an entire set of organically connected

activities aimed at the logical resolution of a problem, constitutes for the
operators a process of the accomplishment of their work.
Conclusion
The conversation analyzed here occurred on several levels simultaneously. The
conversational aspect with the sequencing that it implies is fundamental, because
it maintains the progressive organization of the logical structure of the debate.
The logical aspect that develops simultaneously on the illocution level and on
that of proposals, which they are aimed at, is no less essential, because its process
through the conjunction of the analyses is that of the problem resolution .
107
It would still be necessary to add a textual aspect where the connectors (e.g.,
but, and, because) and the verb tenses playa very important role. Indeed, the entire
enunciative structure contributes to the resolution of the problem . For example,
"it still lifts the sheet " is stated in the present tense and indicates that, for B, the
situation has not really evolved and hence reinforces the illocutionary function of
the objection. On the other hand, "1didn't get as much pulp as yesterday, 1 didn't
have any pulp after" indicates a finished event that belongs to the past. This
suggests that a transformation took place and, in a certain way, outlines the
analysis that will follow. Furthermore, after having given an account in the past
tense of the adjustments he carried out and their results, A moves suddenly to the
present as if he is suddenly repositioning himself in the interaction he is having
with his interlocutor , meaning by that that things have now changed.
The superimposition of these different levels of operation contributes both to
the conversational production of the theory of recuperation relative to the
malfunction of the machine elaborated by the participants, and to the resolution of
the technical problem facing them. On the whole, the in situ accomplishment of
rationality is a global interactive process. In addition, what we have logically
formalized is also more or less expressed in the textual content of the
conversation; we now have to state that the distributed intelligence in the
conversation is overdetermined.
Furthermore, the conversational process creates implied social relationships
through shift work on the various operational levels. The two laborers working
on a machine in 3 eight-hour shifts must exchange information and share their
respective actions. They are required to construct a common history regarding their
relationship with the machine, which is not too far removed from reality. The
intricacy of the textual and logical processes of this operative conversation is thus
also the overall accomplishment of their work, this time, however, on a social
and economic level. It is a global social process.
108
Alain Trognon and Corinne Grusenrneyer
References
Brassac, c ., & Trognon, A. (1993, July). The logic of speech acts as a universal
grammar of intercomprehension . Paper presented at the 4th International Pragmatics
Conference, Kobe, Japan .
Davillerd, C., & Grusenmeyer , C. (1993). La releve de poste: Une periode specifique de
travail collectif [Shift changeover: A specific period of collective work] In F. Six, &
X. Vaxevanoglou (Eds.), Les aspects collectifs du travail [Collective aspects of
work] (pp. 81-88) . Toulouse , France: Octares .
Evans, J. S. B. T. (1984). Heuristic and analytic processes in reasoning. British
Journal of Psychology, 75, 451-468 .
Evans, J. S. B. T. (1989). Bias in human reasoning: Causes and consequences. Hove,
England: Lawrence Er1baum.
Garfinkel, H. (1990) . The curious seriousness of professional sociology. In B .
Conein, M . de Fornel, & L. Quere (Eds.), Les formes de la conversation [Forms of
conversation] (Vol I, pp . 69-78). Paris: CNET.
Ghiglione, R., & Trognon, A. (1993) . OU va la pragmatique? [Where is Pragmatics
going?] . Grenoble, France: Presses Universitaires de Grenoble.
Goffman, E. (1969) . Strategic Interaction. Philadelphia, PA: The University of
Pennsylvania Press .
Grusenmeyer, C. (1991) . La releve de poste: Une phase critique du travail en equipes
successives [Shift changeover: A critical phase in shiftwork] . Cahiers de Notes
Documentaires de l 'INRS, 144, 407-418 .
Grusenmeyer, C. (1992, September, 6-11). Interest of the notion of shared functional
representation in shift changeover phase . Paper presented at the 6th European
Conference of Cognitive Ergonomics , Budapest, Hungary.
Grusenmeyer, C. (1995a). De l 'analyse des communications a celle des representations
fon ctionnelles partagees. Une application a la releve de poste [From
communications to shared functional representation analysis. An application to
shift changeover] . These de doctorat , Universite Rene Descartes (Paris V).
Grusenmeyer, C. (1995b) . Shared functional representation in cooperative tasks . The
example of shift changeover. International Journal of Human Factors in
Manufacturing, 5(2), 163-176.
Hastie, R., & Pennington, N. (1991) . Cognitive and social processes in decision
making. In L. B. Resnick, 1. M. Levine, & S. D. Teasley (Eds.), Perspectives on
socially shared cognition (pp. 308-327). Washington, DC: American Psychological
Association.
Heritage, J. (1990) . Interactional accountability: A conversation analytic perspective.
In B. Conein, M. de Fornel, & L. Quere (Eds.), Les formes de la conversation [Forms
of conversation] (Vol. 1, pp. 23-49). Paris: CNET .
Hoc, J. M. (1987) . Psychologie cognitive de la planification [Cognitive psychology
of planning]. Grenoble , France: Presses Universitaires de Grenoble.
Hughes , S. (1990). Planning-in-action . Cognitive phenomenon or social
accomplishment? In B. Conein, M. de Fornel, & L. Quere (Eds.), Les formes de la
conversation [Forms of conversation] (Vol. 2, pp. 217-236) . Paris : CNET .
Johnson-Laird, P. N., & Byrne, R. M. J. (1991). Deduction. Hove, England: Lawrence
Erlbaum.
To Resolve a Technical ProblemThroughConversation
109
Karsenty, L., & Falzon, P. (1993) . L'analyse des dialogues orientes-taches:

Introduction a des modeles de la communication [Task-oriented dialogues:
Introduction to communication models] . In F. Six, & X. Vaxevanoglou (Eds.), Les
aspects collectifs du travail [Collective aspects of work] (pp. 107-118). Toulouse,
France : Octares,
Lacoste, M. (1993) . Interaction situee et dimension collective du travail [Situated
interaction and collective dimension of work]. In F. Six, & X. Vaxevanoglou (Eds.),
Les aspects collectifs du travail [Collective aspects of work] (pp. 29-49) . Toulouse,
France : Octares.
Lacoste , M., & Rogard, V. (1988). Mediatized interaction between experts in
maintenance of automated machine. Paper presented at the IFACIIFIPIIEAlIFORS
conference, Oulu, Finland.
Larrue, J., & Trognon, A (1993) . Organization of turn-taking and mechanisms for
turn-taking repairs in a chaired meeting. Journal of Pragmatics, 19, 177-196.
Levinson , S. (1983) . Pragmati cs. Cambridge, England : Cambridge University Press .
Moeschler, 1. (1985) . Argumentation et conversation [Argumentation and
conversation] . Paris : Hatier.
Nisbett, R. E., & Ross, L. (Eds.). (1980) . Human inference: Strategies and sho rt
comings of social judgment . Englewood Cliffs, NJ: Prentice Hall.
Piaget, J. (1965) . Etudes sociologiques [Sociological studies] . Geneve : Droz.
Resnick, L. (1991). Shared cognition : Thinking as social practice. In L. B. Resnick, J.
M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 120). Washington, DC: American Psychological Association .
Roulet, E., Auchlin, A , Moeschler, J., Rubattel, C., & Schelling, M. (1991) .
L'articulation du discours en francais contemporain [Discursive structures in today's
French] (3rd ed.). Berne, Switzerland: Peter Lang.
Schegloff, E. (1992) . To Searle on conver sation: A Note in Return . In H. Parret & J .
Verschueren (Eds.), (On) Searle on conversation (pp. 113-128) . Philadelphia, PA:
John Benjamins Publishing Company.
Searle, J. (1985) . Intentionality . Cambridge, England: Cambridge University Press.
Searle, J. , & Vanderveken, D. (1985). Foundations of illocutionary logi c . Cambridge,
Suchman, L. (1987) . Plans and situated action. The problem of human/mach ine
Trognon, A (1990) . Relations intersubjectives dans les debats [Intersubjective
relationships in debates]. In A Berrendonner & H. Parret (Eds.), L'interaction
communicative [Communicative interaction] (pp . 195-213) . Berne, Switzerland:
Peter Lang.
Trognon, A (1991) . L'interaction en general : Sujets, groupes, cognition s,
representations sociales [Interaction in general : Subjects, groups, cognitions,
social representations] . Connexions, 57, 9-27 .
Trognon, A (1992). Psicologia cognitiva e analisi delle conversazioni [Cognitive
psychology and conversational analysis] . In C. Galimberti (Ed.), La conversazione:
Prospettive
sull'interazione psico-sociale
[Conversation:
Interactionnist
prospective in social psychology] (pp. 115-157). Milan, Italy : Guerini Studio .
Trognon, A (1993) . La negociation du sens dans l'interaction [The negotiation of
meaning in interaction] . In J. F. Halte (Ed.), Inter-act ions [Interactions] (pp. 91121). Metz, France : Presses de I'Universite de Metz.
110
Trognon, A. (1994). La pragmatique et la vie de tous les jours [Pragmatics of ordinary

life] . In A. Trognon, U. Dausendschoen-Gay, U. Krafft, & C. Riboni (Eds.), La
construction interactive du quotidien [Daily interactive construction] (pp 10-55).
Nancy, France : Presses Universitaires de Nancy.
Trognon, A., & Brassac , C. (1992). L'enchainernent conversationnel [Conversational
sequencing]. Cahiers de linguistique francaise, 13, 67-108.
Trognon, A., & Larrue, J. (1994) . Pragmatique du discours politique [Pragmatics of
political discourse] . Paris : Armand Colin .
Trognon , A., & Retornaz, A. (1989). Clinique du rationnel: Psychologie cognitive et
analyse des conversation [A clinical approach to the rationale: Cognitive
psy chology and conversation analysis] . Connexions, 53, 69-91.
Vanderveken, D. (1990) . Meaning and speech acts (Vols 1 & 2). Cambridge, MA:
Cambridge University Press .
Wason , P. C., & Johnson-Laird, P. N. (1972) . Psychology of reasoning. Structures and
content. Cambridge, MA: Harvard University Press .
Chapter 5
The Blackness of Black:

Color Categories as Situated Practice
Charles Goodwin
Department of Anthropology, University of South Carolina
Columbia, USA
Abstract
In what rem ains one of the central accomplishments of cognitive anthropology, Berlin
and Kay (1969) demonstrated that the diversity of human color systems was built on a
univer sal infras tructure, with black and white being the most basic colors in all
systems. The analytical focus of their work is a structural system divorced from the
messy tasks of actually using color terms to make relevant distinctions within specific
courses of action situated within the concrete settings that constitute the lifeworld of a
particular society. By way of contrast, Wittgenstein' s later philosophy argues that it is
precise ly such endoge nous activities that provide the necessary framework for the
analysis of human langua ge. Using as data video tape of chemists attempting to
determine when to stop a reaction by deciding when the material they are working
with is jet black , this chapter explores (1) the diverse practices they deploy to
establish what can count as black; (2) how such a distinction is embedded within a
local activity system lodged in turn within a relev ant community of practice; and (3)
the embodied apprenticeship required for new members to become competent in the
use of such a category. For the chemists, jet black (e.g., the most prototypical example
of black) is not a preformulated, context-free universal color category, but instead a
problematic j udgme nt to be artfully accomplis hed through the depl oyment of a
collection of systematic work practices. This analysis contributes to the development
of a practice-based theory of knowledge and action.
112
Charles Goodwin
Two of the central and enduring topics in the analysis of cognition are the study
of vision (in neurophysiology, vision is the cognitive system that is best
understood, and its architecture provides a point of departure for the analysis of
how the brain organizes other types of representations) and semantic categories
(which at times have formed the essential subject matter for whole fields such as
cognitive anthropology) or, more generally, processes of classification. One
crucial place where these two lines of research intersect is in the analysis of color
categories, terms provided by language that are used to codify and structure
perception of the visual field. Different languages classify the color spectrum in
different ways. This has been argued to provide evidence for the Sapir-Whorf
hypothesis that language structures perception of the world that a particular
society inhabits (Bruner, Oliver, & Greenfield, 1966; Greenfield & Bruner,
1966). In what remains one of the central accomplishments of cognitive
anthropology, however, Berlin and Kay (1969) demonstrated that the diversity of
human color systems was built on a universal infrastructure, one almost certainly
linked to structures in the brain.
The focus of analysis in such work is an abstract structural system, divorced
from the messy tasks of actually using color terms to make relevant distinctions
within specific courses of action situated within the concrete settings that
constitute the lifeworld of a particular society. In contrast, Wittgenstein' slater
philosophy argues that it is precisely such endogenous activities that provide the
necessary framework for the analysis of human language.! Analysis of category
use from such a perspective has been a major focus of research by
ethnomethodologists and conversational analysts (e.g., Cicourel, 1964; Garfinkel,
1967; Heritage, 1984; Jefferson, 1987 ; Lynch, 1991; Sacks, 1992; Schegloff,
1972; Suchman, 1987) . Recently, some scholars have begun to analyze seeing as
a social process lodged within endogenous communities of practices (Goodwin,
1994, in press; Goodwin & Goodwin, in press; Heath, in press, this volume;
Heath & Luff, in press; Saljo & Bergqvist, this volume).
Color Categories as Cognitive Universals: Divorcing Cognition

from Practice
Berlin and Kay (1967, 1969) demonstrated that all languages locate the foci of
thei r basic color labels at roughly the same place in the color spectrum and ,
Iproblems with Wittgenstein's initial treatment of color perception provided the Achilles '
heel that led to the dismantling of the Tractatus (Hacker, 1986, p. 109) and its replacement
with a theory of color terms that emphasized their situatedness within the grammars of
diverse natural actitivies.
The Blackness of Black
113
moreover, that a universal pattern exists for adding basic color terms to the
language:
Black
White
~ Red ~ Green ~
Yellow
Blue
~ Brown ~ Purple
Pink
Orange
Gray
Fig. 5.1. Berlin and Kay's Universal Sequence of Color Terms
Berlin and Kay's Universal Sequence of Color Terms

The contrast between black and white is found in the color systems of all
languages. If a language has only three basic color terms, the third term will be
red ; if it has four, the next term to be added will be either green or yellow, etc.2
Language universals, most probably based on a neurological infrastructure, have
emerged from more detailed examination of phenomena that initially seemed to
provide some of the strongest evidence for cultural and linguistic relativity.
All of this research used a model of language and cognition that was consistent
with Saussure's (1959, pp. 6-15) formulation of the distinction between langue
and parole. Two features of this model are especially relevant to the presen t
analysis. First, all phenomena of interest are located inside a Cartesian mind that
must be examined in isolation from the act of speaking. The content of this mind
is socially produced.' but most of the heterogeneous phenomena implicated in
acts of speaking (e.g., the processes of social interaction within which talk is
embedded) must be excluded from analysis (Saussure, 1959, p . 9). Second,
Saussure's conceptualization of langue explicitly excludes human agency :
"Language [langue] is not a function of the speaker; it is a product that is
passively assimilated by the individual... . It is the social side of speech, outside
the individual who can never create or modify it by himself' (Saussure, 1959, p.
14).
Urhe work of Berlin and Kay generated a substantial body of subsequent research, some of
which led to modifications of their original typology. Such modifications, however, are not
consequential for the analysis in this chapter.
3According to Saussure, langue "is a storehouse filled by the members of a given
community through their active use of speaking, a grammatical system that has a potential
existence in each brain, or, more specifically , in the brains of a group of individuals . For
language is not complete in any speaker; it exists perfectly only within a collectivity"
(1959, pp. 13-14).
114
Charles Goodwin
Despite the brilliance of Saussure's insights here and the very great payoff they
have had in the subsequent development of linguistic theory, the way in which he
formulated langue had the inevitable effect of divorcing cognition from practice.
All of the cognitive work involved in coordinating talk and meaning with the
actions of coparticipants and of using language to build a relevant social world in
actual settings is treated as epiphenomenal. Cognitive phenomena, including
categories for the organization of perception , are situated analytically within the
structural system of the language as a whole, and the process of constituting their
meaning lies beyond the grasp of speakers.
Situated Activity Systems

Implicit in the work of Berlin and Kay and in the paradigm created by Saussure
that it builds on are crucial assumptions about the geography of human cognition,
for example, specifications of where relevant cognitive phenomena are to be
found (in the brain and structured systems of mental representation) and where
such phenomena are not located (e.g., in actual speech, multiparty discourse,
material objects, the environment around a group of human actors; see Saussure ,
1959). Within this research tradition, there is a theoretical and methodological
emphasis on coherent, self-contained, modular units: for example, the sound
system of a language , or, in the realm of semantics, taxonomies of structurally
related phenomena, such as kinship systems (Goodenough, 1956) or color terms.
The tremendous advantage of restricting data to categories in a single, bounded
taxonomy is that it permits analysis to focus not on the properties of individual
items (e.g., a specific color term), but instead on a more basic structural system.
By circumscribing a single, internally consistent domain of phenomena, the
system provides for the constitution of meaning through systematic contrast
within a well-defined set of possibilities .
However, a price is paid for that analytical clarity. The very properties that give
the system its precision and coherence - its boundedness and restriction to a
single kind of phenomena - make it impossible to investigate aspects of cognitive
organization that cross such boundaries. Thus, although Conklin (1955) had
demonstrated convincingly that color terms frequently incorporate information
from other sensory modalities (e.g., among tropical forest hunters and gatherers, a
term for green may include notions of succulence and freshness), Berlin and Kay
systematically excluded such phenomena from their study . By doing this, they
were able to analyze color terms as a bounded, self-contained system but could
not (and did not want to) in any way take into account either how color terms
might be shaped by systematic patterns of situated use, or the possibility that
actors might deploy a range of different kinds of criteria in order to categorize
color.
115
Moreover, work in a variety of different fields has called into question the
assumption that human cognition operates within such neatly bounded packages .
Thus , in different ways, both Heidegger (1962) and Wittgenstein (1958) argue
that human cognitive activity is inextricably lodged within the activities and
settings of the lived social world: that is, that knowledge is intrinsically situated.
Strong support for such a position has come from the investigation of how
scientists actually do their work (Latour, 1987; Lynch & Woolgar, 1988;
Pickering, 1992), studies of cognition in the workplace (Heath & Luff, in press;
Middleton & Engestrom, in press; Rogoff & Lave, 1984; Suchman , 1987),
anthropological investigations of systems of cognition encompassing multiple,
differentiated actors and tools (Hutchins, 1993), practice theory (Chaiklin &
Lave, 1993), and conversation analysis (Heritage, 1984; Sacks, 1992; Schegloff,
1972, 1992b). All of this work demonstrates that a diverse collection of
heterogeneous phenomena and processes are implicated in human cognitive
activity : for example, not only mental representations but also material tools,
historically shaped and socially distributed forms of knowledge, processes of
social interaction and the forms of social action they produce , and recognizable
patterns of activity in a specific setting.
Restricted taxonomies provide analytical coherence by restricting research to
the study of patterned variation in a single domain of possibilities: for example,
ways of naming colors. If one wants to move beyond single taxonomies ,
however, the order provided by a data set structured in terms of comm on
underlying features is lost. The question thus arises concerning how a somewhat
ad hoc collection of very different kinds of entities can be related to each other
within a common analytic framework; for example, how can objects as diverse as
language categories, physical tools such as the vats and sticks that will be
examined later in this chapter, and social distributions of knowledge and power in
a specific work setting be studied as interdependent components of an integrated
cognitive process? The solution chosen in this chapter is to focus on what I call a
situated activity system." that is, the range of phenomena implicated in the
systematic accomplishment of a specific activity within a relevant setting . An
example of a situated activity system is provided by a game such as hopscotch
(M. H. Goodwin, in press), which integrates into a common framework of action
4-rhe term situated activity system was introduced by Goffman (1961 , pp . 95-99) to
describe repetitive encounters in social establishments in which an individual is brought
"into face-to-face interaction with others for the performance of a single joint activity, a
somewhat closed , self-compensating, self-terminating circuit of interdependent actions"
(1961, pp. 95-96). Goffman 's interest in moving role theory in new directions is different
from my use of the concept to investigate how properties of an encompassing, situated
activity shape cognition within specific settings. Such differences do not, however, detract
from my complete agreement with Goffman 's central argument: "The point about looking
at situated activity systems is that the complexities of concrete conduct can be examined
instead of by-passed" (1961, p. 99). For an analysis of situated activity systems constituted
through talk in local settings, see M. H. Goodwin (1990).
116
Charles Goodwin
and socially organized perception a collection of very different kinds of events,

including physical inscriptions in a public, material environment (e.g ., the
hopscotch grid), roles for different kinds of participants, rules differentiating
successful from unsuccessful action , game-relevant tasks of seeing and moving,
specifications for how actors should hold their bodies, and systematic language
practices for calling and contesting "outs."
In this chapter, I focus on the activities of a team of geochemists who are trying
to figure out when to terminate a chemical reaction they are monitoring. In their
work, the job of discriminating colors is posed as a consequential task. The
situated activity system to be examined is the ensemble of practices and tools
deployed to determine when the materials in the reaction vat arejet black, that
being the diagnostic sign that the process being scrutinized has run its course and
should be immediately quenched. As described in more detail throughout this
chapter, this activity system provides organization for the cognitive work
occurring in this setting in a number of different ways. Thus, it sets parameters on
what color shades are encomp assed by an appropriate definition of black. By
virtue of its consequences for practical projects (e.g., producing something that
will work), what will count as valid instances of the color category is established
within a public, socially constituted world of relevant activity, rather than in the
mental processes of an isolated actor. This does not mean, however, that the
cognitive operations of young chemists, trying to figure out if the material they
are working with is black yet, are irrelevant; instead, the activity creates an arena
for situated apprenticeship as newcomers train both their bodies and workrelevant perceptual structures to the demands of the activity (e.g. , become
competent practitioners, through interaction with both more experienced chemists
and the materials being manipulated). Material objects and mental representations
are integrated into a common cognitive process by the situated activity system.
The task of successfully bringing the reaction to completion provides a focus for
the perceptual activity of the chemists and motivates them to scrutinize their
material in terms of its color rather than, for example, its weight or any of a range
of other equally available attributes. By virtue of the encompassing activity,
chemists involved in this task are not disinterested observers; they are extremely
interested actors whose perception is being shaped by orientation to a set of
relevant contingencies posed by the tasks they are attempting to accomplish.
Analysis of situated activity systems' provides one way of investigating how
5Such systems are frequently lodged within settings, such as the geochemists' lab, in which
collections of tools have been brought together to deal with particular kinds of tasks. Other
examples include centers of coordination (Suchman, this volume; Heath & Luff, in press)
such as Operations Rooms , and business and educational establishments. The same
personnel work repetitively in such settings (although with considerable differences in
experience because newcomers arrive and oldtimers leave) and thus develop skill in both
handling the tools located there and dealing with the range of tasks that the setting is
organized to accomplish. Such a setting constitutes a historically shaped environment of
possibilities for action. Although this chapter focuses only on a particular situated activity
117
cognitive phenomena, such as color categories, are constituted through the social
deployment of a collection of diverse practices lodged within the lifeworld of a
relevant community of practice.
Scientific Description as Embodied, Situated Knowledge

Saussure's vision of the social nature of language as a shared system of
prespecified meaning and structure that is internalized by each speaker had
consequences far beyond linguistics. Thus, "the institutionalization of commo n
meanings for symbols in advance of their use in particular situations" (Heritage,
1984, p. 28) lies at the heart of Parsons and Skills' (1951) solution to the problem
of how socia l actors can know a world in common. Like Saussure, Parsons treats
the actual cognitive work engaged in by actors to build situated meaning and
action as epiphenomenal.
Scientific writing is one place in which the assumptions made by both Saussure
and Parsons about the location of meaning in a prespecified system of symbols
can be sharply tested. For Parsons, rationality and intersubjectivity,
uncontaminated by error , are possible only when the knowledge used by social
actors coincides with the findings of science (Heritage, 1984, pp. 24-30). More
generally, in both philosophy and the social sciences, scientific writing has
traditionally been treated as the prototypical example of rational description.
However, recent research on the sociology of scientific knowledge has strongly
and clearly demonstrated that, like other domains of activity, scientific knowledge
is constituted through the deployment of a range of socially organized practices
(Knorr-Cetina, 1981; Latour, 1987; Lynch, 1985; Pickering, 1992; Star, 1988a) .
Building on this work, I focus on a particular genre of scientific writi ng : a
description of basic laboratory procedures. If the assumptions made by Saussure
and Parsons about prespecified meaning are correct, they should apply here. If
knowledge is, in fact, abstract and disembodied (i.e., capable of being comp letely
formulated in the language that appears in a journal article), anyone with access
to proper equipment should be able to use that description as a recipe and to carry
out the procedure themselvesv Other competent scientists can use a description in
this fashion, and such replicability is one of the hallmarks of science as an
system, this larger setting must be kept in mind. Methodologically, the way in which such
workshops develop through time and practice, relevant toolkits and systematic solutions to
the repetitive problems they encounter (such as classifying color for the chemists)
demonstrates the importance of investigating situated activity systems within the natural
environments where a society's work is done.
&rhe work of Collins (1985) on the situated work required to build an existing scientific
instrument from scratch in a new lab demonstrates how much practical knowledge is
required to translate a written description into a working machine.
118
Charles Goodwin
institution . However, the ability to translate such instructions into workable

products frequently builds on an ensemble of embodied competence and tacit
knowledge (Polanyi, 1966) acquired by newcomers to a profession in the labs and
field settings where the work of their discipline is done.?
To further explore how language is used in such a process, I examine a specific
work situation:
Fig. 5.2. Work in a Geochemistry Lab
The participants in this setting are a geochemist and his students who are making
a scientific instrument. I videotaped what they were doing (from a specific
location that gave me better access to some aspects of what they were doing than
it did to others) but did not help in the work. By depositing manganese oxide on
acrylic fibers, the participants create a tool that is capable of extracting virtually
all of the radium ions present in a sample of water. Different bodies of water
(e.g., different rivers) have distinctive radium signatures. By using the fiber, the
7See Goodwin (1994) for an analysis of the socially organized practices of seeing that must
be mastered by a young archaeologist in order to make one of the standard documents of
her discipline : a map of an excavated section of dirt. Although the young archaeologist
know s the linguistic meaning of rules telling her where to take measurements (e.g.,
"wherever there ' s a change in slope") , finding what counts as such an event in the
complicated perceptu al field provided by the landscape in front of her is a complicated,
contingent process. Mastery of such ability is something that all competent archaeologists
expect of each other, and it constitutes part of the embodied infrastructure required for
proper understanding of the writing practices that constitute archaeology as a profession .
119
geochemists can make extremely subtle measurements about the distribution of

water from different sources in the world 's rivers and oceans (e.g., determining
how much of the water in the seas off Puerto Rico came originally from the
Amazon river) . Because the fiber is one of the crucial links in the chain of
scientific work that leads to such findings, explicit instructions describing how
the fiber is manufactured are published in the scientific literature. The following
is the standard reference:
To prepare the improved fiber, acrylic fiber (such as Monsanto "Acrilan , " 3.0 denier,
type B-16) is immersed in one fiber volume of 0.5M at 70 to 80 C. The
Permanganate solution partially oxidizes the fiber and deposits Mn02. on it. The
reaction is stopped after about 10 min . by removing the jet-black fiber and washing it
in deionized water. The exothermic reaction is rapid and produces considerable heat;
therefore the transfer from the reaction solution to the wash solution should be
completed quickly. After washing and partially drying, the fiber is separated (fluffed)
and is then ready to use. The fiber can be prepared at lower temperatures using longer
contact times. At 30 C about 3 days are necessary to blacken the fiber completely.
(Moore, 1976, p. 647)
This fiber is used by approximately 10 to 20 geochemists worldwide . The lab that

was taped is that of the geochemist who invented the fiber and wrote this
description. He is making a batch for upcoming cruises with a group of students
who have never made it before. Although these participants were not actually
working from the printed article (the students had the author himself to learn
from), for purposes of the present analysis, it is useful to use the journal
description as point of departure for investigating their work. First, this is the type
of model (e.g., descriptions published in scientific journals) that has typically
been used by philosophers and social scientists when they point to the rational,
disembodied properties of science. Second, it is here that the properties of
language noted by Saussure and Parsons can be most clearly examined. For
clarity, analysis will focus on how the term black in the following sentence is
interpreted by those making the fiber: "The reaction is stopped after about 10
min . by removing the jet-black fiber and washing it in deionized water."
Restricting analysis to what counts as black in this process has a number of
advantages. For example, although some language in the description consists of
esoteric craft terms (e.g., "exothermic reaction," "0.5M potassium permanganate
solution") that require competence in chemistry for their proper understanding,
the term black is known by everyone who speaks English. This makes one aspect
of the phenomena that the participants in the lab are working with available to
readers with little background in chemistry; and it also constitutes a hard case for
the social construction of the entities used by scientists, by focusing not on a
conceptual object (e.g., a quark) that is brought into existence and changed as
theory develops (such that changes in theory can be used to demonstrate changes
in the object being attended to), but instead on the description of something being
worked with by the scientists (a jet-black fiber), that is apparently available to
naked perception .
120
Charles Goodwin
Seeing Jet Black as a Problematic, Situated Task

As indicated earlier, color terms have received extensive study in anthropological
linguistics. Note that the clear and strong findings of Berlin and Kay (1967, 1969)
make it extremely difficult to treat black, the color term used to describe the
fibers, as arbitrary or idiosyncratic to the color system of a particular language or
group. The contrast between black and white sits at the very apex of their
hierarchy of universal terms: That is, all languages will make this basic
distinction and make it before they make any others . Moreover, according to
Berlin and Kay, all languages locate the focal point of a color at roughly the same
place. The fiber being made by the scientists is described in the journal article as
"jet-black." This expression seems to be designed precisely to declare that the
black at issue is the blackest of blacks, that is, the focal point for defining what
constitutes the color, the most prototypical case. If any color term could claim the
status of a context-free universal, this would seem to be it.
However, as noted by Vincent van Gogh in a letter to his brother Theo:
. . . we of course agree perfectly about black in nature. Absolute black does not really
exist. But like white, it is present in almost every colour, and forms the endless variety
of grays, - different in tone and strength. So that in nature one really sees nothing else
but those tones or shades . (reprinted in Roskill, 1983; p. 158, emphasis added by
current author)
As van Gogh recognized, the existence of a term such as black within the
semantic space of a particular language in no way solves the problem of how
what counts as black in nature is to be determined. Practitioners, such as the
geochemists being investigated here, who wish to use the category to locate
something relevant to their work are not given a solution to that problem by the
term itself. Instead, they are faced with a task: that is, how to find a specification
for black that can distinguish tones within "the endless variety of grays" in a way
that is appropriate to the activities in which they are engaged.
Several general processes provide organization for the work involved in the
task of determining a relevant specification for jet black.
Situated Activities as Frameworks for Motivation and Precision
Determining what will and will not count as a proper referent for a category in a
specific setting is lodged within larger activity structures . Establishing when the
fiber is jet black is important to the geochemists, because that color is the
diagnostic sign that the chemical reaction has proceeded to the point where it
should be quickly stopped. If the fiber is put into the quenching bath before it
exhibits the proper shade of black, it will absorb less radium ions when used later
121
as a tool for measuring different bodies of water. The color change is the simplest
measurement that can be made to indicate the progress of the reaction.f The
larger activity of making the fiber thus provides a motivational framework that
leads those involved in the activity to make particular perceptual distinctions in
the first place (i.e., it establishes a texture of relevancies, a focus for perception).
The activity also establishes the parameters of what will count as a correct
solution to the task of identifying black in these specific circumstances (i.e., a
range of shades that will lead to usable fiber if the reaction is stopped when they
appear). Clearly, other tasks would set other parameters. Moreover, different
tasks will set relevant standards for accuracy and precision at different places?
For some, a very wide range of shades might count as acceptable solutions,
whereas, for the geochemists being investigated here, a much more limited,
precise sense of what can count as an acceptable black is necessary if they are to
succeed in terminating the reaction at the proper moment. This precision arises
not from the status of their work as science, but rather from the specific task at
issue. In other scientific tasks involving the fibers, measurement could encompass
a much wider range of variation. Thus, when the fibers were used to gather data,
the scientists simply loosely filled the collection tube with fiber. Because the
fibers were so efficient in extracting radium from the water, in essence getting all
of the radium in a given sample of water, it was not considered necessary to
measure precisely the amount of fiber being used. For more extensive discussion
of measurement as a situated phenomenon, see Cicourel (1964), Lynch (1991),
and Sacks (1989).
The Social Organization of Practice and Apprenticeship Within Situated

Processes of Human Interaction
Although the encompassing activity sets constraints on what can count as a

solution to the perceptual tasks it makes relevant, it does not specify the solution.
Finding what will count as the proper black must be discovered by the
participants as they engage in the activity. This might be accomplished in a
number of different ways. For example, someone not yet familiar with the process
might withdraw fibers colored in slightly different ways and then see which ones
did and did not work. More typically, learning what will count as a proper
solution to a problem such as this is embedded within organized processes of
social interaction. In the lab examined here, one of those present (the senior
81[ necessary, other measurements can be taken later to check the color judgment.
However, as a way of working skillfully and efficiently, this professor teaches his students
to try to use the easiest measurement to get the necessary result and to check that later.
9See Heritage (1984) for extended discussion of how the prevasiveness of the documentary
method of interpretation as a resource for making definite sense with indefinite resources
leads to "an inherently approximate relationship between a description and the range of
states of affairs it may be used to describe " (p. 145).
122
Charles Goodwin
professor) had already successfully made the fiber (he had invented it). His
memory (as the sedimented product of prior practical engagement in the process)
can be used to shape not only his own actions but also those of the newcomers he
is supervising. In the lab, this is formalized into an organized system of
apprenticeship (e.g., a professor guiding the work of his students). In the lab
process examined here, the professor let his students carry out the tasks involved
in making the fiber (and also helped himself), while monitoring what they were
doing and evaluating their decisions.
Let us examine several examples of their interaction. Talk is transcribed using
the system developed by Gail Jefferson (Sacks, Schegloff, & Jefferson, 1974).
Talk receiving some form of emphasis (e.g., talk that would be underlined in a
typewritten transcript using the Jefferson system) is marked with bold italics .
Punctuation is used to transcribe intonation: a period indicates falling pitch; a
question mark, rising pitch; and a comma, a falling rising contour such as would
be found after a nonterminal item in a list. Comments (e.g., descriptions of
relevant nonvocal behavior) are printed in italics. Numbers in parentheses mark
silences in seconds and tenths of a second. To make it easier for the reader to find
a place in the larger transcript that is being discussed in the current analysis, I
have sometimes highlighted that talk by drawing a box around it.
In Example 1, Billy notices that one of his students, Gina, is about to check the
color of the fiber in the batch she is working on. He shakes his head from side to
side (i.e., No) (line 2) and says, "It' s not-" (line 3). She interprets this as an
evaluation of the current state of the fiber and asks (line 6), "It' s not even clo:se?"
Before she has completed the word clo:se, he overlaps with a No, affirming that
the fiber is notyet ready.
Here judgments about what constitutes the proper shade of black are calibrated
within the work group. The participants treat applying the category to the fiber as
something to be artfully accomplished, a topic for discussion in its own right.
Note that, although the view of the professor prevails, his word is not blindly
accepted . In lines 13-14, Gina offers a mild challenge to his assessment of the
situation ("I don know. It's not looking that far... "). He counters this proposal by
noting that the fiber "still has that tinge" (line 17). Through this exchange, her
attention is drawn to finer perceptual discriminations that she should take into
account in judging the color of the fiber. The notion of what can count as black is
not static but rather something that is progressively shaped and modified as
participants inspect the changing materials they are working with, while
interacting with each other.
123

Example 1
1
2
3
4
5
6
7
(Gina positions squirtbottle over Vat

BillyShakes head "no"from side toside)
Billy:
It's not-
(Shakes head "no"from side toside again)

(1.5)
Gina:
Billy:
It's not even cl0rse?

No: :.
8L...---'-----;::7.':"---:------,;--;--;-----,:;-------:----;------'
(Billy turns and picks up thermometer)
10
11
12
13
(1.0)
Billy:
(1.9)
Gina:
14
15
16
17
18
Billy:
Gina:
19
20
21
22
23
24
I'don'know
it's not lookin that far (
How is it.
(An::: d)
It still has that-rimae.
L(Yamean tal
Ye(h)a(h)h.
(1.7)
Gina:
But it's gettin a little- (points withstick)

the lumpy texture.
Billy:
We:ll,
Billy:
The temperature is doin we:ll.
(1.2)
25
26
(Maybe we'd better take another one)
(Gina sprays water onfiber)
(0.9)
Fig. 5.3. Example 1
Inventing New Category Systems Tailored to the Local Setting
Upon encountering the limitations of off-the-shelf tools (the general lexicon of a

language), participants can tailor those tools for specific tasks or make new ones
more relevant to their needs through a process of situated improvisation (BrunCottan et al., 1991; Suchman, 1992; Star, 1988b). Although the term jet-black
serves to officially record the process within the scientific literature, it does not
make salient the precise distinctions between shades of closely related colors that
those involved in assessing the fiber are required to make . To highlight the
relevant perceptual distinctions, Billy, the senior scientist who invented the fiber,
coined another set of terms to guide the looking of those making the fiber within
124
Charles Goodwin
his lab. Fiber that had reached the desired color was referred to as gorilla fur,
whereas fiber that was not yet the right color was called orangutan hair. Rather
than focusing discrimination entirely on two very similar shades of color, this
new set of terms contrasted two distinct types of animals, each of which
incorporated within its gestalt one pole of a color distinction that was relevant to
the activity at hand. The perceptual distinction that counted for those making the
fiber was thus highlighted. The humor in the new contrast set not only facilitated
memory and heightened salience but also incorporated an affective stance into the
perceptual distinction . In addition to color, the gorilla fur/orangutan hair contrast
encodes another dimension of the material being manipulated: its fibrous
qualities. It thus provides a richer evocation of the sensory environment of the
task than a color term alone would.
By being more salient, specific, concrete, and humorous, the new contrast
provides a tool that is simultaneously more powerful and better adapted to the
specifics of the environment within which it will function than the more abstract
jet black was. However, much research on the language used by scientists has
taken as its point of departure precisely the opposite set of assumptions. Thus,
Bernstein (1972) distinguishes elaborated codes from restricted codes specifically
in terms of how context-bound they are. Restricted codes that "sensitize their
users to particularistic meanings" are inferior to elaborated codes, the language of
science, which "orient their users toward universalistic meanings" (p. 164).
Similarly, Parsons' view of science emphasizes "conceptual abstraction from the
concrete" (Heritage, 1984, p. 19). In these data we find scientists moving in
precisely the opposite direction, actively inventing particularistic, restricted codes
when they already have access to far more universal, less context-bound
categories. Moreover, they have very good reasons for doing this: Categories
attuned to the particulars of the work that they are doing make relevant features of
that work more vivid and salient and thus help them to perform that work.
Viewing such issues from a broader perspective, Schegloff (1972) has argued that
the really difficult and interesting issues posed in the analysis of cognition
concern not the development of abstractions but rather the analysis of systematic
procedures capable of building the particulars of local events in a way that is
sensitive in detail to the structure of relevant context in those events .
If, following Vygotsky, we think of language as a tool for mediating our
relationship with the world (e.g., a term such as black mediates our perception of
the materials being worked with), the gorilla fur/orangutan hair contrast provides
an example of a second level of mediation being bootstrapped on top of a first.
These terms mediate between black and the fibers whose color is being assessed
by tailoring the general color distinction encoded in black to the perceptual tasks
faced in a specific local situation.
Recent work in the sociology of science has called into question traditional
notions of authorship by focusing attention on the contributions of workers, such
as lab technicians, whose crucial, highly skilled practice was central to the
findings reported in publications but made invisible there (Shapin , 1989). Here
125
we find that similar processes occur with cognitive structures as well. The
humorous contrast between gorilla fur and orangutan hair provides organization
for the perceptual work involved in making the fiber but disappears in
publication, to be replaced by a less useful but more abstract and general
category.
Highlighting and Positioning for Perception

A range of work is required simply to make the object available for relevant
perception. In the lab, the chemical reaction occurs when the fibers are immersed
in a deep purple potassium permanganate solution .
Fig. 5.4. Fibers in a Vat of Chemicals
Fig. 5.5. Positioning Fibers to Judge Their Color
126
Charles Goodwin
It would be impossible to make a fine color discrimination while the fibers were
sitting in this liquid. Before any color judgment can be made, they must be
positioned for perception. To do this, a stick was used to lift a sample of fiber
from the liquid and hold it over the vat. The purple solution clinging to the fiber
on the stick was then washed away by spraying water on it.
Only after the fiber had been extracted from the very complex background
within which it was otherwise enmeshed could its color be evaluated. The activity
of assessing the fiber is thus supported not only by work-relevant cognitive
structures (e.g., an encompassing activity, category systems of various types), and
frameworks for interaction, but also by sets of situated practices for appropriate
manipulation of artifacts in the setting and by a tool kit that makes those practices
possible (e.g., plastic spray bottles of deionized water, sticks, vats, buckets,
chemicals). Although operating in the physical world, the practices engaged in
here are analogous to the cognitive structures given such prominence by the
gestalt psychologists (e.g., extracting a figure from a ground) and more recently,
used by anthropological linguists to shed new light on the nature of reference and
indexicality (Hanks, 1990, 1992) . Moreover, such processes of enhancement are
central to the production of knowledge in science more generally, constituting
what Lynch (1988) has termed an externalized retina. They are very important in
other work situations as well. For example, some of the most pervasive activities
found at the airport studied by Xerox PARC's Workplace Project involved tools
and practices designed to highlight phenomena (e.g., the use of yellow highlighter
to make information relevant to the task at hand stand out on a document) and
amplify perception (e.g., video links that allowed visual access to distant
locations). A major topic that has been virtually overlooked in the analysis of the
organization of gaze and other body behavior in face-to -face interaction is the
range of movements and activities involved in positioning for perception . The
cognitive task of assessing the color of the fiber in the lab would be impossible
without practices designed to extract the fibers from an irrelevant background and
to massage them into phenomena that are available for work-relevant perception.
Seeing Activities
The systematic incorporation of these practices into the activity of manufacturing
the fiber produced a framework for intelligibility that enables one party to make
inferences about what another is doing. In Example 1, Billy's negative head
shake, which was interpreted by Gina as proposing that the fibers were not ready,
occurred just after Gina positioned her squirt bottle over the vat. By seeing this
action within the larger framework of the activity system, Billy was able not only
to infer what she was doing but also to make an assessment about her perceptual
competence. These practices thus form part of the texture of intelligibility that the
participants are deploying to infer intersubjectivity and to make sense out of the
activities that they are collaboratively performing.
127
Embodied Cognition
Overwhelmingly, theories of cognition have divorced the mind from the body,
treating the latter (with the exception of the brain) as irrelevant to analysis of how
human beings think. This prejudice extends to other aspects of human social and
cultural behavior. Thus , Scheper-Hughes (1994) has noted critically that "The
body in social anthropology emerges as a passive, inert, dead weight attached to a
lively , responsive, nomadic mind, the true agent of culture" (p. 231) . Although
science is frequently depicted as the prototypical exemplar of disembodied
abstract thought, these geochemists were consistently attuned to what the
experience of working with the fiber might reveal to their bodies in an ad hoc
fashion as the activity unfolded. In Example 2, after making several batches of
fiber, a process that has required her to squeeze and manipulate the fibers at
various stages within the production cycle, Gina comments to the professor, Billy,
that, when the fiber is done, it has not only a particular color but also a distinctive
texture:
Exarnple2
Gina :
En it gets a certain
""1
to it
Billy Starts to Nod

3
4
5
6
7
8
9
10
11
12
(0.3)
Billy:
Gina :
[Yes.
Also,
When it getstght. (fingers "feeling"fibers gesture)
Billy:
Gina :
Billy:
Gina:
Billy:
It certainl}1[~oes.
It gets sorta
It certainly does.
En that wasn't quite enough. (repeats "feel" gesture)

Yeah.
Uhhuh.
Fig. 5.6. Example 2
Billy immediately agrees with Gina's observation (he starts nodding before she
has finished the word texture). Alertness to the sensations experienced by her
body as she manipulates the fiber while evaluating its current state of progress
has revealed to Gina the possibility of another diagnostic criterion, one that is
available through touch rather than sight. This possibility was made known to her
not by instruction from her professor , but rather through the embodied process of
physically working with the fiber. The presence of her mentor is not, however,
irrelevant. By talking to him about what she has experienced, she is able to
transform what might otherwise remain private sensations and hypotheses into
128
Charles Goodwin
public events that can be evaluated and confirmed (or denied) by a more
competent practitioner. Note that an informed evaluation of her observation is
possible only from another body that has also physically worked with the fibers .
The embodied nature of the phenomenon that Gina and Billy are constituting
together is aptly demonstrated by the way in which both specify the experience of
the texture with hand gestures rather than words. A main focus of recent work on
gesture has been on how gestures externalize internal mental representations
(McNeill, 1992). The gestures performed here reveal a way of knowing that flows
in the opposite direction, from the hand as a sensory actor alive to the ad hoc
sensations it encounters as it works with external materials, to theories about how
those sensations are relevant to the accomplishment of the activities in progress.
The gesture points not to some hidden image lodged within the speaker's brain,
but instead to the hand as an agent of experience in its own right, encountering
specific phenomena in the world within which it is working . It is true that Gina is
trying to make Billy aware of something she has experienced, (e.g., a mental
event), a sensation of texture. However, rather than constituting a private point of
origin for the gesture, that experience is embedded within and emerges from the
embodied activity that the gesture makes visible. Indeed her interlocutor's ability
to recognize and evaluate the sensation she is talking about requires coparticipation in that same activity. The frameworks that make possible mutual
understanding of this gesture and of the sensation it makes visible are not
constituted by preformulated representations, but through co-membership in a
relevant community of practice.
The way in which Gina learns from her ad hoc engagement with a relevant
environment (i.e., the fibers she is manipulating), within a situation where she is
able to talk about what she finds with a more competent practitioner, is
compatible with Vygotsky's notion of a zone of proximal development (Cole,
1985). The way in which the professor sometimes learns from his students and
changes aspects of the manufacturing process in light of what they discover is
also compatible with Engestrom's (1987) expansion of the zone of proximal
development beyond individuals to processes of change within organizations.
Using Diverse, Serendipitous Criteria to Constitute a Category

The geochemists thus use their bodies as media that experience the material being
worked with through a variety of different modalities, as one primary framework
for uncovering and shaping the organization of the process in which they are
engaged. Insofar as they remain open to what unanticipated sensations can reveal,
their thinking contains an essential element of situated improvisation that can be
incorporated into subsequent practice. Example 1 occurred approximately half an
hour after the conversation about the texture of the fibers in Example 2. Looking
again at Example 1, we find that, just after Billy counters Gina by noting that the
fibers still have a tinge, she provides further grounds for her assessment (note the
129
But that prefaces her turn in line 21) by drawing his attention to the current
texture of the fibers (lines 21-22):
Example la
16
Billy:
17
18
19
Gina:
20
21
22
23
24
25
26
Ye(h)a(h)h.
1.
But it's gettin a little- (points with stick)
Billy:
(0.9)
Billy:
the lumpy texture.

(1.2)
We:ll,
The temperature is doin we :ll.
Fig. 5.7. Example la
Assessment of the current state of the fiber is not made by the simple application
of a single category, the meaning of which is known in advance, but instead
emerges within a situated matrix of action encompassing multiple perspectives. In
this single sequence, color , texture, and temperature are all used to contest a
diagno sis. Some of the criteria being used to make the assessment are
pr ogressively changing (e.g., Gina's sense of what can count as an acceptable
black ), whereas others have only just been discovered by some of the participants
(e.g., Gina's recognition of the importance of texture). Rather than being
explici tly taught, many of these criteria are acquired through embodied
participation in the activity. The use of multiple criteria in this fashion cannot be
accounted for within the analytic framework of Berlin and Kay , who, like many
other cognitive anthropologists, carefully isolated for study a well-bounded
taxonomy restricted to a single perceptual domain . It is, however, consistent with
Wittgenstein's (1958) proposal that many categories are organized, not via
underlying essences, but rather in terms oi family resemblances: "a complicated
network of similarities overlapping and criss-crossing" (66) . When faced with
the practical task of locating a category, participants artfully make use of and
creatively discover a range of different methods drawn from a variety of sources.
The Social and Practical Constitution of Accountable Knowledge
Assessments of the fiber are lodged within a web of accountability encompassing

at least two different orders of phenomena:
130
Charles Goodwin
(1) The task itself and its material infrastructure (i.e., will the fiber actually
work?). Participants are not free to ignore the fit between the decisions they make
and the usefulness of the tools that are thus produced .
(2) What others will hold one responsible for.
There is thus a reciprocal relationship between the development of new tools
within a discipline and the development of socially organized structures of
perception by practitioners of that discipline. By working in concert with others
on relevant tasks, the body of the geochemist is transformed into a tool of the
trade. When Gina, one of Billy's graduate students, reached the shore of the
Amazon, she wanted to know approximately how much sea water it contained. To
determine this, she scooped a handful of the river water into her mouth and used a
chemist's sense of taste to evaluate its salinity.
Making judgments that others can not only recognize as appropriate but also
rely on constitutes being a competent practitioner. That knowledge can be gained
only through embodied practice, by working with the relevant materials and
having the judgments made in such circumstances evaluated by other competent
practitioners. Thus, by the end of the day (approximately one hour after Example
1), Gina was in a position to justify her decision to give a batch of fiber more time
by noting tinges in it (i.e., precisely the criterion to which Billy had earlier drawn
her attention) and to have Billy agree with that assessment after inspecting the
fiber himself:
Example 3
1
Billy:
2
Gina:
3
4
Billy:
6
7
8
9
10
11
12
13
Gina:
Billy:
Billy:
Gina:
Billy:
Billy:
Let's look at it.

Huh-ohI was givin it like two more mi:nutes here .
Because it's been in the solution
(Let's go ahead)
(11.5)
(Randy inspects thefiber)
There[is a couple a little bro[wn tinges
Yeah.
Good.
Uhhuh.
So I was just gonna leave it[until it's done.
Okay.
Okay .
Fig. 5.8. Example 3
The crucial importance of such embodied practice for being able to know what
black means within this activity system is well illustrated by my own inability to
make such a decision competently. I was physically present throughout the entire
fibermaking process . Several hours of that time were spent less than a yard from
131
the vat where these judgments were being made . Because I was videotaping the
process, I spent a lot of time looking at both the fiber and those working with it. I
was, however, outside the structures of accountability that linked Billy and Gina
to the fiber. I was not required to make decisions about when to withdraw the
fiber, did not have my color judgments evaluated by others, and did not feel the
fiber repetitively as it was undergoing the reaction .
Although I did not share with these geochemists a workable notion of jet-black
fiber so that they could rely on my judgment, this does not mean that their use of
the term was idiosyncratic. Instead, as Billy's careful work with Gina amply
demonstrates, the ability to make such an assessment was very much a social fact,
something that competent practitioners could hold each other accountable for and
that they were careful to teach to newcomers. The problem lies in assuming that
the language as a whole, as a cognitive structure abstracted from the messy details
of practice, is the place to study how categories encoding perceptual judgments
are organized. Instead , as argued by M.H. Goodwin (1990), the proper locus for
the analysis of culture, including the categories and practices through which it is
constituted, is not the society or disembodied language but situated activity
systems. By defining langue in the way that he did, Saussure found it impossible
to include human agency in the constitution of meaning or structure. For the
geochemists being analyzed here, determining what could count as black, so that
they could successfully perform the tasks in which they were engaged, was a
contingent, ongoing achievement, something that they had to work at to
accomplish (i.e., a prime locus for the analysis of human agency as a socially
embodied process).
Conclusion
The analysis in this chapter has focused on one of the central topics in
anthropological studies of cognition: semantic categories used to encode
perception of color. How does the view of cognition developed here differ from
that in the classic treatment of the subject by Berlin and Kay (1969)? Rather than
contradicting their findings, it uses as a point of departure a notion of what counts
as human cognition, and where it is located, that leads to the investigation of a
range of phenomena that were systematically excluded from the domain of
scrutiny they so ingeniously probed. It is useful to note explicitly some of these
differences.
132
Charles Goodwin
The Methodology of Berlin and Kay
For Berlin and Kay, the primary objects of study , the phenomena they are
attempting to uncover and describe, are universal structures divorced from the
messy contingencies of situated practice. These structures are located in two
related places : the human brain and the semantic systems of particular languages .
This theoretical agenda led to a specific methodology : First, the relevant units of
analysis are separate languages (e.g., English is compared with Japan ese and
Tzeltal). Second, within each language, analytical criteria (e.g., that the term be
monoleximic, that its application not be restricted to a narrow class of objects) are
used to locate a small set of basic color terms. One effect of this is to isolate color
distinctions as a self-contained semantic domain; criteria from other sensory
modalitiesl? and from task-relevant use of color vocabulary are eliminated . Third,
in line with accepted experimental procedure in psychology, a standard stimulus
was prepar ed: an array of 329 Munsell color chips mounted on stiff cardboard.
Fourth, basic color terms were elicited from speakers of different languages.
Finally, these native speakers were asked to locate on the chart both the best
example of a specific color term and the boundaries of that term (e.g., all chips
that the term could validly designate).
These procedures were used to systematically collect data from a wide sample
of languages from many different parts of the world. However, the use of color
terms in locally relevant , endogenous activities was never probed . All speakers
were performing exactly the same experimental task, and, with the exception of
the Tzeltal speakers, all of the speakers resided in the San Francisco Bay area.
The notion of a relevant community of competent practitioners was completely
irrelevant to Berlin and Kay's analysis; for many languages, only a single speaker
was used. Thus, in a very strong sense, the basic analytic unit in these studies was
a context-free component of langue, located in discrete languages that were
treated as relatively homogeneous rather than as a set of endogenous speech
communities .
Phenomena Made Available for Analysis by a Situated Activity System
In contrast to this, the basic unit of analysis for the study in this chapter was the
situated activity system. Investigation was focused on a group of geochemists
who had to determine when a fiber they were working with was jet black in order
to know when to stop a relevant chemical reaction . Black and white are the most
basic color names in Berlin and Kay's analysis. However, for the geochemists, jet
black (i.e., the most prototypical example of black) was not a context-free
10Although as Berlin and Kay (1969) acknowledge in the very fIrst paragraph of their
book, Conklin (1955) had drawn attention to the presence of non-colorimetric information
in Hanunoo color words.
133
universal color category that pointed automatically to a specific set of color

shades; instead, the term constituted a point of departure for a problematic
judgment to be artfully accomplished through the deployment of a collection of
systematic work practices. It might be argued that such practical work is
irrelevant to the constitution of color categories as abstract entities, for example,
mere performance details that have no bearing on the underlying system of
idealized competence where semantic categories should properly be analyzed.
However, limiting the scope of analysis in such a fashion arbitrarily excludes by
fiat a host of issues and phenomena that are central to the organization of human
cognition. Thus, Wittgenstein (1958) notes, "If language is to be a means of
communication there must be agreements not only in definitions but also (queer
as this may sound) in judgments" (242). The practices used by the geochemists
to assess the color of the material they are working with are central components
of the process through which the consequential judgments required for the proper
use of the category black are organized as systematic phenomena by a community
of relevant practitioners. Moreover, for Wittgenstein, the meaning of a name is
not its bearer (e.g., a range of shades named by a color term) but rather mastery of
the practices required to use that category competently within a relevant language
game (Baker & Hacker, 1980). Consistent with Hutchins' (1991, this volume)
analysis of how cognition does not reside exclusively in the individual brain but
is instead distributed throughout a relevant setting, focus on such practices opens
up for systematic study not simply mental representations but also external
cognitive artifacts, tools shaped by a prior history of engag ement in the tasks
being performed in the setting, the social distribution of knowledge, and the
processes of human interaction and apprenticeship (Lave & Wenger, 1991;
Rogo ff, 1990) through which relevant judgments are calibrated within an
endogenous workgroup. From such a perspective, analysis of a semantic category
is not restricted to discovering fixed, essential features common to all situations
of use . Instead, the usefulness of a category, as a tool capable of being
continuously appropriated to accomplish novel tasks, lies in its inherent,
contextually constituted, flexibility (Heritage, 1984). Once this is taken into
account, investigation of how such appropriation is accomplished emerges as a
topic of study in its own right.
The situated activity system, within which the color judgments being examined
here are lodged, provides organization for a range of phenomena. For example, as
has long been noted by conversation analysts (Sacks, 1972; Schegloff, 1992a), a
central issue posed in any analysis of human category use is that of relevance.
Any entity can be accurately categorized in an indefinite number of different
ways (e.g., a person can be described in terms of his or her weight, height, date of
birth, gender, religion). Issues such as how a category is to be defined or whether
it is being accurately applied are thus analytically subordinate to the prior
question of what organizes the selection of a particular category system (e.g., why
do these parties choose to attend to these fibers in terms of their color, instead of,
for example, their weight?). The answer is provided by the relevance of that
134
Charles Goodwin
specific category system to the activity they are engaged in: When the fibers
reach jet black, the reaction being monitored has to be terminated. The
encompassing activity thus provides a motivational framework within which
color discrimination becomes a relevant and expected thing to do .
Simultaneously, the structure of that activity sets parameters for what will count
as an acceptable solution to the task set by the relevant use of a color term (e.g.,
those shades of black that will produce usable fiber) . In turn the successful
accomplishment of that task leads to the deployment of a range of other practices
and tools (e.g., the invention of new category systems that highlight subtle
differences among similar colors, tools for extracting the material being examined
from a confusing background and positioning it for perception). The use of these
tools within the framework of the activity provides the participants with a visible
texture of intelligibility, enabling them to make inferences about what each other
is doing . By virtue of the encompassing activity, a heterogeneous collection of
very diverse phenomena - color categories, spray bottles, descriptions of animal
fur, sticks - is integrated into the accomplishment of a common cognitive task.
Proper use of these tools (what counts as "proper" is defined by the encompassing
activity) requires the mastery of socially organized embodied competencies (e.g.,
the ability to see, feel, smell, and taste as a geochemist). Rather than being private
perceptual structures lodged within the individual brain, such professional vision
(C. Goodwin, 1994) is socially organized by the tasks set by activities, such as the
one investigated in this chapter, and is something that members of the
communities responsible for doing these activities hold each accountable for if
one is to be recognized as a competent practitioner. Using the situated activity
system as a basic framework for analysis thus opens up to systematic study an
expanded view of human cognitive activity.
Using General Structures to Build Locally Relevant, Situated Action
In searching for cognitive universals, Berlin and Kay were reacting against
ethnographic particularism. It might be argued that focusing on the situated
activity system leads right back to particularism; for example, instead of
examining color distinctions common to all speakers of English or even to all
geochemists, this chapter has investigated a perceptual discrimination used in the
work practices of a small group of geochemists. However , other ethnographic
work has demonstrated that versions of many of the practices described here are
found generally. Consider, for example, highlighting: the way in which the
geochemists extracted the fiber they wanted to examine from a confusing
background so that its perceptual salience was emphasized, and it was positioned
for focused, intense scrutiny. Similar practices constitute part of the professional
craft of archaeologists who both lift objects from the dirt to examine them and
annotate the earth with lines drawn with such tools as trowels, brightly colored
flags, and bits of string in order to make dim features stand out from a confusing
135
background (for a detailed analy sis, see Goodwin, 1994; Lynch, 1988) . Such
highlighting is an instantiation in concrete practice of a most general cognitive
structure, the figure-ground relationship. As noted earlier, highlighting of
documents, so that information of relevance to a particular workgroup is made
salient, was one of the most general work practices found at the airport studied by
the Xerox PARe Workplace Project. By virtue of the way in which such
highlighting structures the perception of others, by reshaping a domain of scrutiny
so that some phenomena are made salient, whereas others fade into the
background, it can have strong rhetorical and political consequences. The lawyers
defending the policemen who severely beat an African American motorist,
Rodney King, highlighted the videotape of the beating through gesture, category
systems, and by drawing white lines around Mr. King's body in order to focus the
attention of the jury on "aggressive" body movements of Mr. King and away
from the actions of the policemen beating him (Goodwin, 1994) . In brief, human
cognitive activity characteristically occurs in environments that provide a very
complicated perceptual field. It is, therefore, not surprising that a general class of
cognitive practices consists of methods for structuring that perceptual field so that
phenomena relevant to the activity in which participants are engaged are made
salient, a process that simultaneously helps classify those phenomena (e.g., as an
archaeological feature rather than an irrelevant patch of color in the dirt, or as an
aggressive movement). However, such processes remain outside the domain of
what can be studied, if the notion of what counts as cognition is restricted to
structures hidden inside the brain . Moreover, a central component of this process
is the framework of relevance provided by the situated activity system within
which the act of highlighting is embedded. Practices such as highlighting
precisely link relevant features of the setting to the activity being performed in
that setting. When setting and activity are lost, these cognitive practices
disappear. The issue, therefore, is not particularism but rather access to a range of
basic cognitive processes that requ ire for their analysis detailed study of actual
work in endogenous settings (Scribner, 1984).
An excellent example of how scientists use highlighting to make complex
phenomena amenable to rigorous investigation can be found in the procedures
developed by Berlin and Kay to extract the domain of basic color terms from a
very confusing environment. To accomplish this, Berlin and Kay had to proceed
on several fronts, developing, on the one hand , semantic criteria that would apply
across languages to systematically separate a small set of basic terms from the
much larger color vocabulary found in each language and, on the other hand,
constructing a relevant perceptual target, the Munsell chart, that would enable
explicit comparison between languages. Assembling this package of procedures
so that basic color terms uncontaminated by extraneous phenomena could be
rigorously measured and compared was a major accomplishment that led to new
knowledge about how the brain and language work together to structure
perception. As such practices reveal, however, the context-free universals of
Berlin and Kay were themselves shaped into distinct kinds of entities and made
136
Charles Goodwin
available for a particular form of analysis , through an artfully crafted situated

activity system.
Even if one's goal is analysis of how the brain organizes categories, questions
can be raised about whether extracting phenomena from the rich context of
situated activity is the only or even best way to proceed. For example, recent
neurological research investigating how categories are organized in the brain
reveals that information from a range of different sensory modalities is integrated
in a concept. Thus, the concept of a cup includes not only a visual image but also
a sensation of weight and an association with liquids (Damasio & Damasio,
1992). Possibilities for investigating such a network of overlapping criteria were
eliminated in Berlin and Kay's analysis by the very procedures that extracted pure
color terms from their encompassing background. However, when the task of
color discrimination posed for the geochemists was investigated as a relevant
component of an encompassing activity (i.e., the geochemists were not assessing
color in a controlled environment as disinterested observers, but instead they were
working hard to figure out when to stop their reaction), it was found that they
creatively brought to bear whatever information their embodied ad hoc work with
the fibers made available to them (e.g., noting that, when the fibers reached the
proper color, they also had a distinctive texture). As an embodied practitioner,
Gina found the black that she was seeking not only with her eyes but also with
her hands. Such complementary use of the information provided by different
modes of sensation is consistent not only with recent work on how the brain
structures categories but also with Wittgenstein' s (1958) argument that categories
can be organized through a network of family resemblances rather than core
common properties. In brief the analysis of situated activity systems provides a
rich arena for the study of a diverse range of basic practices central to the
organization of human cognition.
Acknowledgments
I am deeply indebted to Dr. Willard S. Moore for making possible the research in
his lab and to Alessandro Duranti, Cathryn Houghton, Ed Hutchins, Candy
Goodwin, Elinor Ochs, Billy Moore, Clotilde Pontecorvo, and Curtis Renoe for
helpful and insightful comments on an earlier version of this analysis.
137
References
Baker, G. P., & Hacker, P. M. S. (1980). Wittgenste in: Understanding and meaning .
Chicago : The University of Chicago Press.
Berlin, B., & Kay, P. (1967). Universality and evolution of basic color terms. (Working
Paper #1). Berkeley, CA: Laboratory for Language Behavior Research.
Berlin, B., & Kay, P. (1969) . Basic color terms : Their universality and evolution.
Berkeley, CA: University of California Press.
Bernstein, B. (1972) . Social class, language and socialization. In P. P. Giglioli (Ed.),
Language and social context (pp. 157-179). New York: Penguin.
Brun-Cottan, F., Forbes, K., Goodwin, C., Goodwin, M. H., Jordan, B., Suchrnan, L., &
Trigg, R. (1991). The workplace project: Des igning for divers ity and change.
(Videotape) Palo Alto, CA: Xerox Research Center.
Bruner , J., Oliver, R. R., & Greenfield, P. M. (1966). Studies in cognitive growth. New
York: John Wiley and Sons.
Chaiklin, S., & Lave, J. (Eds.). (1993). Understanding practice : Perspectives on activity
Gild context. Cambridge, England: Cambridge University Press.
Cicourel, A. V. (1964). Method and measurement in sociology. New York: Free Press.
Cole, M. (1985). The zone of proximal development : Where culture and cognition create
each other. In J. Wertsch (Ed.), Culture, communication, and cognition: Vygotskian
perspectives (pp. 146-161). Cambridge, England: Cambridge University Press.
Collins, H. M. (1985) . Changing order: Replication and induction in scientific practice.
London: Sage.
Conklin, H. (1955). Hanunoo color categories. Southwestern Jounal of Anthropology, 11,
339-44.
Damasio, A. R., & Damasio, H. (1992 , September). Brain and language . Scientific
American, 89-95.
Enge strom, Y. (1987) . Learning by expanding : An activit y-theoretical approach to
developmental research. Helsinki: Orienta-Konsultit Oy.
Engestrom, Y., & Middleton, D. (Eds.). (in press). Cognition and communication at work.
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Goffman , E. (1961). Encounters: Two studies in the sociology of interaction. Indianapolis,
IN: Bobbs-Merrill .
Goodenough, W. H. (1956). Componential analysis and the study of meaning . Language,
32, 195-216.
Goodwin, C. (1994). Professional vision. American Anthropologist, 96(3), 606-633.
Goodwin , C. (in press). Transparent vision. In E. Ochs, E. A. Schegloff, & S. Thompson
(Eds.), Interaction and grammar. Cambridge, England: Cambridge University Press.
Goodwin , c., & Goodwin , M. H. (in press). Formulating planes : Seeing as a situated
activity . In Y. Engestrom & D. Middleton (Eds.), Cognition and communication at
work. Cambridge, England: Cambridge University Press.
138
Charles Goodwin
Goodwin , M. (in press) . Co-construction in girls ' hop scotch. Research on Language and
Social Interaction.
Goodwin, M. H. (1990) . He-said-she-said: Talk as social organization among black
children. Bloomington, IN: Indiana University Press .
Greenfield, P. M., & Bruner, J. S. (1966). Culture and cognitive growth . International
Journal ofPsychology, 1, 89-107.
Hacker , P. M. S. (1986). Insight and illusion: Themes in the philosphy ofWittgenstein (rev.
ed.). Oxford, England: Oxford University Press.
Hanks, W. F. (1990). Referential practice: Languag e and lived space among the Maya.
Chicago: University of Chicago Press.
Hank s, W . F. (1992) . The indexical ground of deictic reference. In A. Duranti & C.
Goodwin (Eds.), Rethinking context: Language as an interactive phenomenon (pp. 4376) . Cambridge, England: Cambridge University Press.
Heath, C. (in press) . Virtual looking : Spatial transformation and communicative
asymmetries. In P. Pelligrino (Ed .), Proceedings of the Colloquiem on the Semiotics of
Space. Geneva, Switzerland: University of Geneva .
Heath, C., & Luff, P. (in press). Convergent activities: Line control and passenger
information on the London underground. In Y. Engestr6m & D. Middleton (Eds.),
Cognition and Communication at Work. Cambridge, England: Cambridge University
Press.
Heidegger, M. (1962). Being and time (1. Macquarrie & E. Robinson, Trans.) . New York :
Harper & Row.
Heritage, J. (1984). Garfinkel and ethnomethodology. Cambridge, England : Polity Press.
Hutchins, E. (1991). The social organization of distributed cognition . In L. Resnick, 1. M.
Levine , & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 283-307) .
Washington, DC: American Psychological Association .
Hutchins , E. (1993). Learning to navigate . In S. Chaiklin & 1. Lave (Eds.), Understanding
practice: Perspectives on activity and context (pp . 35-63) . Cambridge, England:
Cambridge University Press.
Jefferson, G. (1987). Exposed and embedded corrections. In G. Button & J. R. E. Lee
(Eds .), Talk and Social organisation (pp. 86-100) . Clevedon, England: Multilingual
Matters Ltd.
Knorr-Cetina, K. (1981). The manufacture ofknowledge. Oxford, England: Pergamon.
Latour, B. (1987). Science in action: How to follow scientists and engineers through
society. Cambridge, MA: Harvard University Press .
Lave, J., & Wenger, E. (1991). Situated learning : Legitimate peripheral participation.
Lynch, M. (1985).Art and artefact in laboratory science. London: Routledge and Kegan
Paul.
Lynch, M. (1988). The externalized retina : Selection and mathematization in the visual
documentation of objects in the life sciences . Human studies, 11, 201-234.
Lynch, M. (1991). Method: Measurement - ordinary and scientific measurement as
ethnomethodological phenomena. In G. Button (Ed.), Ethnomethodology and the human
sciences (pp. 77-108) . Cambridge, England: Cambridge University Press .
139
Lynch, M., & Woolgar, S. (Eds.). (1988). Representation in scientific practice. Cambridge,
MA : MIT Press .
McNeill, D. (1992) . Hand &: mind: What gestures reveal about thought. Chicago:
University of Chicago Press.
Moore, W. S. (1976) . Sampling Ra<228> in the deep ocean . Deep-Sea Research, 23, 647651.
Parsons, T., & Shills, E. (1951) . Toward a general theory of action . Cambridge, MA:
Pickering, A. (Ed.) . (1992) . Science as practice and culture. Chicago: The University of
Chicago Press .
Polanyi , M. (1966) . The tacit dimension. Garden City, NY: Doubleday .
Rogoff, B. (1990) . Apprenticeship in thinking. New York : Oxford University Press .
Rogoff, B., & Lave , J. (Eds .). (1984) . Everyday cognition: Its development in social
context. Cambridge, MA: Harvard University Press.
Roskill , M. (Ed.). (1983) . The letters of Vincent van Gogh. London: Fontana Paperbacks.
Sacks, H. (1972) . An initial investigation of the usability of conversational materials for
doing sociology. In D. N. Sudnow (Ed.), Studies in Social Interact ion (pp. 31-74) . New
York : Free Press .
Sacks, H. (1989). On members ' measurement systems. (Edited by G. Jefferson from
unpublished lectures : 1966, 1967, 1970.) Research on Language and Social Interaction ,
22,45-60.
Sacks, H. (1992) . Lectures on conversation: Volume I. (Edited by G. Jefferson, with an
Introduction by E. A. Schegloff. ) Oxford , England : Basil Blackwell.
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974) . A simplest systematics for the
organization of turn-taking for conversation. Language, 50, 696-735 .
Saussure, F. d. (1959) . Course in general linguistics . (Edited by Charles Bally & Albert
Sechehaye, in collaboration with Albert Riedlinger; W. Baskin, Trans .) New York:
Philosophical Library .
Schegloff, E. A. (1972). Notes on a conversational practice: Formulating place. In D.
Sudnow (Ed.), Studies in social interaction (pp. 75-119) . New York: Free Press.
Schegloff, E. A. (1992a) . On talk and its institutional occasions . In P. Drew & J. Heritage
(Eds .), Talk at work: Interaction in institutional setting (pp. 101-134). Cambridge,
England : Cambridge University Press .
Schegloff, E. A. (1992b) . Repair after next turn: The last structurally provided defense of
intersubjectivity in conversation. American Journal ofSociology, 97(5), 1295-1345.
Scheper-Hughes, N. (1994) . Embodied knowledge: Thinking with the body in critical
medical anthropology. In R. Borofsky (Ed.), Assessing cultural anthropology (pp. 229239). New York: McGraw -Hill.
Scribner, S. (1984). Studying working intelligence. In B. Rogoff & J. Lave (Eds.),
Everyday cognition: Its development in social context (pp. 9-40) . Cambridge, MA :
Shapin , S. (1989) . The invisible technician . American Scientist, 77, 554-563.
Star, S. L. (1988 a). Introduction: The sociology of science and technology . Social
Problems, 35(3), 197-205 .
140
Charles Goodwin
Star, S. L. (1988b). The structure of ill-structured solutions: Heterogeneous problemsolving, boundary objects and distributed artificial intelligence. In M. Huhn s & L.
Gasser (Eds.), Distributed artificial intelligence 2 (pp. 37-54). Menlo Park, CA: Morgan
Kaufmann .
Suchman, L. (1992). Technologies of accountability: Of lizards and airplanes . In G. Button
(Ed.), Technology in working order: Studies of work, interaction and technology (pp.
113-126) . London : Routledge.
Suchman, L. A. (1987) . Plans and situated actions: The problem of human machine
Wittgenstein, L. (1958). Philosophical investigations. (Edited by G. E. M. Anscombe & R.
Rhees ; G. E. M. Anscombe, Trans ., 2nd ed.). Oxford , England : Blackwell.
Part Two
Negotiating Identities: The Construction of

Sociocognitive Communities
Chapter 6
Reasonable Uncertainties:
Parents' Talk About Caring for Children
with Chronic Renal Failure
David Middleton
Department of Human Sciences, Loughborough University, LEI I 3TU, United Kingdom
Abstract
Examples of parents' talk about their care of children with chronic renal failure are
discussed. These are taken from a series of parent support group meetings run over a
32-month period at a Regional Paediatric Renal Unit within the U.K. National Health
Service. All the par ents are directly involved in dialysizing their children at home
rather than in hospital settings. The success of these home-b ased treatments is
dependent on paren ts achieving a sophisticated level of paramedical skills and dietar y
knowledge. The analysis examines the discursive functioning of these support groups
in establishing and maintaining what parents argue as ordinary within the extraordinary circumstances of their lives. Parents tell of routine dilemmas they face in the
care of their children. The analysis demonstrates the way parents use these dilemmas
as a resource for arguing the basis of what, for them, constitutes reasonable outcomes
and circumstances in the care of their children- reasonable both as a form of jointly
accomplis hed reasoning over the salience of their experiences as parents, and
reasonable as acceptable in terms of the circumstances of the lives they live.
144
David Middleton
Introduction
This chapter examines the talk of people who have been projected into a
technologically supported environment by force of circumstance. They are
parents of children who have suffered various forms of kidney failure. To
maintain the viability of their children 's lives, either prior to kidney
transplantation or after any episodes of rejection, these parents are offered the
option to be trained to engage in complex medical and dietary procedures. This
involves the use of technologically sophisticated medical equipment, associated
consumables , drugs, and the maintenance of a carefully monitored dietary regime .
Examples of parents ' talk about their involvement in this demanding and timeconsuming health care are the focus of analysis . My aim is to examine how
parents argue the basis of what, for them, constitutes reasonable outcomes and
circumstances in the care of their children-reasonable both as a form of jointly
accomplished reasoning over the salience of their experiences as parents, and
reasonable as acceptable in terms of the circumstances of the lives they live.
The Data and Context of Parental Conversations
The examples of transcribed talk are taken from a series of parent support group
meetings run over a 32-month period up to the point of this analysis at a Regional
Paediatric Renal Unit within the U.K. National Health Service. All the parents are
directly involved in management of the dialysis of their children at home rather
than in hospital settings. The particular forms of peritoneal dialysis used allow for
the treatment of infants from birth onwards.
The success of these home-based treatments is dependent on parents
developing their competence in a range of paramedical skills and dietary skills .
To achieve this, systematic training in the necessary procedures is offered by
members of a Regional Paediatric Renal Team. The direct involvement of parents
in peritoneal dialysis , either as a precursor to transplantation or as a result of
kidney transplant rejection , locates medical care as part of domestic rather than
hospital-centered routines (see Collier & Watson, 1994).
One of the main benefits for the children is that their exposure to long periods
of hospitalization are reduced. However , the extra burden of care places heavy
demands 'on the parents , both in terms of the disruption to conventional routines
of domestic life, and in terms of the emotional stress of adapting to and living
with a child who has a life-threatening chronic illness. This increased burden of
care sets families apart from the normal run and organization of domestic and
community life and makes for a form of daily routine and experience that is by
definition out of the ordinary.
To help ameliorate this burden of care, particular attention is paid by the
Paediatric Renal Team to what is termed psychosocial support. Such support
Reasonable Uncertainties
145
addresses a nexus of social and psychological issues and is a recognized feature

of multidisciplinary care in paediatric nephrology. The parents' support group
discussed here is one of a range of psychosocial support initiatives developed as
part of the practice of a multidisciplinary team coordinating the regional services
for paediatric renal care. 1 This group was initiated to address the particular need
identified in discussions with parents for some form of exchange of experiences
that could help alleviate their expression of feelings of "social isolation."
Although parents came into contact with each other during the routines of clinic
attendance and periods of acute hospital care (e.g., to treat peritoneal infections in
their children), such occasions did not offer the best circumstances for detailed
discussions. In response to this, a series of day-long parent meetings held
approximately every 4 months was instigated to occur outside the routines of
clinical attendance and dialysis (see Argles, MacKinlay, Middleton, & Watson,
1994, for practical details).
Analytical Approach
The analytic approach adopted here to the parents' accounts of health care
experiences represents an inversion of the conventional methodological positions
in health-related psychological studies. The conventional approach in
psychosocial studies of health care is to take description as mapping in some
indirect way onto peoples' perceptions of the realities of health care experience
(cf. Abraham & Hampson, in press). In contrast, the focus of analytic inquiry here
is the to-be-established nature of peoples' experiences accomplished in their talk
about the care of their children. This research forms part of a body of work
concerned with developing a discursive psychology informed by conversational
analysis and discussions in ethnomethodology (see Atkinson & Heritage, 1984;
Billig, 1987; Billig et aI., 1988; Buttny, 1993; Edwards & Potter, 1992; Potter &
Wetherell , 1987; Sacks, 1992). Within such an approach, what and how people
communicate in talk and text about the world in which they live, and their
psychological states, are treated as categories and interpretative resources that
they themselves use to construct accounts of their lives and their accountabilities
within those lives.
This chapter examines the construction of health experiences through peoples'
descriptions and accounts of those experiences. Such accounts are not treated as
representing the facts of the matter but as functional constructions that handle
exigencies of the interactive situation of their involvement in health care. The
particular focus is on the way parents voice uncertainties about the circumstances
and outcomes of their children's care. Their care experiences are examined as
IThe Paediatric Renal Team incorporates within its core membership a permanent renal
social worker, a clinical psychologist, and a health psychologist, in addition to specialized
nursing staff, consultants, associated doctors, dietitians, and a hospital teacher.
146
David Middleton
socially intelligible through an analysis of the way they construct accounts that
are functional to the circumstances in which they take themselves to be in the
support groups (see also Middleton, 1996a).
Summary Aims
Overall, the analysis examines the discursive functioning of these groups in terms
of what parents argue as "ordinary" within the extraordinary circumstances of
their lives . The aim is to demonstrate how the narration of parental experience
affords a discursive context for exploring dilemmas and difficulties in the care of
their children , and how parents use these dilemmas as a resource for arguing the
basis of what, for them, constitutes reasonable circumstances in the shared care of
their children and health-related outcomes for their children.
The analysis takes up these issues in a number of ways. First, examples are
presented of how parents formulate their memberships of the support group .
These include the way they report the support group's value in terms of the way
participation enables sharing of common problems with other parents and the
development of common understandings concerning the processes and outcomes
in the type of treatment they are all involved in. Their working of group
membership is also examined in terms of the way they contrast their
circumstances and understandings with other people they come into contact with
(including medical experts, the general public, acquaintances, and family
relations). Second, the organization of the parents ' talk is examined for the way
collaborative completions accomplish and display affiliation between parents as
group members.
The analysis then moves to examine directly how parents' reasoning about the
circumstances and outcomes of their care of their children is organized in terms
of the uncertainties about what to expect. Uncertainties of circumstance are
illustrated with an example of talk about the practical dilemmas of mundane child
care practices in the unusual circumstances of child care that parents face as a
matter of routine experience . Uncertainities of outcome are illustrated in terms of
the sorts of expectations it is reasonable to assume might attend their children's
care. Finally, the analysis illustrates three recurring problems in parents' attempts
to grapple with uncertainties in presenting and re-presenting the circumstances
and outcomes in the care of their children: uncertainties in representing the
implications and consequences of transitions and changes in the regimes of care
children are subject to (e.g., transitions from home to hospital); uncertainties
about the circumstances families find themselves in now (e.g., are we doing it
right? how far can we deviate from prescribed care schedules?); and uncertainties
of trustworthiness (e.g., to whom can parents entrust the care of their children).
147
Formulating Membership and Belonging

Sequence 1 details parental views on the group at Session 7 (13 parents: 3
mothers and 5 couples). Toward the end of the day, the parents were requested by
the renal social worker (Anne) to express what, for them, appeared to be the
benefits of their participation in the group sessions.'
Parents' Views in Group Participation
Sequence 1 (Session 7)
1
Anne
Mrs . T
3
4
Mrs . T
Mr. 'IN
Mrs. C
Mrs . C
Mrs . TH
Mr. I
does this sort of group fulfill a need for parents

on the unit or do you think we could do without it
(general laughter)
yes it does fulfill a need
(general laughter)
I'll answer you Anne
the very fact that someone's corne back a second or
third time is an indication of the fact that it's
useful to them
I mean at clinic visits we don't really get a
chance to sit down and talk $
(general agreement)
$ and the kiddies are listening at least they're
out of earshot here and you can sort of sit down
and talk to other parents about What you really
feel about things because at clinic you're fraught
and you're rushed
also when when you're new and you corne to a
meeting like this you meet a lot of people at once
and you 're bound to bump into someone at cl i ni c
and you can say oh we'll see you at the meeting
etcetera and er er it's a good introduction I
think for new people especially to feel that er
they belong to something
when people find out about oh my child's got
kidney problems you tend to feel as though you're
an odd one out [nobody else knows what I'm going $
[(general agreement)
2 Transcription conventions are found at the end of the chapter.
148
David Middleton
Mr. I
10
Mrs. I
11
12
13
Mrs . ?
Mrs. I
Mr. 'IN
14
15
Mrs . I
Mrs . N
16
Mrs . I
17
Mrs. TH
18
19
20
Mrs . I
Mrs . TH
Mrs. IJ
21
22
Mrs. TH
Mrs. IJ
23
Mrs . IJ
24
Mrs. I
24
Mr. I
$ through and I've got all these problems and it's

so nice to be able to sit and talk about your
problems and realize that there's a lot of other
people going through the same problems and be able
to talk about your experiences and realize that
you know there's other people that can help you
talk to you about things that's er you're having
to put up with
even if you didn't know you're not looking useless
you know (.) we've all got the knowledge that our
child's going to go through dialysis and
transplant ( ... ) whereas if you talk to someone
outside here they just look at you .
(general agreement)
it's a whole different world isn't it really
i t is yeah
I had to pick up some new medicine or something
from the chemist a few weeks ago and I simply said
to him quite naturally oh can I give it to her
through the gastrostomy he hadn't a clue what I 'm
talking about that working in a chemist shop I
mean that makes your point I think
yeah
you've got people to talk about the same sorts of
things
yes so it's nice to be able to get together in a
chat really ( . . . ) and know thatyou're not boring anybody because they can relate
to it [all where $
[yes
$ somebody outside you know you can [
[tend to shut
off
[shut off
[shut up now yeah
they shut off to it don't they they don't know
what you're talking about so they don't listen
well even your own family shuts off
(general agreement)
they do (1) and they either want to know so you go
through it and then they understand or they just
don't want to know and they just don't want to
know and they just just go away
they leave it to you
149
Establishing Criteria of Value
The argument for the benefits of participation unfolds in formulations of a series

of warrants justifying the value of the group. Initially these are listed to include:
the self evidency of rates of attendance attesting to their value (4: Mr. TN: "the
very fact that someone's come back a second or third time is an indication of the
fact that it's useful to them"); identifying the groups as a time and place where
the pressure of medical clinic routines do not pertain, and where it is possible to
talk "to other parents about how you really feel about things" (6: Mrs. C); as a
context for establishing introductions and engendering feelings of belonging (7:
Mrs. SH: "it's a good introduction I think for new people especially to feel that er
they belong to something") . The bases for accomplishing membership and
belonging are then further elaborated as a topic of concern for participants. These
are worked through a series of arguments concerned with the problematic status
of their identities as parents of children with chronic renal failure.
Problems in Common
The changing status of becoming a parent of a child with "kidney problems" (8:
Mr . I) is constructed as entering a condition of being the "odd one out" (8). His
contribution takes up what it is to be a parent in these circumstances and
expresses an argument concerning the potential benefits of group attendance.
Mrs. TH has already enumerated the benefit of the group as a conversational
resource in the organization of contact beyond the meeting . Potential attendance
at the meeting provides a topic people can orient to in their incidental contacts, in
addition to being an actual location where "new people especially can feel that
they belong to something" (7). Mr. I's contribution does more than declare his
particular feelings, or at least make a claim that his feelings can be construed as
such. His contribution is constructed in such a way that it provides an opportunity
for other parents to align themselves with his conclusion and in so doing provides
the basis for corroborating his claim as being generally legitimate for parents in
such circumstances.
Mr. I presents the changing status of becoming a parent of a child with "kidney
problems" as entering a condition of being the "the odd one out" (8). His reported
speech of the news "oh my child's got kidney problems" is the premise for a
claim in general with others that "you tend to feel as though you are the odd one
out" (9). This drew agreement from others present. Its form as a piece of reported
speech is also interesting. In moving from a general claim in the second personal
plural ("you tend to feel") to something that can be heard as him reporting his
own experience ("nobody else knows what I'm going through and I've got all
these problems") makes it available as something others might care to go along
with without taking as a given that everyone shares the same experiences. The
first person reporting ironises his experiences and in so doing makes them
150
David Middleton
available as something others might care to accept as theirs. Such agreements

provide the basis for working up the affiliative consequences of group
participation in terms of defining problems in common. In other words , the
declaration of being the "odd one out" is part of a social argument that contributes
to the developi ng consensus within the group concerning what people accept as
being the common property of being such a parent.
Processes and Outcomes in Common
The value of membership and its affiliative consequences is elaborated further by

Mrs. H. ( 0) in terms of the specialty of being involved in the processes
associated with dialysis and transplantation. Again there is an interes ting
rhetorical structure in the affiliative argument. The conditionality of "even if you
didn't know you are not looking useless" (10), makes available the inference that
you might be displaying uselessness without necessarily accepting or rejecting
that you are "useless" in these circumstances . However, it is participatio n in the
group, "we've all got the knowledge that our child's going to go through dialysis
and transplant" (10), in contrast to contact with those beyond the group who are
taken as being incapable of comprehending the implications of being embroiled
in the dialysis and potential outcome of transplantation ("whereas if you talk to
someone outside they just look at you"; 10), that confirms the affiliative
consequences of group membership. By implication, it is the talk with insiders
that works mutual recognition concerning equivalencies of experience. This
argument is granted general assent by others and summarized by an
unrecognizable speaker in the rhetorically structured invitation for confirmation :
"it's a whole different world isn't it really" (11). This is not simply a matter of
declaring feelings of living in a different world. It is rather that living in a whole
different world is part of what the conversational action in the group achieves.
Problems in common and the affiliative consequences of group membership are
elaborated by Mrs . H.(10) in terms of the specialty of being involved in the
processes associated with dialysis and transplantation.
Contrasts With Other People
The group then proceeds to identify further aspects of what it is to occupy such a
unique status as parents handling the routines of dialysis and medical care, a
status that marks them as different from others in ways that only insiders can
understand. A whole range of social contacts with others (be they medical
professionals, other folks in general, and even family members) are argued as
challenging the assumptions on which contact with such people is usually
conducted. Theirs is a different world for all sorts of reasons.
151
Contrasts in Their Relations With Medical Experts
Even professionals with medical knowledge can fail to understand the nature and
normality for them of their expertise. Me. TN (13) presented a short account of
his contact with a chemi st (pharmacist). The specific details concerning the
reason he needed to visit the chemist are imprecise : "I had to pick up some new
medicine or something" (13). This serves to emphasize the ordinariness of his
task. As part of such a mundane commercia l transaction, the ordinariness of his
position in having access to the use of the gastrostomy button' was expressed as
"simply said to him quite naturally" (13). The claim that "he hadn 't a clue" (13)
about what he was referring to despite "working in a chemist shop" (13) is
presented as the evidence for the specialty of the world they as parents inhabit.
Even professionals associated with medical care do not know about the most
basic procedures they as parents have to deal with routinely. What is ordinary to
them is extraordinary and unknown to others, even those who have medically
related expertise. These parents inhabit a world whose normality is defined in
terms of uncommon and irregular circumstances .
Contrasts in Their Relations With Other People in General
Again , in representing the group as a place where you can exercise very basic
forms of social engagement, "yes so nice to be able to get together for a chat
really" (16: Mrs. I), while at the same time knowing that the person you are
talking to "can rel ate" (Mrs. TH : 17) to what you are talking about, the
functionality of group attendance is contrasted with a social contact with others
from the outside who "tend to shut off' (Mrs. 11: 20). (The organization of this
particular sequence will be examined in more detail shortly.)
Contrasts in Their Relations With Family Members.
Even in their relations with family members, uncertainties contingent on their

special circumstances come into play. There is no guarantee how people close to
you will react. Mrs . 1 (24) identified a polarized reaction on the part of relatives,
one of detailed engagement contrasted with one of apparent indifference and
detachment: "they either want to know so you go through it and then they
understand or they just don't want to know and they just don't want to know and
3To overcome medical and behavior problem s associated with poor appetite, feeding
supplements and medication can be achieved via a gastrostomy button. This is a small
silicone rubber valve that enables direct introduction of feeds and medication into the
stomach without interfering with normal feeding and drinking. The use of such buttons is
the preferred method at the Regional Centre for Paediatric Nephrology where the children
of these parents attended (Collier & Watson, 1994).
David Middleton
152
they just just go away" (24). Again, dilemmas in the extraordinariness of their
circumstances provide a resource for identifying criteria for accomplishing
commonalities of experience in terms of what is reasonable to expect.
Organization of the Talk

Collaborative Completions in Accomplishing Affiliation
Apart from the argument concerning what the group affords and the joint
construction of warrants for such arguments in terms of insider/outsider
boundaries, there are some interesting features in the way the organization of
their talk accomplishes and displays affiliation among the parents in their group
participation. In Sequence 1 (Sections 15-21), the parents do not just "talk about
the same sorts of things" (Mrs. N: 15); they mutually complete and repair each
other's talk. Their interruptions do not have to do with establishing alternative
outomes but with contributing to build up the same project.
Collaborative completions in accomplishing affiliation

Sequence 1 (Sections 15-21)
15
Mrs . N
16
Mrs . I
17
Mrs . TH
18
19
20
Mrs . I
Mrs. TH
Mrs. IJ
21
22
Mrs. TH
Mrs. IJ
you 've got people to talk about the same sorts of

things
yes so it's nice to be able to get together in a
chat really ( . .. . j and know thatyou 're not boring anybody because they can relate
to it [all where$
[yes
$ somebody outside you know you can [
[tend to shut
off
[shut off
[shut up now yeah
they shut off to it don't they they don't know
what you 're talking about so they don't listen
The projected termination of Mrs. I's turn (concluding what she may, in fact
"know" about the consequences of group participation) is preempted by Mrs. TH:
"you're not boring anybody because they can relate to it" (17). This warrant for
chatting , knowing that you share like concerns, is ratified by Mrs. I's expression
of agreement. Mrs. TH's continued comparison with the "outside" is, in turn,
153
interrupted by another speaker, Mrs. Il, who offers a reason it is difficult to talk to
those outside: because they "tend to shut off' (20). This, in turn, is both repaired
and ratified by Mrs. TH: "shut up now yeah"(21). However, her repair is qualified
in Mrs. IJ's next turn, where she qualified the implication of "shut off' as not
listening because "they don't know what you're talking about so they don't
listen" (22). The conclusion to be drawn from no response is that people have no
common experience to share with those in the know.
Such close coupled joint reasoning accomplishes a mutuality of common-sense
understandings of the situation. These interruptions do not signal that speakers
have alternative communicative projects to pursue. They do not interrupt to
propose alternative conversational consequences. Their completions of what
someone else has begun display sharing the same description of the world rather
than an alternative one (Perakyla, 1994). Lerner (1993) has examined how such
collaborative completions are part of the way in which people make relevant to
their action their memberships of groups. The collaborative completions in this
sequence mutually display the parents' involvement in common circumstances.
This no longer has to be argued for. In their talk, they practically constitute a
community of experience, what it is that they hold in common.
Reasonable Circumstances and Outcomes

The group is formulated as a context in which defaults concerning what can be
expected in the parenting and care of their children can be accomplished. This
discursive accomplishment in terms of what it is to be affiliated with this
extraordinary regime of care is a significant achievement of group participation.
The parents' accounts of their experiences were constructed with respect to
uncertainties about what they might or might not expect to face as parents . Such
uncertainties are expressed both in terms of the circumstances they find
themselves in and in terms of the potential outcomes for themselves and their
children . For example, even the most mundane practices associated with child
rearing challenge the conventional wisdom of such circumstances. In the
following sequence, a father is describing his experience in changing the nappy
(diaper) of his then very young child.
Practical dilemmas of doing ordinary caring practices in extraordinary

circumstances.
1
Mr . I
(did he) manage any kidney function?

urn (.) [yeah
154
David Middleton
3 Mrs. I
4 Mr. I
5
6
7
8
[yeah he was right 'til he was eight and a half

we got him home at 5 or 6 weeks . and he's go :ne right
til then. for the first two years of his life were in
and out of hospital. er. various operations erm. they
brought tubes out to the side . and er . >i t were bad
enough putting a nappy on for urine coming out of one
place but trying to put a nappy on coming out there .
coming out both sides at the same time . err bit of an
experie : :nc e ehh
??
[ehh:hh[
Mrs . TH?
[mm
Mrs . TH I can believe it >y eah<
Mr. I
and er . they reinplanted the tubes into his bladder
and they blocked these off ( . . . ) he's been OK . he had
to go actually- we'd just finished all the operations
( ) his kidney ( . .. ) and er . he had to go in for a
double hernia. well hernia hydrocele and erm. since he
was ~ years old he's not been too bad . and we've
gone right up to (1) this last year or so ( ... )
The "tellability" of this experience is worked in terms of the practical dilemmas

of putting on a nappy in difficult circumstances. It is the doing of the ordinary
within the extraordinary: in other words, what it is to do this type of extranormality with a child in this condition, the practical dilemma of coping with a
baby that leaks urine not just in "one place" but "out of the sides" (MR . I: 4) as
well . The surgical insertion of the connection points necessary for peritoneal
dialysis had led to unforeseen complications in the routines of child care that are
demanding enough : "it were bad enough putting a nappy on for urine coming out
of one place but trying to put a nappy on coming out of there . coming out both
sides at the same time . er[ bit of an experie: :nce ehh" (4). The laughter does
interesting work in this sequence. Jefferson (1984) has argued that laughter in
trouble talk can mark speakers as "troubles resistive" and that joint laughter
establishes both "time out" and "buffers" between topics. Laughter is certainly a
key feature throughout all the sessions recorded within the parent groups.
However, in this particular instance, although this father's commentary on the
consequences of tube insertion, for what he presents as the less desirable features
of ordinary parenting, initially engenders laughter from someone else ("ehh:hh"
(5)), and he also offers the basis for the irony of the situation he found himself in
through his own laughter ("bit of an experie: :nce ehh"), it did not in that instance
lead to further joint laughter. Rather than leading to time out or acting as a buffer
to the next topic, it enjoined ratification of his version of his feelings about his
experiences of care as appropriate, with agreement from Mrs . TH: "mm (6); I can
believe it >yeah<" (7). This parent's experience is ratified and normalized in this
context as a believable experience by another parent. Such ratifications clearly
155
work a membership that is accountable through reasoning about what was

unreasonable about the circumstances in which they found themselves.
Similarly, it is possible to find examples of reasoning concerning what to
expect as possible outcomes of their experiences of care.
Uncertainties of expectation
Mrs . TIS but they were very urn . (.) 0 : h serious when she
was in urn baby unit I expected her to be enn well
have loads of things round her she was actually in
an incubator for about 24 hours. but having it
been three weeks I expected her to be in there
three months you know it was a bit of a shock when
they said she could corne horne urn but they were
very ..ill:.:.ious I mean they were ( . .. )
2?
huh huh huh
3
Mrs . TIS yes so serious you thought you know is she going
to pop off tomorrow or the day a : :fter uh huh you
know you didn't know quite how how to feel urn but
I got used to how doctors talk so huh huh huh
4
Mrs . ST when we brought Jeffrey in we were prepared for
quite a long stay so surprised when he said we
could go horne you know
5
Mrs . TIS I suppose it's nice if you are prepared for the
worst but things aren't that bad then. (.) you
know
6
Mrs . ST Doctor Smith said to us how long do you think
you'll be in for I said oh about 'til Christmas he
says no you 'll be back before then before
Christmas and it was such a shock when he said you
can horne you know
7
Mrs. I
we found that when we carne ' caus e we carne on the
Monday afternoon at tea time and we'd been in er
we'd had 5 weeks? 6 weeks in {City 2} in and out
of hospital . staying for a fortnight at a time and
we carne here on the Monday and we thought <I mean
to say I brought clothes for a fortni : :ght huh hu
$
8
Mrs . TIS hah hah hah
156
9
10
11
12
13
14
15
16
17
David Middleton
Mrs. I
$ packed me chucked things in I thought this was

it we're down there for months you know> and er
especially because {City 2} said it would have to
be haemodialysis i f {City 1} it would be
peritoneum. and we just come prepared for a month
didn't we? at least a ffiQilth. and got here on the
Monday teatime and on the Thursday after dinner
time Dr Smith said we can go ho: :me huh huh huh .
you wha: : t? you can go home come back on Saturday
mor :ni : :ng huh huhuh you know urr:: you look at
them like a fish thats dried out of water don't
you? got ( .) . you know (.) [couldn't believeMrs. TIS
[its a nice surprise if
you're prepared to spend longer here it is a nice
surprise if you find you can go home
Mrs. I
and we go home and you phone up me sister and you
say it's me I'm at home . what you doing at home.
how long have you got? well we've got Jeffrey at
home as well . you wha: : t huh huh huh you know
even nicer than that we'd actually booked to go
Mr. I
over to Jersey on the saturday. {City 2} had said
we'll give you all the letters for your insurance
and everything. you can forget your holiday and
all the rest of it. transferred us down here on
the Monday er:m with us having lots of problems
with Jeffrey draining out and er within a few
hours of us being down here they whipped his tube
out he was of- off dialysis all together and on a
low protein diet and er. wi thin the week we'd got
clearance to go to Jersey for us holiday
Mrs . I
couldn't believe it
gosh
?
( . .. ) having us holida : :y huh huh. (.) [without the
Mr . I
hassle of all the dialysis
[yes
Mrs. TI
[yes
?
These contributions developed out of Mrs. TIS's preceding account of the

diagnosis of her daughter's renal difficulties during her pregnancy. It turns
initially on the trauma associated with learning for the first time the diagnosis of
her child's condition. Her account of her uncertainties concerning what to expect
is built around her reading of the seriousness of that condition, the immediate
outcome being one of long stays in hospital: "I expected her to be in there three
months you know it was a bit of a shock when they said she could come home urn
but they were very ~ious I mean they were" (1).
157
We also see voiced the uncertainty of outcome in terms of her daughter's

apparent tenuous hold on life: "yes so serious you thought you know is she going
to pop off tomorrow or the day a::fter uh huh you know you didn't know quite
how how to feel urn but 1 got used to how doctors talk so huh huh huh" (3).
She claims competence now as a skilled reader, not only of the implications for
her child's survival, but also as someone capable of setting into context the
vocabulary of feedback in "how doctors talk." Her laughter potentially displaces,
misaligns, her utterance with her current position in these affairs, where she can
now claim to have shifted to a more sanguine view on what the implications of
her child's condition are . She claims competence in understanding the
implications of the way doctors give information in a "serious" manner and
ironically displays her previous view on the fragilities of her child's grasp on life.
Uncertainties of what people expected then became the currency for the
ensuing dialogue. The tellings, their news worthiness in the context of the group,
develop the topicalization of uncertainties in what to expect in relation to going
home. Other members' entitlements to equivalent experiences of uncertainties of
outcome are worked through their identifying equivalencies of place in the receipt
of news about the discharge of their children from hospital (see Sacks, 1992Lecture 4 , Spring 1970, for a detailed discussion on the management and
organization of second story telling). But more than this, there is a normativeness
being worked about the sort of experiences they have all had . They all expected
long stays and offer details in their accounts that ratify their warrants for those
expectations.
Mrs. ST does more than merely report an equivalent experience: "when we
brought Jeffrey in we were prepared for quite a long stay so surprised when he
said we could go home you know" (4). She elaborates her account in terms of the
reported rhetorical construction of the consultant's interrogation of their
expectations: "Doctor Smith said to us how long do you think you'll be in for 1
said oh about 'til Christmas he says no you'll be back before then before
Christmas and it was such a shock when he said you can home you know" (6).
She claims an equivalence of surprise and experience through her justification
of being in an equivalent position. And so it builds up. Mrs. 1 (7; 9) reports a
similar story of surprise:
"we found that when we came' cause we came on the Monday afternoon at
tea time and we'd been in er we'd had 5 weeks? 6 weeks in {City 2} in and
out of hospital. staying for a fortnight at a time and we came here on the
Monday and we thought <I mean to say 1brought clothes for a fortni::ght huh
hu packed me chucked things in 1 thought this was it we're down there for
months you know> [details edited] and we just come prepared for a Illiillth
didn't we? at least a.IIl.QD.1h. and got here on the Monday teatime and on the
Thursday after dinner time Dr Smith said we can go ho::me huh huh huh. you
wha::t? you can go home come back on Saturday mor:ni::ng huh huhuh you
158
David Middleton
know urr:: you look at them like a fish thats dried out of water don't you? got
(.). you know (.) couldn't believe-" (7; 9)
This is no mere repetition of the previous speakers' reasons and experience. She
deploys evidence in terms of her preparations and her stunned reactions as
detailing the equivalence of her experience. But "even nicer than that, "Mr. I.
(12) caps the telling through the unexpected outcome of being able to go on
holiday:
"even nicer than that we'd actually booked to go over to Jersey on the
Saturday. {City 2} had said we'll give you all the letters for your insurance
and everything. you can forget your holiday and all the rest of it. transferred
us down here on the Monday er:m with us having lots of problems with
Jeffrey draining out and er within a few hours of us being down here they
whipped his tube out he was of- off dialysis all together and on a low protein
diet and er. within the week we'd got clearance to go to Jersey for us
holiday" (12)
Detailed information about his place in sequences of events is offered in support

of his claim to having shared equivalent feelings concerning the unexpected
consequences in the treatment of their child, unexpectedly being allowed to go
home. These "second stories" accomplish more than the claim for equivalencies
in experience; they show those experiences to be equivalent (cf. Sacks, (1992);
see also Buchanan & Middleton, 1995). Intriguingly, the dilemmas of uncertainty
and the working of equivalent reactions of experience through warrantable
evidence for having been in an equivalent position highlight the extraordinariness of the normalities of the lives they lead. For it is in coming "off
dialysis" with "tubes whipped out" (12) that we see possibilities for being just
ordinary, that is a family on holiday away from all the paraphernalia of life
support that is in present reality a mundane feature of their lives. There is always
a present hope in the dialogue of parents in the group of being one day truly
ordinary.
"Reasonable" Dilemmas of Uncertainty

We can now turn in more detail to the use of these uncertainties as a resource in
the parents' arguments concerning what, for them, constitutes reasonable
circumstances and outcomes in the care of their children. The analysis illustrates
three of the recurring problems in parents' attempts to grapple with uncertainties
in presenting and re-presenting the circumstances and outcomes in the care of
159
their children that were identified in the seven sessions that are the subject of
analysis in this chapter. They involved uncertainties in representing the
implications and consequences of transitions and changes in the regimes of care
children are subject to; uncertainties about the circumstances families find
themselves in now; and uncertainties of trustworthiness (e.g., to whom can
parents entrust the care of their children).
Uncertainties in Representing Transitions and Changes

Transitions from hospital to home. What to expect?
1 Mr. TN we were sort of based here in the hospital for a long

long time until after one or two false we finally took
her and the machine home three weeks ago?
2 Mrs . 'IN ey e s that's rightO
3 Mr. 'IN and er touching wood <and I'm saying this> but living
with the dialysis at home has been nothing like as
awful as : as we thought it was going to be we had urn
we had a sort urn t he prospect of taking the machine
home I think four or five weeks before it actually
happened . urn but they decided she was doing better
than they thought and she: was coming back to hospital
two nights a week f or quite a f ew weeks Which I know
you know we ( . . .. )
4 Mrs . 'IN yeh
5 Mr. 'IN almost a frie : endly thing wasn't it sort of meeting up
on these occasions and erm we did have one days
training with Anne but then they- changed their minds
about sending us home in that way and when we got
around to training the second time it was nothing like
as
6 Mrs. 'IN y ehO
7 Mr. 'IN awful at least for me and we'd obv i ous l y picked up a
lot in hospital we hadn't realized how much we'd
learned. and urn (1) I guess. there was there was an
awful lot to worry about . to start with . and er it was
unfortunate that t he first week or so with Daisy at
home she had a foul cold <sh e spends an lot of her
time having dreadful colds it seems> and she did wake
up an awful lot during the night and of course that
left us wondering $
8 Mrs . 'IN y e hO
Sequence 5 is taken from the initial part of Mr. and Mrs. TN's account of their
initial experiences in dealing with the first few weeks of their preparation for
160
David Middleton
home-based peritoneal dialysis . Mr. TN voices his previous fears in anticipating

the transition from dependence on hospital care to the interdependencies of care
that home-based dialysis appeared to entail. At the time of recording, they were
three weeks into home dialysis . His conclusion that "dialysis at home has been
nothing like as awful as: as we thought it was going to be" (3) is premised on
previously anticipated fears in coping with the transition from hospital to home
care. He mitigates his tempting of fate-"touching wood" (3). His quickly added
qualification, "<and I'm saying this>" foots (Goffman , 1981) his utterance in
such a way that the responsibility for this conclusion is made available as his. His
argument concerning the uncertainties associated with the transition to the home
is ratified by his wife, with continued qualification by him that, although this was
jointly experienced as a couple, he could only really claim a view insofar as it
related to him: "the second time it was nothing like as awful at least for me" (5).
However, having voiced his argument that the anticipated transition had not been
as difficult as he or they had been anticipating (his wife interjects agreement), and
provided reasons for why that should have been the case "we'd obviously picked
up a lot in hospital we hadn't realized how much we'd learned"(7), the transition
home had not been totally without things "to worry about" (7). There remained a
large degree of uncertainty about how to evaluate the consequence of what, for
most parents, would have been an irritation rather than portentous of something
more serious (their child having a cold and waking during the night). They were
left "wondering" (7). Here fears are both expressed and allayed in the arguments
the parents themselves deploy in the representation of their uncertainties.
Uncertainties About the Circumstances They Find Themselves in Now

The practicalities of the home-based regime continually face parents with a range
of practical dilemmas. Although their management of the equipment and the
associated techniques of feeding are carefully introduced and monitored by the
Renal Team, ultimately the parents are responsible for the delivery of the care
when they are at home. They have to become experts in day-to-day management
of the regime of care. Their own evaluation of their performance was a topic of
concern, and it introduced issues concerning the extent to which they were doing
things correctly, whether other parents had shared the same problems, and
whether and to what extent it was possible to deviate from the guidelines laid
down in their training.
Are we doing it right?

l?
what time does she go on
2
Mrs. T about seven half past seven and $
3?
right
161
Mrs. 'IN $ then she comes off at half past seven "half seven"
Mr. 'IN at least with the dialysis machine we find that we
can jygge when to put it on and when we want to get
it off but we've compared notes with other people
about using the overnight fe :ed $
6
?
tuh
7
Mr. 'IN $ and i t seems to have a mind of it's own $
8
?
[yeh
9?
[of course
10??
[huh huh huh
11 Mr. 'IN $ and it's very difficult
5
In the preceeding transcribed sequence, others are recruited into a ratification of

Mr. and Mrs. 'IN's problem in handling the overnight machine delivering feed via
the gastrostomy button . We see articulated a default understanding that these
machines are far from mere technical objects: "it seems to have a mind of its own
and it is very difficult" (Mr. 'IN 7; 11). This is immediately corroborated by other
people. It is ratified and ironically commented on as the nature of such things and
engenders laughter: "yeh" (8); of course (9); "huh huh huh" (10). What it is to
"judge" the probity of their actions is both reported as being based on "compared
notes with other people about using overnight fe:ed" (5) and achieved as such in
that reporting.
How far can we deviate?

1
2
3
4
Mrs L
you were talking about gastrostomy ( . . . )$
Mr. 'IN ermn ermn
Mrs. L $ do you find it's not finished when you want to take
her off in the morning
Mrs. 'IN yeah but on the quiet occasion we want to get up we
just 0 ( . . ) 0 don 't tell Jean hu huh
?? ?
hah huh huh
Mrs. 'IN ( ... ) seven hundred or so she's on seven hundred a
night and it says seven hundred and twenty nine and
you're only putting about seven hundred in so it's
not as if ( .. . ) but she eats a bit of it
?
( ... )
Mrs .?
it takes a while to just get it right
9
Mr. 'IN nm nm
10?
it depends on what bag you use as well
8
162
David Middleton
Admission of deviation is also used as a resource in developing a common basis

for accountabilities within the group. This is illustrated in Sequence 7. Mrs . L
offers the basis of mutual commentary in her request for evaluation concerning
the status of residues of unadministered feed via the gastrostomy button: "do you
find it's not finished when you want to take her off in the morning" (3). In Mrs.
L's move to ground this topic in the concern of Mrs. TN's previous talk, "you
were talking about gastrostomy" 0), and in the location of her request within the
experience of Mrs. TN, Mrs. L implicitly acknowledges and makes available for
others the potential generality that feed residue might remain in the morning. Mrs .
TN's admission that this is the case, her marking of it as a deviation "yeah but on
the quiet" (4) , and her whispered details (which are so quiet as to be
untranscribable) topicalize the dilemmas of deviation with reference to the
authority of the dietitian on the renal team. The problem of how much authority
and responsibility parents may take is acknowledged and recognized in the
laughter of others (5). Although they are deemed expert enough to handle the
regime of care, the basis for flexibly managing contingent problems, one of the
hallmarks of expert in contrast to novice practice, remains problematic. Their
examination of deviancy, in this case from the prescriptive norms of feeding laid
down by the dietitian, provides a resource for establishing the norms of their own
expert practice, although in this particular case, information about the particular
circumstances of Mrs . TN's child is offered in mitigation of deviancy from the
feed requirements: "but she eats a bit" (6). In addition, other candidates for
mitigating schedule deviation are made available: "it takes a while to just get it
right" (8); and variation in the type of equipment can lead to other variations: "it
depends on what bag you use as well" (10) . In all this, the contingent
uncertainties and accounts of deviations from procedures are the resource for
establishing common understandings of what it is to care.
Uncertainties of Trustworthiness
In addition to the working of extraordinary normalities, there are also discussions
about uncertainties of trustworthiness: whom from the outside can they trust, be
they medical, lay persons, acquaintances, or relatives:
Uncertainties of trustworthiness
1
Mr . 'IN
2
3
4
?
???
?
we're feeling particularly good about that because

we've had our first babysitting [last night we
actually got out for an evening
[(
hah hah hah
I think you need tha :t
. .. )
5
6
7
8
9
10
11
12
13
14
15
16
17
18
163
yeh yes it was nice

was Daisy on the machine last night
yes we thought if we have a nurse she 's go: i : ng to
be dia:aly :sed
? ? ? ??
hah hah hah hah hah
Mrs . 'IN so we changed the night off to Friday's
???
he huh huh
Mrs. 'IN but we thought if something goes wrong at least up
to eleven or midnight there is actually a nurse
here
Mrs . l i t ' s nice to be able to go out and think well i f
anything goes wrong they can pick the phone up $
yeh
Mr .?
Mrs . ?
yeh
Mrs. I
$ and phone somebody and talk to somebody at the
other end that- they know that she knows
Mrs . 'IN yeh althQiliIh that's quite right as she was filling
ever so slowly for the first fill and I was sitting
in the theatre thinking ah has she got peritonitis
is there a blocka: :ge $
Mrs. I
huh huh huh
Mrs. 'IN $ and you know thinking are they going to get in
touch wi th us you know as h e was okay"
Mrs . 'IN
Mrs . I
Mrs . 'IN
Within any normal pattern of child-rearing, issues of whom to trust your children
with is a typical concern. For parents in these circumstances, the basis for giving
over the responsibility of the care of their child to others is very problematic. In
the preceding sequence, the parents are describing their first use of the babysitting service that utilizes people skilled in dialysis. This afforded both the
possibility of going out and the possibility of handing over the responsibility for
dialysis for one night. The description of taking advantage of such an opportunity
was met with considerable laughter by the group (8). They clearly recognized the
trade-off in responsibilities being reported. In addition the issue is one of
knowing that those with whom you leave your child are known to have expertise
in the special regime of care. However, it is not enough for you to know that they
might be competent. Others at the hospital who might be called on to give advice
in an emergency need to know that the person left in charge is competent: "it's
nice to be able to go out and think well if anything goes wrong they can pick the
phone up and. phone somebody and talk to somebody at the other end that- they
know that she knows" (Mrs . I: 12; 15). Notwithstanding this, in the particular
circumstance being reported by Mr. and Mrs. TN, their intimate knowledge of the
child 's recent reactions and difficulties still places them in a dilemma concerning
whether they should or should not have relinquished temporary responsibility: "I
was sitting in the theatre thinking ah she has got peritonitis in there a blocka::ge
[huh huh huh ] and you know thinking are they going to get in touch with us you
164
David Middleton
know she was okay"." The uncertainty is never eradicable; it might be

provisionally dispelled ; however, it remains a feature of her reporting of what it is
to become part of other normalities (e.g., going out and socializing) that are the
givens of less extraordinary circumstances of child care. Again, the reporting of
their experiences of care are organized in terms of uncertainties in representing
what it is reasonable to expect and of the working out of the conditions under
which you might or might not trust others with the care of your child . Common
understandings of what can be taken to be the basis of trusting others are both the
resource and the product of the talk in these groups.
Conclusion
This chapter has aimed to demonstrate how joint ratifications of entitlements to
experience accomplish commonalities concerning the extraordinary
circumstances of home-based peritoneal dialysis of children with renal failure.
The groups do more than consolidate some parent support network. The
communicative action in the groups works consensus about the ordinariness of
the extraordinary circumstances within which parents live their lives. The
descriptions of experience become organizable in terms of a consensus
concerning the legitimacy of that experience as commonly understood. Common
ownerships are worked." The point is that what it is to be ordinary within these
extraordinary circumstances is something that is the accomplishment of the group
discussions: A normativeness is accomplished that has no basis beyond the
parents' accountabilities as carers in these extraordinary circumstances.
The talk that makes up those discussions is organized through accounts that
express dilemmas that handle uncertainties about what is reasonable to expect as
parents who find themselves coping with the rigors of this form of shared care .
This is not to claim that the uncertainties outlined and illustrated exhaust the
possible dilemmas that parents may express in the future or those they might have
expressed in the past. Indeed, that is precisely where the generative potential of
this type of communicative action resides. Uncertainty concerning the nature of
their care of their children is a recurring feature of their talk. Resolving such
uncertainty is always provisional; alternative versions and interpretations are
continually made available (see also Middleton, in 1996b; Middleton & Curnock,
1995). This gives such talk about care its contradictory and dilemmatic quality.
4 See Perakyla and Silverman (1991) for a conversation analytic examination of owning
experience in multiparty AIDS counseling sessions.
165
Working the Boundaries Between the Extraordinariness and Ordinariness
As a parent dealing with the idiosyncrasies of chronic illness, you are not part or
at least cannot assume to be part of some "default" consensus about what, how,
when, with whom, for what purposes you should proceed in the care of your
child. This uncertainty of how you should represent the circumstances you are in
is definitional of what it is to be the parent of a child with a chronic illness. Under
the more routine circumstances of parenthood, we can on the whole assume there
is a solution , a way, something acceptable, appropriate in the way we conduct
ourselves. Such default assumptions are precisely what is missing in caring for
children who have developmental difficulties or are diagnosed as having some
form of chronic illness. The accomplishment of the talk in the parents' support
groups discussed here is the ratification they afford of forms of "ordinarinesses"
in the extraordinary circumstances of the parents' child care and in terms of
outcomes and regimes of treatments. In accounting for their actions, and their
action in telling their experiences, parents articulate the basis for understandings
held in common, understandings that can be claimed as warrants in future
elaborations of the significance of their struggle to define new forms of what
constitutes the normal.
Sacks (1992) pointed out that what it is to be an ordinary person in the world
cannot be taken for granted. Being ordinary is something we work at in our
communicative relationships with others. Ordinariness is not an inevitable
product of being an ordinary person. Rather we expend much effort in
demonstrating in our communicative relationships with others that our lived
experience of life can be taken as being the experience of an ordinary person.
In the context of parents who care for children with chronic renal failure, what
is it to be ordinary? Their doing of being ordinary has to be done with the
resources that are the common currency of their unusual circumstances. It is not
that they are ordinary parents of children with an unusual condition. Rather they
have to establish, they have to do in their talk with others , what it is to be
ordinary as parents-of-children-with-renal-failure. We have seen that there are
recurrent uncertainties in doing this that are a key resource in doing being
ordinary. One of the practical consequences of the analysis presented is an insight
into why narrating health care experiences provides such an effective device in
group meetings.
This is psychosocial support as examined within the communicative actions of
parents. We see joint reasoning in settings of extreme uncertainties about just
what to expect. The parents' experiences of isolation are narrated in terms of
these uncertainties. Uncertainties over what is normal in their extraordinary
circumstances, about just how to represent what was, is, or might be, give their
talk its dilemmatic quality . Their normality is an extreme of uncertainty. The
group affords joint reasoning about what is reasonable both in terms of
circumstance and outcome: that is, in terms of their incorporation and
David Middleton
166
engagement in the process of care for their children and in terms of the sort of
outcomes they might reasonably expect.
Transcription Conventions
distinctly quieter than surrounding
talk?
continuing talk $
interru [ption
[or spoken in
[unison
sequence spoken faster than

preceding talk-c
<sequence spoken slower than preceding

talk>
emphasized
(3) pause in seconds
(sounds like - guess)
[additional information]
. a stop with a fall in tone
: stretched preceding sound or letter
{details omitted to preserve anonymity} ? rising inflection

'hh in breath
hh out breath
eAPITALS louder than preceding talk
- cut off prior word or sound
C) short pause in talk <0.5 seconds
C..) indecipherable
167
Acknowledgments
Grateful thanks are extended to the parents and families who consented to have
their conversations recorded; to Dot MacKinlay and Judith Argles for their
collaboration in recording the parent groups; to Jacqueline Collier for the initial
transcription of the conversational material; and to members of the Nottingham
Paediatric Renal Unit for their support and enthusiasm. Other people have been
most generous in their critical comments and suggested amendments, including
members of the Department of Communication Studies (Tema K) and Child
Studies (Tema B) at Linkoping University, Sweden ; members of the
Loughborough Discourse and Rhetoric Group; delegates at the 1993 NATO
Workshop on "Discourse, Tools and Reasoning," and Anssi Perakyla, Clotilde
Pontecorvo, Roger Saljo, and Barbara Burge.
168
David Middleton
References
Abraham, C. S., & Hampson, S. E. (in press). A social cognition approach to health
psychology: Philosophical and methodological issues. Psychology and Health.
Argles, J., MacKinlay, D. R. E., Middleton, D. J., & Watson A R. (1994) . Parent support
groups : A local account. Maternal and Child Health : The Journal ofFamily Medicine.
19(5), 152-158.
Atkinson, J. M., & Heritage, J. (Eds .). (1984) . Structures of social action : Studies in
conversational analysis. Cambridge, England: Cambridge University Press .
Billig, M. (1987) . Arguing and thinking : A rhetorical approach to social psychology.
Cambridge , England : Cambridge University Press .
Billig, M., Condor, S., Edwards, D., Gane, M., Middleton, D., & Radley, A. (1988) .
Ideological dilemmas: A social psychology ofeveryday life. London : Sage Publications .
Buchanan, K., & Middleton, D. (1995). Voices of experience: Talk, identity and
membership in reminiscence groups. Ageing and Society, 15,457-491.
Buttny , R. (1993) . Social accountability in communication. London : Sage Publications .
Collier, J., & Watson, A R. (1994) . Renal failure in children : Specific considerations in
management. In H. McGee & C. Bradley (Eds.), Quality of life following renalfailure:
Psychosocial challenges accompanying high technology medicine (pp. 225-245). Chur,
SwitzerlandlPhiladelphia, PA: Harwood Academic Publications .
Edwards, D., & Potter J. (1992). Discursive psychology. London: Sage Publications.
Goffman, E. (1981). Forms oftalk. Oxford, England: Basil Blackwell.
Heritage, J. (1984) . Garfinkel and ethnomethodology. Cambridge, England: Cambridge
University Press.
Jefferson, G. (1984) . On the organisation of laughter in talk about troubles. In J. M.
Atkinson & 1. Heritage (Eds.), Structures of social action: Studies in conversational
analysis. Cambridge, England: Cambridge University Press.
Lerner, G. H . (1993). Collectivities in action : Establishing the relevance of conjoined
participation in conversation . Text, 13(2), 213-245.
Middleton, D. (1996a) . A discursive analysis of psychosocial issues : Talk in a "parent
group" for families who have children with chronic renal failure . Psychology and
Health, 11, 243-260.
Middleton, D. (1996b). Talking work: Argument, common knowledge and improvization
in multi-disciplinary child development teams . In Y. Engestrom & D. Middleton (Eds.),
Cognition and communication at work (pp. 233-256) . Cambridge, England: Cambridge
University Press.
Middleton, D., & Curnock, D. (1995). Talk of uncertainty : Doubt as an organisational
resource for co-ordinating multi-disciplinary activity in neonatal intensive care. In G.
Loomes (Ed.), Risk in organisational settings . York, England: Economic and Social
Research Council.
Perakyla, A, & Silverman D. (1991) . Owning experience: Describing the experience of
other persons . Text, 11(3),441-480.
Perakyla, A (1994). Personal communication.
Potter, 1., & Wetherell, M. (1987). Discourse and social psychology: Beyond attitudes and
behaviour. London: Sage.
Sacks, H . (1992) . Lectures on conversation. (Vols 1 & 2; G. Jefferson, Ed.), Oxford,
England : Basil Blackwell.
Chapter 7
Syncretic Literacy in a Samoan American

Family
Alessandro Duranti' and Elinor Ochs'
'Department of Anthropology, UCLA, USA
2Department of Applied Linguistics , UCLA, USA
Abstract
After examining three misconceptions of the concept of multiculturalism, we

introduce the concept of syncretic literacy to deal with how diverse cultural
frameworks inform the organization of literacy activities by members of the
Samoan American community in urban Los Angeles . On the basis of our earlier
work in a rural Western Samoan community, we discuss the ways in which the
strategies used by caregivers in traditional learning environments are also found
in a Samoan American household in urban Southern California. By analyzing a
sequence in which a young child asks for help with his homework assignment, we
show that (a) the code used (e.g., English) is not always a good predictor of the
participants' cultural orientation and (b) members of multicultural communities
can be in more than one culture at a time.
Multiculturalism
In this chapter, we examine the multicultural organization of literacy instruction
within the Samoan American community of urban Los Angeles . In particular, we
introduce the concept of syncretic literacy as a framework for analyzing how
170
Alessandro Duranti & Elinor Ochs
diverse cultural frameworks inform the organization of literacy activities of this

and other ethnic communities.
Multiculturalism is a pervasive social reality whose complexities have boggled
the minds of social scientists and educators for decades. In characterizing and disentangling the diverse cultural threads that compose heterogeneous communities,
three common misconceptions of multiculturalism persist. We examine these
misconceptions before discussing syncretic literacy.
Misconception #1: Language Is a Precise Indicator of Cultural

Orientation
Multiculturalism is commonly identified with multilingualism. It is entirely
intuitive to link language and cultural orientation. After all, as we acquire
linguistic competence, we are at the same time being socialized into cultural
competence, a process we call language socialization (Ochs & Schieffelin, 1984;
Schieffelin & Ochs, 1986). Language is the most important semiotic tool for
representing, transmitting, and creating social order and cultural world views.
Furthermore, in linguistically heterogeneous communities, choice of a particular
language is often intimately tied to a desire on the part of interlocutors to
instantiate for that particular interactional moment a set of sociocultural
relationships, institutions, activities, topics, concepts, ideologies, expectations,
and values (Blom & Gumperz, 1972; Gal, 1987; Hill & Hill, 1986; Kroskrity,
1993; Macpherson, 1991). In these cases, code-switching is an analog of cultureswitching. In other cases, however, cultural orientation may not correspond to
code orientation . As will be demonstrated in the present study, multiculturalism
may in fact pervade the use of what appears as a single code. In the Samoan
American community, for example, one may use English in a distinctly Samoan
manner or may use Samoan in a manner appropriate to mainstream American
interactions. Although language is an important symbol and tool of culture, the
researcher cannot count on language as a privileged key to how cultures interface
in the literacy activities of a person or of a community . We need to revise
currently dominant notions of language as code and see it instead as a set of
practices, including specific ways of speaking and of interpreting the world (words
included), as well as a means of interacting with human, symbolic, and material
resources available in the environment (Goodwin & Duranti, 1992; Hanks, 1990;
Hymes, 1974).
Misconception #2: Members of Multicultural Communities Are in

One Culture at a Time
A second misleading tendency in analyzing multiculturalism is the assumption
that members of multicultural communities shift from one cultural orientation to
Syncretic Literacy in a Samoan American Family
171
another in the course of conducting their daily social lives. As noted earlier, in
this view, just as one switches between codes, so one switches between systems
of social values, beliefs, emotions, practices, identities, and institutions.
Although it may be useful to treat cultures as coherent and separate and although
in some cases members of multicultural communities do draw boundaries between
what they consider traditional and what they consider new, more typically cultural
threads from diverse sources are interwoven into a single interactional fabric. In
the Samoan American community, members tend not to abruptly culture-switch
as they move from one activity to the next or one setting to the next; rather, they
blend cultural orientations in the course of carrying out a single activity,
including the activities of reading and writing Samoan and English .
Our research in this area is consistent with recent studies of immigrants in the
United States that stress the importance of seeing these groups as operating
simultaneously in two communities (Chavez, 1994; Rosaldo, 1989; Zentella,
1990). Our work, however, differs from these studies in our emphasis on the need
to study the daily, moment-by-moment confluence of multiple cultural models
and language-mediated practices. We believe that multiculturalism as a sense of
belonging to more than one community is not only imagined as an ideology of
connections (Chavez, 1994), but it is also enacted in daily routines that need to be
unpacked, if we want to uncover the different cultural threads they both imply and
sustain. A substantial part of this chapter is dedicated to the discussion of some of
these routines and threads in a Samoan American family.
Misconception #3: Each Culture Is Homogeneous and

Uncontaminated
Another misleading picture of multiculturalism is that each ethnic community in
the multicultural stew is itself internally undifferentiated and untouched by the
influences of other communities. In the struggle to illuminate ethnic
communities in the United States, for example, it is sometimes assumed that the
traditional culture brought from one's homeland is/was shared and untainted by
multinational economic and ideological forces. This perspective, however, is
belied by centuries of trade and missionization in third and fourth world and even
highly insular communities. Literacy activities in particular are probably always
complex structures organized by multinational interests, as well as indigenous
frameworks for carrying out social life. In many societies, literacy was introduced
and transmitted by foreigners committed to establishing religious ideologies
and practices among the unconverted (Grillo, 1989; Mannheim, 1991; Romaine,
1994). To understand multiculturalism in literacy activities within urban ethnic
communities, we need to (a) assume that, even in communities of origin, a
complex heterogeneity of traditions informs the practices of reading and writing
and (b) be prepared to analyze how this heterogeneity structures the literacy
practices of ethnic communities in the United States. To that end, we draw upon
172
more than a decade of research that we have conducted on language activities in

rural communities in Western Samoa to illuminate the multicultural organization
of literacy activities among Samoan Americans.
In this chapter, we examine the activity of doing homework, and in so doing,
we discuss the first two misconceptions outlined earlier. The third one (each
culture is homogeneous and uncontaminated) is examined in Duranti and Ochs (in
press) , where we discuss the uses of a particular literacy tool , the Samoan
alphabet table, in Western Samoa and in the United States.
Syncretic Literacy
In counterpoint to these three misconceptions of multiculturalism, we propose
considering literacy activities in all communities as syncretic . By syncretic
literacy, we mean that an intermingling or merging of culturally diverse traditions
informs and organizes literacy activities. This use of syncretism is drawn from a
number of sources, including studies of the interaction of religions in contact
situations (Apter, 1991; Herskovits, 1937, 1952, 1966), Bakhtin's work on
heteroglossia and hybridization (Bakhtin , 1981, 1984; Voloshinov, 1973), and
contemporary studies of multilingual communities and postcolonial discourse
(Hanks , 1986, 1987; Hill & Hill, 1986; Kulick, 1992).
In the study of religion, especially in Africa and the New World, syncretism is
generally defmed as the combining, intermingling, or merging of different belief
systems (Droogers, 1989). Herskovits (1952) identified syncretism as a form of
reinterpretation, a cultural process that affects all aspects of cultural change. This
anthropological use of the concept of syncretism differs from the way in which
the term has been used in psychology, where it connotes a somewhat incoherent
and transitional developmental stage, as in the acquisition of word meaning
(Vygotsky, 1986).
Our work is also inspired by Bakhtin's view of how different dialects,
perspectives, and voices can coexist within the same national language or even
within the same utterance. Bakhtin (1981) and his collaborator/alter ego
Voloshinov (1973) expressed a harsh criticism of synchronic investigations that
took monolingual, monological speech genres as the normal or unmarked form of
communication. They believed that any approach that takes an individual system
or code as the ideal object of inquiry is problematic because it does not allow us
to uncover the richness of language used by speakers as well as novelists. As
pointed out by Hill and Hill (1986) , bilingual situations in which codes are
routinely mixed in creative and sometimes contested ways are ideal contexts for
testing out Bakhtin's theory and, in particular, his notion of hybridization: "a
mixture of two social languages within the limits of a single utterance, an
encounter, within the arena of an utterance, between two different linguistic
173
consciousnesses, separated from one another by an epoch, by social differentiation

or by some other factor" (Bakhtin, 1981, p. 358). If we accept Bakhtin's view that
"any living utterance in a living language is to one or another extent a hybrid"
(Bakhtin, 1981, p. 361), the study of multilingual, multicultural communities
can become the ground on which to test current theories of language use rather
than the exception to be explained by special principles that do not necessarily
apply to monolingual communities .
For us, syncretic literacy is not necessarily restricted to a blending of
historically diverse literacy traditions; rather, syncretism here may include
incorporation of any culturally diverse values, beliefs, emotions, practices,
identities, institutions, tools, and other material resources into the organization of
literacy activities . The main idea behind this notion is the belief that, when
different cultural systems meet, one rarely simply replaces the other. This means
that, as pointed out by Hanks (1986, 1987) for the Maya, as soon as contact takes
place, any pre-existing indigenous tradition is bound to be affected by the new
tradition proposed (or imposed) by the newcomers. At the same time, it is
counterproductive to conclude that the blending is such that one cannot trace the
influences of different traditions or the culture-specific strategies used by
participants. Hanks's analysis, for instance, makes ample use of two distinct
traditions in order to describe their blending in the Maya discourse of the last four
centuries. We follow a similar path by differentiating between modes of
interaction in the bilingual situation that resemble those found in most adult-child
interactions in Western Samoan households and those that are found in school
settings and literacy activities in both the U.S. and Western Samoa.
The term syncretic has been used by Hill and Hill (1986) to characterize the
hybridization of codes used by the Mexicano speakers who rely on Nahuatl and
Spanish . We wish to extend syncretism to include hybrid cultural constructions of
speech acts and speech activities that constitute literacy. In our case, the
availability of an audiovisual record of the interactions we are studying allows us
to widen the concept of syncretism (as well as the concepts of hybridization and
heteroglossia). Rather than focusing on a code- or text-centered notion of
syncretism, we analyze the merging of different activities, acts within the same
activity, and tools that originated in and are indexically related to different cultural
traditions.
In developing this approach to syncretic literacy, we are inspired by recent work
of linguistic anthropologists and other students of face-to-face communication
who emphasize the need to recognize the multiplicity of contexts that give
meaning to the same strip of interaction (Cicourel, 1992; Goodwin & Goodwin,
1992; Keating, 1994; Kendon, 1992).
In our view, the syncretic nature of literacy activities is due to a number of
factors, which include (a) the need to use tools that originally were not
specifically designed for literacy activities, (b) the need to connect different
cultural traditions and expectations, and (c) the ability to coordinate among
competing simultaneous activities.
174
The simultaneous occurrence of multiple, sometimes competing activities has

often been commented on by researchers studying literacy events in the home
environment:
Our data clearly indicate that literacy events function not as isolated bits of human
activity but as connected units embedded in a functional system of activity
generally involving prior, simultaneously occurring, and subsequent units of
action. (Anderson, 1984, p. 28)
A crucial feature of familial organization is the fitting together of multiple
simultaneous activities. Analyzing these parallel activities raises the broader
questions of how literacy is embedded in activities that are carried out for purposes
other than instruction, and how reading and writing are combined with other
modes of communication. (Leichter, 1984, p. 44)
Learning as we ordinarily recognize it in our lives with children seems to be a
subordinate activity in most homes . . . . learning seems to occur during the
performance of other and more important tasks essential to the daily table of
organization. (McDermott, Goldman, & Varenne, 1984, p. 399)
These studies, among others (Heath, 1983; Lave, 1988; Scribner & Cole, 1981),
emphasize the importance of shifting the almost exclusive focus on writing (or
print) as a generalized technology characteristic of earlier studies of literacy
.(Goody, 1977; Goody & Watt, 1968) to an appreciation of the importance of
specific activities organized around and through print. In this chapter, we continue
in this tradition by integrating the emphasis on activities with an
ethnographically informed attention to the physical environments in which
literacy activities take place and the tools and artifacts that are made available in
such environments. When we look at literacy practices outside the school
environment, we find that the material resources utilized for literacy tasks have a
crucial role in establishing the breadth as well as the limits of what literacy can do
or mean for the children engaging in it. To be able to participate successfully in
literacy activities, children must learn to utilize resources that were not originally
designed for literacy practices. Even in those cases in which the materials used
were originally conceived as literacy-tools, the context in which they are used
forces a new interpretation of their original meaning. Placed in a new context, old
tools not only bring in remnants of the past but also force participants to face
issues of tradition, change, and social identity.
175
Field Research
Starting in 1993, we have been video recording Samoan American children and
adults interacting in a variety of settings, including four households and a church
compound in the greater Los Angeles area. The present study of children's literacy
activities is part of a more comprehensive study of how Samoan American
families educate their children into problem-solving discourse activities. This
project is a component of a larger project funded by the U.S. Department of
Education and coordinated by the National Center for Research on Cultural
Diversity and Second Language Learning at the University of California, Santa
Cruz.
In carrying out this urban multicultural and multilingual study, we are building
on a number of projects previously conducted by the two principal investigators,
including (a) a study of language socialization and language use across a number
of social settings in a rural village in Western Samoa (Duranti & Ochs , 1986;
Ochs, 1982, 1988); (b) a study of problem-solving discourse in decision-making
councils in Western Samoa (Duranti , 1981, 1990, 1994); and (c) a study of
problem-solving discourse during dinner conversations among white families in
the Los Angeles area (Ochs, Smith, & Taylor, 1989; Ochs & Taylor, 1992a,
1992b; Ochs, Taylor, Rudolph, & Smith , 1992).
Features of Traditional Samoan Learning Environments

Before introducing the syncretic literacy activities we have been documenting in
Los Angeles, we provide a rough outline of the basic features of what we call
traditional Samoan learning environments.

We are aware of the fact that the label traditional, in this context, is potentially
misleading, especially in light of our own stance against rigidly placing
interactions into one category or another. Ultimately, categorical differences are
given by differences in degree, as well as by the ideology framing the choices
available to social actors. We are, however, confident that there exist interactional,
symbolic , and material features of the children's learning environment that
(a) are more closely associated by the participants themselves with the home
country and a traditional way of life (often lexicalized as thefa'aSamoa or
Samoan way in opposition to thefa'apalagi or Western way);
(b) characterize informal household interactions involving children in
Western Samoa as described in some detail by Ochs (1988); and
176

(c) are analytically distinguishable from patterns of interaction found in
school-based literacy practices in Samoa and in the United States (see
discussion to follow) .
The historical, psychological, and interactional reality of a distinct Samoan way is

necessary to entertain the very idea of syncretic literacy and to discuss those
features ascribable to one or the other culture or to both cultures. It is important,
therefore, to realize that, in proposing our notion of syncretic literacy, we are not
arguing in favor of fuzzy notions of culture and language but in favor of a more
fluid analytical understanding of the ways in which specific elements that
originated in different sociohistorical contexts come together in the daily life of a
Samoan American child and (more generally) of any multicultural child in the
U.S .
In our observations of traditional learning environments, we have found the
following four principles to be at work.
Caregiving: Hierarchical and Distributed

The organization of caregiving is closely related to the political organization of
the village. In particular, the stratified nature of Samoan society, where people
are distinguished in terms of status (titled vs. untitled) and rank (high chief vs.
orator), is reflected at the household level in the hierarchical organization of
caregiving, a feature of Samoan socialization that has been previously described
by Ochs (1982; 1988). The basic principle here is that, given two potential
caregivers, the more senioror higher ranking one will expect the younger and
lower ranking one to be the active caregiver. This is realized either through other-
selection - the higher caregiver directs the lower caregiver to carry out a task on
behalf of the child - or through self-selection - the lower ranking caregiver carries
out the task without being so instructed. Thus, for example, the practical care of
even young infants - rocking the infant to sleep, picking her up upon awakening,
changing clothes and bringing her to be breastfed, or, in the case of older infants,
feeding the child whole foods as well - tends to be the responsibility of older
siblings of the child or perhaps a younger sibling of one of the parents, if these
persons are not at school or called for other chores. If no one younger is present,
the child's mother or other adult takes charge of the child's care. Even when the
child is cared for by her siblings, an adult is usually within relatively close
proximity and intermittently monitors the caregiving activity (see Figure 7.1).
In this sense, caregiving is not only hierarchically organized but also socially
distributed. For example, caregiving is distributed when a sibling carries a hungry
infant to her mother and the mother breast feeds the child . Similarly, when an
adult directs a sibling on how to carry the infant or to wash the infant, the task is
collectively accomplished.
177
Fig. 7 .1. Young boy oversees sleeping baby while mother sits nearby (Western
Samoa, 1979)
Child Expected to Accommodate to Situation

In Samoan households and communities , there is an expectation that a child must
accommodate to the situation rather than an expectation that other people present
will change their goals or activities to accommodate to the child. We refer to the
first set of socialization practices as situation-centered and the latter as childcentered (Ochs & Schieffelin, 1984).
Fig. 7.2. Baby is fed facing outward

(Western Samoa , 1981)
Fig. 7.3. Baby sits facing outward

(Western Samoa, 1981)
In rural Western Samoan villages, children are socialized to attend to what is

taking place and who is present in a variety of ways. For Samoans, the most
178
important quality a child must display is faaaloalo [respect]. The term includes
the term ala, meaning "(to) face (in the direction of)." From infancy, children are
held or placed facing outward, paralleling the orientation of the caregiver toward
some interaction at hand. This is shown in Figure 7.2, where a baby is fed facing
outward, and in Figure 7.3, where he is made to sit up to look at the rest of the
interaction in the house (where his mother and other young women are sewing
colored feathers to sleeping mats) .
Young children are expected to position themselves at the edge of dwellings to
observe and ultimately report on what is taking place outside. In addition, toddlers
are explicitly socialized to identify and call out names of family and community
members as they pass by (see Figure 7.4).
Fig. 7.4. Young boy observes at the edge of a dwelling (Western Samoa, 1979)
Socialization Through Repeated Demonstration, Prompting, and

Action Imperatives
In every community, members interact with children and other novices in ways
that allow them to appropriate cultural knowledge, beliefs, stances, expectations,
preferences, and practices (Rogoff, 1990). Further, by and large, in every
community, members rely on a similar core of strategies for facilitating the
appropriation of these cultural orientations. However, communities differ in their
reliance on one or the other socializing strategy. That is, they differ in terms of
179
strategies they usually employ and in terms of the activity settings in which they
opt for one or another strategy. In traditional Samoan households, three
socializing strategies prevail: repeated demonstration of an activity, prompting,
and action imperatives.
Generally, Samoan children are allowed a prolonged period of observation of
repeated denwnstrations of actions and activities before being expected to assume a
central role in them. For example, young children are not pressured into talking
before they articulate words and show themselves to be interested in
communication. Children are given ample opportunities to observe a wide variety
of culturally relevant activities . They accompany older siblings, parents, and
grandparents in work and, to some extent, in formal school environments. As
siblings and adults weave mats, carve boats, braid string, shuck coconuts, cook,
do laundry, read the Bible, and deliberate political matters, toddlers are close at
hand. Often a sibling will be carrying out some other work activity as he watches
over a younger brother or sister. The toddler may stand close by, attending to the
activity at hand.
Children and adolescents spend a long period of time as overhearers of
ceremonial speeches. Only much later in life are they expected to display their
own oratorical skills. Young children are often brought to ceremonial events by
their grandparents to sit by them and occasionally fetch cigarettes and other items .
Adolescents are expected to be present and ready to act as messengers, food
bearers, and gift announcers in their traditional role of servant in political,
religious, and other formal activities. Similarly, children spend long periods of
time as audience to dance performers as they practice dance routines over and over
again before they are themselves public performers. Young children may sit on
the laps of older community members, occasionally imitating the movements of
the dancers, but the focus of audience and dancers is on the performance. In these
ways, by the time they are asked to participate more centrally in an activity,
children have usually had many occasions to witness its completion (see Figure
7.5).
Similar to the strategy of repeated demonstration is the socializing practice of
prompting. Whereas repeated demonstration involves a member reproducing a
cultural activity in the presence of a novice, prompting involves the novice
reproducing or attempting to reproduce an act or activity. Prompting is a
socializing interaction in which a (usually) more knowledgeable interlocutor
elicits the repetition of an action or activity by a (usually) less knowledgeable
interlocutor. Although prompting is universal, some societies, such as rural
Western Samoan communities, rely heavily on this practice as a form of
instruction (see also Demuth, 1986; Miller, 1986; Schieffelin, 1990; WatsonGegeo & Gegeo, 1986).
180
Fig. 7.5. Samoan children and caregivers watching a ceremonial exchange (Western
Samoa, 1979)
In Samoan, the activity of prompting what to say is usually keyed by prefacing

the utterance to be repeated with the verb fai meaning "say" (and also "do").
Prompting may also be keyed prosodically through varying amplitude and raising
pitch height. Prompting is common both at home and in formal school
environments. At home, prompting is a means of involving young children as
participants in multiparty interactions. More particularly, prompting offers a
model of age (and status) appropriate behavior for children to then reproduce. For
example, a sibling or adult may prompt a child to tell a third party a piece of
news or a directive on behalf of the sibling or adult. In these cases, the child is
being socialized through prompting to act as messenger, a role expected of
younger untitled persons. In delivering"messages, the child serves the author of
the message. Much like an orator in relation to the high chief the orator
represents, the child represents the older family member and thereby manifests the
high status of that member.
A third related socialization strategy is to direct explicitly the actions of the
child or novice through the use of imperatives. Action imperatives lace the
interactions of family members. As in all communities, children are ordered to
assume age and gender appropriate demeanors (e.g .,faitai 'ou vae ! or "cross your
legs!") or to carry out particular tasks (e.g., 'aumai se vai a le tama! or "bring
some water for the boy!"). Commonly a toddler will be ordered to carry out what
Lave & Wenger (1991) call a peripheral task within the larger work effort taking
181
place. For example, a mother sweeping the household compound will give a
basket to her two-year-old and order her to pick up fallen leaves on the compound
or to rearrange mats in a dwelling. In our observations, generally children are not
ordered to carry out tasks that they cannot easily carry out by themselves.
Emphasis on Task Completion

As noted earlier, in Western Samoan rural communities, children spend
considerable time in close proximity to work activities carried out by others, and,
in some cases, they themselves participate in these activities. As observers and as
participants, they are expected to be mindful of and contribute to the
accomplishment of these activities (Duranti & Ochs, 1986). This expectation is
part of the situation-centered focus of socialization in Samoan households.
When a household activity in which a child participates is completed, family
members tend not to focus on the child to praise or blame him or her. Rather than
emphasizing the child's individual contribution, members of Western Samoan
households tend to focus either on the fact that the task has been correctly
performed or on the task as a collective and collaborative accomplishment.
Linguistically, this is realized in two prevalent forms of assessment. First,
those monitor ing the task may produce an assessment of the action or activity
performed rather than of the performer. These impersonal assessments point out
the appropriateness, correctness, or completion of something rather than the
specific quality of the child's contribution. Thus, frequently heard expressions,
such as fa'apefla! meaning "like that!" (often abbreviated as fla!), predicate a
quality of an impersonal referent: namely, the action or activity .' However,
individual-oriented laudatory expressions such as "very good!" or "good boy!" expressions that are rife in interactions with children in middle class American
families - are not generally directed to children in Western Samoa.
In Western Samoan communities, those monitoring a task may engage in a
ritualized exchange of linguistic assessments of the action or activity that
acknowledges the collaborative nature of that behavior . One person marks
completion of the task with the assessment, such as malO! meaning "Well done!",
and the addressee immediately responds back malO fo il, "Well done also!" In so
doing, the praise is reciprocal and distributed rather than directed to any single
participant. This type of exchange is typical of adult-adult interaction but less so
of child-child or adult-child interaction.
ITo those familiar with Samoan orthography, our transcription will appear, in some
respects, unconventional . We follow the tradition in using the inverted apostrophe C)
to represent a glottal stop (+ ), but we replaced the macron found in traditional Samoan
orthography with the circumflex accent (e.g., a, e). See section on Transcription of
Interaction that follows and Footnote 2 for an explanation of our transcription of
colloquial Samoan .
182
Context of Syncretic Literacy

An important requirement of any ethnographically based study of language use is
the need to have an understanding of the range of events within which a particular
type of phenomenon occurs (Bauman & Sherzer, 1975; Hymes, 1962, 1974).
This applies to biculturalism and bilingualism as well (Hill & Hill, 1986).
Although we will not be able to elaborate on all of the relevant contexts in which
syncretic literacy takes place, we want to mention briefly some of the contexts
that we have been analyzing: (a) church-based literacy instruction of Samoan in a
village in Western Samoa (cf. Duranti & Ochs, 1986); (b) church-based literacy
instruction of Samoan in Southern California; and (c) home literacy instruction of
English in Samoan American families in Southern California.
In our discussion of syncretic literacy, we concentrate on the last setting, while
keeping in mind the relevance of the others. A comparison of the three settings
will be provided in future publications.
Transcription of Interaction
' In our transcripts, wefollow a modified version of the conventions introduced by
Gail Jefferson for conversation analysis (Sacks, Schegloff, and Jefferson, 1974). A
detailed description of our conventions is found in the Appendix.
Here, as elsewhere in our publications on Samoan language and culture, we
have chosen not to modify speakers' colloquial expressions, and thus we have not
hidden any use of what Samoans call "bad speech" (tautala leanai, a phonological
register where the two alveolar sounds represented in traditional Samoan
orthography as t and n are pronounced as the velar sounds Ik/ and 111/,
respectively- (Duranti & Ochs, 1986; Shore, 1982). Whereas bad speech
pronunciation is itself common in formal and ceremonial events in Samoa
(Duranti, 1981, 1994), and for this reason alone cannot be simply labeled casual
or informal, the combination of this pronunciation with the colloquialisms and
morpho-syntactic reductions characteristic of many household interactions gives it
a rather private connotation that many Samoans do not like to see displayed
publicly, especially in a written form. Although recognizing the complexity of
Ufo avoid ambiguity, we replaced the letter g introduced by the missionaries to
represent the velar nasal with the phonetic symbol zn/ (in other Polynesian languages ,
the same sound is represented by ng) . It should be pointed out that the sound IT]I is also
found in good speech, but in bad speech , IT]I replaces all Inl found in good speech .
Thus , certain distinctions found in good speech (e.g., Ifanal "gun , shoot" vs. /fana/
"bay ") are neutralized in bad speech - where bothfana andfaTJa are pronounced [fana],
183
the issues involved in this as in other debates over representation of ethnic

identity through language use (see Morgan, 1994, for a discussion of these issues
in the study of African American English), we feel that the particular style or
register used by the participants is such an important part of the on-going
interaction that it cannot be hidden or modified without altering the nature of the
phenomena being studied. Furthermore, in our case, the range of registers spoken
in the home (which includes several varieties of English and of Samoan) is part of
the linguistic and social reality in which Samoan American children are raised. A
study of the different cultural frames they are part of cannot but include the range
of linguistic resources they are exposed to and are expected to understand.
All of the names of the participants in the interactions discussed here have been
changed, and the visual record of their interactions has been framed and cropped to
protect as much as possible the identity of the participants .
The transcripts, translations, and interpretations of the interactions discussed
here were made in collaboration with Elia K. Taase, a member of the Samoan
American community in this study. Other members of our research team who
have also contributed to the discussion of the themes and issues discussed here
include Jennifer Reynolds, James Soliai, and Edgar Taase.
Syncretic Literacy Instruction Among Samoan Americans

in Southern California
In our view, syncretism is a property that can be found in a number of contexts,
including the display and use of material resources, the spatial arrangement of
human bodies and tools, the organization and content of co-occurrent activities,
and the linguistic means through which specific acts are accomplished. In the
following three sections, we will be looking at syncretism in (a) the environment,
(b) activities, and (c) speech acts.
Syncretism in the Environment: Material Resources and Spatial

Organization
In entering a Samoan-American household in a Southern California suburb, one is
struck by the multiculturalism of the physical surrounding as achieved through a
syncretism of elements of material culture from two traditions, as well as by the
spatial organization of such elements, including the occupation of the space by
the participants' bodies. Our earlier work on the cultural organization of space in a
traditional village in Western Samoa (Duranti, 1981, 1992, 1994; Ochs, 1988)
can guide our perception and interpretation of the organization of space in the
houses occupied by Samoan American families in Southern California.
184
tv ..... /vldc09l1mc
bockyo rd
ki tchen
front door
STREET
Key
semicircle (..)) indicates position of pelvis

(bird's eye view) of participants, Gm=grendmother, Gf=grendfutb er, A.D .=researcher
~ = mats,
= person lying.
Fig. 7.6. Map of section of Siks's house with participants ' locations at the
beginning of first visit
On May 27, 1993, one of us (A. Duranti) went for the first time to visit Sike's
family, one of the four families in our study. The shoes and sandals left in front
of the entrance were the first of a long series of reminders that the external shape
of the building, a two story house with a small front yard and a garage on the
side, was perhaps the only feature in common with the other families on the
block .
185
When the researcher walked in, he found several members of the family in the
living room area, including Sike, his cousin Mata, Sike's grandparents, and
Sike's aunt (his grandparents' unmarried daughter). Despite Duranti's repeated
attempts not to upset the flow of events with his presence, he was immediately
offered some food and encouraged to sit down and eat it. Before sitting down,
however, he set up the tape recorder on the floor and the video camera on a tripod.
Figure 7.6 provides a map of where everyone is located at the moment when
Duranti sits down to eat, and Figure 7.7 shows a frame from the video taken
while he is consuming the meal and talking with the grandparents .
Fig. 7.7. (From left) Grandfather, grandmother, Sike and his cousin, Mata, sitting at
the TV screen, and Sike's aunt (resting on the couch ; only her head is visible).
As can be seen in Figure 7.7, the grandfather is sitting on the floor to the left; the
grandmother is sitting on the floor farther away from the camera. Sike (6 years
old) and Mata (12 years) are sitting facing the opposite corner, playing a video
game; Sike's aunt (who had prepared and delivered the plate with food for the
researcher) is lying on a couch watching the children play while also listening to
and occasionally participating in the conversation among the adults in the room;
the researcher (A.D.), who is not visible in the video frame, is shown in Figure
7.6 to be facing the grandfather and the grandmother.
Several aspects of this setting are worth reflecting on, especially from the point
of view of the syncretism of material resources and the distribution of activities
and roles.
186
Despite the presence of the two TV screens, the telephone in the center of the
room (next to the grandmother), the furniture , the central chandelier (off screen),
and the thick walls, the camera has captured a scene here that, in several respects ,
resembles what one might see in a very different type of house in a Samoan
village. In particular, the seating arrangement closely resembles the type of spatial
organization that we have documented in Western Samoa. If we take as a point of
view the entrance to the house and the road, participants appear seated according to
traditional categorical distinctions, with the guest and the grandparents located in
the front part of the room and the grandchildren and the unmarried daughter in the
back region (an area that in Samoa would be considered an extension of the basic
floor plan of the house). In the front part of the room, a further distinction can
also be drawn between the guest and the grandfather (who is also a chief) who are
seated facing one another, as appropriate to people of equal rank (Shore , 1982), in
the more prestigious region, and the grandmother in the relatively lower position
(the back) . As shown in Figure 7.7 , only the section of the room with the
grandparents and the guest is covered by sitting mats, 3 whereas the back region
with the children and the unmarried daughter shows the house floor rug.
further beck regi on
for food preperettcn.
young women end
children
RACK->
reg ion
for high
chi ef s
FRONT ->
region for wi yes end

orotors of les s er ron"
reg ion
for high
ch iefs
region for guests and

orators of high ran "
rood
Fig. 7.8. Scheme of traditional spatial distinctions in a Samoan house when guests
and other people of high rank are present.
Figure 7.8 represents the hierarchical division of space in a more traditional house
in Western Samoa with an extension in the back (adapted from Duranti, 1992). In
Figure 7.9, the seating arrangement in the living room of the Southern California
3These are an imported brand of industrially produced mats that are not found i n
Western Samoa , where the sitting mats are hand-woven, like all other mats, by the
women of the household.
187
house is graphically matched with the seating arrangement in the house in

Western Samoa.
tv wlvideogame
Ch
iaCr
~Aunt
beck->
front -)
Ibi9l
ITV.-I
CAMERA
RORD
Fig. 7.9. The two maps are matched to show salience of traditional organization of
space.
Syncretism of Activities
A few minutes later, the activities as well as the positions of the same
participants have shifted considerably. Mata and the grandfather are the only two
people left in the living room area. Mata is doing her homework, and the
grandfather is watching television. The grandmother is at the sink in the kitchen;
Sike is sitting at the kitchen table reading his instruction sheets for his homework
assignment; and the aunt is outside in the backyard. While the researcher with
camera moves to the kitchen to follow Sike's actions, Sike gets up from the table
and heads outside where his aunt is . He asks for help . The aunt first gives him
directions on how to proceed; then, just as Sike is heading back toward the
kitchen table, she instructs him to get a box that is behind her, against the
external wall of the living room . After an attempt to get the wrong box (the
plastic crate in front of the sliding door), Sike identifies the box his aunt was
indicating and drags it to her (see Figure 7.10).
188
Alessandro Duranti & Elinor Oehs
Aunt
~:-:-IU- -~-
Sil< e
bOH
Fig. 7.10. Sike gets the box as instructed by his aunt.
The aunt adjusts the box on the side between her and Sike and indicates to him
that he ean write on it. The box has been transformed into a desk (see Figure
7.11).
Fig. 7.11. The box is transformed into a desk
189
As the camera moves closer and eventually outside of the house to follow Sike's
activity in the yard, we get a sense of the multiplicity of activities within the
same physical space and the roles played in them by people and material
resources. The visual record makes evident that homework is not the only activity
occurring in the backyard. Another competing activity is emerging, one that has
the box-desk as one of its essential elements . Next to Sike and his aunt, on a
couch, there are a number of traditional fine mats Cie toga) (see Figure 7.12).
Fig. 7.12. Rolled fine mats ('ie toga).
These are precious goods that are going to be sent back to Western Samoa to be
exchanged in a ceremony, called saofa'i, during which the grandfather will be
given a new and prestigious chiefly title . The box transformed by Sike and his
aunt into a desk is one of the boxes to be filled with fine mats. This makes the
box a tool with multiple but by no means equal or neutral functions. Each use of
the box indexes not only different types of activities within different value
systems (e.g., doing homework vs. packing fine mats for a ceremony), but also
different sets of culturally mediated expectations about children's and adults' roles
and about the goals of socialization. These different expectations are exhibited in
the following two segments. In the first, the aunt is instructing Sike while the
grandmother watches them, apprehensive and somewhat irritated.
190
Segment
1
2
3
Aunt;
4
5
Aunt;
Sik@;
7
8
Gm;
Sik@ ;
okay - count this,

how many books ?
pointing with pen as he counts one. two. three . four .
[
"(three. four.)
okay.
starts to write , turns to look at camera, goes back to write
changes position , leaning harder on box with his right arm
raises eyebrows exhibiting disapprovaljj'i
(2 .0)
10
Gm;
e:! leana le ki( 1J)ipusa!
11
Sik@;
'hey! the empty box (gets) ruined! '

still writing
In line 10, the grandmother expresses her concern about the box with a warning in
fast speech: e:! leana le ki(Tl)ipusa!, literally "the empty box (atinipusa) is
bad (leana)." Sike, however, continues to write on the box. A few minutes later,
the grandmother's fears come closer to reality as Sike leans over and pushes hard
enough with his elbow to cause a dent. She upgrades her warning (line 28),
and this time her words are briefly echoed by her daughter (Sike's aunt) (line 29):
4Pronounced eyebrow movement is here interpreted by members as a pre-disagreement.

This movement must thus be distinguished from the eyebrow flash studied by Eibl Eibesfeldt (1974) which, accompanied by a slight raising of the chin and (sometimes)
by a smile, conveys agreement.
191
Segment 2
27
Sike;
(kneels down on the box

and makes a dent))
28
Gm;
29
Aunt ;
30
31
32
Sike;
Aunt ;
Gm;
le kinipusa i 5 leana!
'the box 's gonna get ruined! '
'aua le so 'oga'don't over-' (i.e. 'don't press too hard')
(l write 'Tom'?)
yeah. write your "(name). write your middle name,
'a'e!
'oh! (disapproving)'
33
Aunt ;
dent
pointing to a point on page ku'u lema i 0

'put (it) slowly there'
Several aspects of these two interactions are worth examining from the point of
view of the syncretic nature of the literacy activity we find in it. First, we will
focus on the different ways in which the aunt and the grandmother interact with
the child. Whereas the aunt accommodates to the child by attending to the task
that he is proposing (doing homework) and letting him use the box that was
needed for another activity (packing fine mats), the grandmother is much less
accommodating and more concerned that the child does not jeopardize the adults'
forthcoming activity (packing) by ruining the empty box on which he is writing.
The aunt and the grandmother are thus following two different models: The
grandmother is situation-oriented (or accommodate-child-to-situation) and the aunt
is more child-oriented (accommodate-situation-to-child; see earlier section, Child
Expected to Accommodate to Situation . The point here is not that the grandmother is insensitive to the child's needs but that she is more concerned with the
5 This is a colloquial expression in which the preposition i before the predicate (leaga)
is used to introduce a likely and unwanted event. In careful good speech, the word
kinipusa would be pronounced atinipusa and in careful bad speech akiuipusa.
192
child learning how to accomplish his goals without interfering with those of adult
family members. At the same time, the aunt is not just assuming a western role
of accommodating caregiver. She is also acting within the logic of traditional
Samoan child-care by assuming the lower ranking role of active caregiver, hence
allowing the grandmother (the higher ranking caregiver) not to be directly engaged
in the task at hand (see earlier section, Caregiving: Hierachical and Distributed).
Syncretism of Acts Within a Single Activity
In the previous section, we have seen how two different caregivers, Sike's
grandmother and aunt, attending to two different activities in contact with one
another, use two different socialization strategies toward Sike to control his
behavior as he does his homework. In Segments 1 and 2, the codes the caregivers
speak match such diverse strategies, with English being used (by the aunt) to help
Sike in his homework (adapt situation to child) and Samoan being used to remind
him of his need to adapt to the situation.
In this section, we show that the same caregiver may produce a syncretic blend
of teaching strategies within the same activity. Sometimes the blending takes
place within the same code (English), other times in two separate codes (English
and Samoan) . In line 2 of Segment 1, for instance, 'the aunt asks a test question
(how many books ?), a strategy typical of American teachers but not of Samoan
caregivers in traditional communities. At other points (in line 1 in Segment 1 and
lines 13, 14, and 24 in Segment 3), she uses explicit action directives to Sike ,
thereby following patterns characteristic of Samoan caregivers in traditional
households (see earlier section, "Socialization Through Repeated Demonstration,
Prompting , and Action Imperatives") .
Segment 3
12
13
14
Aunt;
15
16
Sike;
17
18
19
20
21
22
23
24
Gm;
sns.
Sike;
Aunt;
okay there it goes. okay.

(Pulls sheet away from Sike) this one right here
count how many balls
pointing with pen as he counts one, (1.0) two,
(1.0) three, (1.0) four, (1.0) five , (1.0) six.
[
turns away --- moves metal frame on window
writes answer on paper
(mh :)
stands up
you see? starts to reach for paper
that's how you do your homework.
pulls sheet away from Sike and replaces it on box
okay. write your 19st name
193
Finally, in line 51 of Segment 4 (which takes off where Segment 2 ended), we

find the Samoan expression 'lla (abbreviated form of faapena meaning "like that"),
which we earlier described as a typical recognition of the fact that the task has
been completed (see earlier section, Emphasis on Task Completion). This
segment also shows more clearly the traditional Samoan pattern of instruction and
direct error-correction. The aunt guides Sike step by step, providing instructions
and close monitoring of his actions.
Segment 4
35
36
37
38
39
40
Sike
Aunt;
41
42
43
44
45
46
Aunt;
47
48
49
50
51
Sike;
Aunt;
sne.
Aunt;
okay . write your middle name, (2.0) write it small!

(4 .0)
(write ) (1.0) "(small)
continues to write while Aunt holds hand next to his))
((leans back))
"(ok a-) your last name,
write above this one . write here . 6
[
(?)
continues to write))
(don 't put-) no=no=no=no
(1.0)
over here (1.0) start "(it) here
[
pointing to place on the sheet))
(twritesj)
and write "T" over here
(twritesj)
OCT/d.)
'Like that.'
The homework activity is thus syncretic in that certain acts that comprise it orient
toward western modes of instructions, whereas other acts orient toward traditional
modes of instruction.
6An alternative interpretation of this sequence is : "right above this one. right here ."
194
Syncretism Within a Single Act

But syncretism may also characterize the construction of a single act within a
literacy activity such as homework. Particularly striking is the hybrid
construction (Bakhtin, 1981) there it goes in line 12, which is a blending of the
English there you go, said to a child who has just managed to do what was asked,
and of the Samoanja'ape1]a (or simply 1]a), "like that." The sense of recognition
found in the English there you go, which after all acknowledges the addressee
(you) as the successful agent of an action, is downgraded with the replacement of
you with it, which, once again, focuses on the activity rather than on the person
engaged in it.
Reallocating and Relocating the Task

As we discussed earlier, in a Samoan village, the organization of caregiving is
highly stratified and distributed: Given two potential caregivers, the more senior
or higher ranking one will expect the younger and lower ranking one to be the
active caregiver. In the interaction discussed so far, we saw this principle at work
in that, when both Sike's aunt and grandmother are around, it is the aunt
(younger) who attends to the child's needs (to finish his homework). As shown in
the next segment, the hierarchical and distributed model continues to apply when
Sike's cousin, Mata, comes to the scene.
A few minutes after the interaction just discussed (Segments 1-4), Sike comes
back with his second assignment sheet, which is about English prepositions. In
the meantime, the grandmother has taken control of the box and has started to fold
and carefully place the fine mats in it. While Sike is showing the homework sheet
to his aunt, his cousin, Mata, comes by and starts to follow the interaction. This
is the point at which Segment 5 takes off.
Segment 5
1
2
3
4
Aunt;
is holding a sheet of paper that he shows to Aunt

Mata comes outside and stands on doorway
reaches for paper
okay this one
[
5
6
Mata;
grabs paper with right hand, brings it closer to look at it

moves over to look at paper
Aunt;
8
9
10
11
12
Mata;
sue.
195
oh no this one you have to:(puts paper back on flap of box in front of Sik~))
you have to nllld.
moves behind Sike to be able to read))
well you havepoints to paper)) you go:(t) uhmm underline it.
When Sike seems to have a hard time explaining the assignment to his aunt, his
cousin , Mata, becomes more involved in the activity. In line 16, which follows,
she reads the relevant passage from the assignment sheet; in line 20, she indicates
where to find the relevant information (the picture of the cave); and in line 22, she
proposes the solution to the problem.
13
14
15
16
17
Aunt;
Sik~ ;
Mata;
18
Aunt
19
20
Mata;
21
22
Mata;
Sik~ ;
underline d' what?

bus or thepoints to three different places on the sheet))
(leans down and over to read))
reading)) "the bus will go in"
[
grabs paper and tilts it to have a better view))
tha:- bus ::-,
pointing)) right here .
(1.0)
the bus is going in the cave.
196
At this point , the aunt instructs Mata and Sike to get another box, implicitly
suggesting that they can go and work on it, creating a second desk, but the
grandmother this time explicitly sanctions the use of a new box by suggesting
that they use something different. The solution is a nearby bench, where Mata and
Sike are eventually sent by the aunt to complete the homework assignment (see
Figure 7.13).
Fig. 7.13. The two activities of packing mats and doing homework become
separate.
With this new spatial configuration in place, the two activities that earlier
coexisted in one location, sharing one tool (the box), are now divided. This is a
very traditional solution whereby the adults are free to continue with their work
(in this case, folding mats and placing them in boxes) while the older children in
the household take care of the younger ones. The syncretism is this time realized
through the blending of a western task (English homework) with a form of social
organization that is part of the traditional Samoan learning environments
(hierarchical division of labor in caregiving).
Conclusions
Our work in a Samoan-American community in Southern California has been
used here to make two general points about the relationship between
multilingualism and multiculturalism and, thus, to rectify what in our view
are common misconceptions about immigrant communities in the United States
and elsewhere. By concentrating on a homework task carried out in one of the
families in our study, we have shown that, in a multilingual , multicultural family
setting:
197
(a) language (in the sense of the specific code used at any particular time in an
interaction) is not always a good predictor of the cultural orientation or
interpretive frame that is being activated by the participants. Thus, in a
Samoan American family, English may be used fluently but in ways that
are consistent with the socialization practices typical of traditional learning
environments in the home country.
(b) Members of multicultural communities can be in more than one culture at
a time. This becomes apparent in the homework sequence analyzed here as the
same space and material resources are being used for two very different tasks
by different participants. Whereas the grandmother sees the space in the
backyard and the boxes placed in that space as needed for the accomplishment
of a task oriented toward a rather traditional aspect of Samoan culture (the
preparation and exchange of ceremonial Objects), her grandchild sees it as a
new location on which to accomplish his math and English homework
assignments. The child's aunt (and grandmother's daughter) accommodates to
both cultural orientations by mediating between them. She first transforms a
material object designed for one task (a box) into a tool for the other (a desk)
and then, when a younger potential caregiver arrives on the scene, redirects the
child to a different location so that she can accommodate to her mother and the
accomplishment of the traditional task.
An implication stemming from the concept of syncretism employed here is that
becoming an English speaker does not necessarily entail adopting strategies
characteristic of other groups who use English. Although many Samoan
American family members use the dominant language of the United States, they
may do so primarily at the level of the communicative code rather than at the
level of communicative conduct.
198
Appendix
Transcription Conventions
Aunt;
Speakers' names or kinship relation to the target child are

separated from their utterances by semicolons , followed by
a few blank spaces.
(1.0)
Numbers between parentheses indicate length of pauses in

seconds and tenths of seconds.
A square bracket between turns indicates the point at which
there is simultaneity of actions, either in the form of
overlapping speech or by concurrent actions (e.g., speech by
one person and gesture by another).
you have
to read
Underlining is used for emphasis , often accompanied by

higher volume.
(mh)
Material between single parentheses indicates uncertainty of

transcription.
"(oka-)
Material between parentheses preceded by a degree symbol

was uttered in low volume or whisper.
points
Material between double parentheses provides information

about bodily movements.
e leona
Samoan utterances and expressions appear in italics an d

reproduce the actual pronunciation. The letter g used in
Samoan orthography for a velar nasal (ng in other
languages) has been replaced by the phonetic symbol n.
(??)
Blank spaces inside parentheses with occasional question

marks indicate uncertain or unclear talk of approximately
the length of the blank spaces between parentheses.
no=no=no
The equal signs indicate latching, that is, no interval

between the beginning of one turn (in this case, a monosyllabic expression, no) and the next.
SyncreticLiteracy in a Samoan American Family
199
References
Anderson, A. (1984). Social and institutional influences on the development and
practice of literacy. In H. Goelman & A. Oberg (Eds.), Awakening to literac y (pp.
24-37) . London: Heinemann .
Apter, A. (1991). Herskovits 's heritage: Rethinking syncretism in the African
diaspora. Diaspora, 13, 235-60 .
Bakhtin, M . (1981) . The dialogic imagination: Four essays. (M. Holqui st, Ed.; C.
Emerson & M. Holquist, Trans .). Austin : University of Texas Press .
Bakhtin, M. (1984). Problems of Dostoevsky 's poetics (C. Emerson, Ed. & Trans.) .
Minneapolis: University of Minnesota Press .
Bauman, R., & Sherzer, J. (1975) . The ethnography of speaking . Annual Reviews, 4 ,
95 -119 .
Blom, J.-P ., & Gumperz, J. J. (1972). Social meaning in linguistic structures: Codeswitching in Norway. In J. J. Gurnperz & D. Hymes (Eds.), Directions in
Sociolinguitics: The Ethnography of Communication (pp . 407-434) . New York:
Holt.
Chavez, L. R. (1994). The power of the imagined community: The settlement of
undocumented Mexicans and Central Americans in the United States . American
Anthropologist, 96 , 52-73.
Cicourel, A. V . (1992). The interpenetration of communicative contexts: Examples
from medical encounters . In A. Duranti & C. Goodwin (Eds.), Rethinking context:
Language as an interactive phenomenon (pp. 291 -310). Cambridge, England :
Demuth , C. (1986). Prompting routines in the language socialization of Basotho
children. In B. B. Schieffelin & E. Ochs (Eds.), Language socialization across
cultures (pp, 51-79). Cambridge, England: Cambridge University Press.
Droogers, A. (1989). Syncretism: The problem of definition, the definition of the
problem. In J. Gort, H. Vroom, R. Fernhout, & A. Wessels (Eds .), Dialogue and
syncretism: An interdisciplinary approach (pp. 7-25) . Grand Rapids , MI: Eerdmans .
Duranti, A. (1981). The Samoan fono : A sociolinguistic study. Pacific Linguistics
Monographs, Series B, 80. Canberra: Australian National University, Department of
Linguistics .
Duranti, A. (1990) . Doing things with words : Conflict, understanding and change in a
Samoan fono . In K. Watson-Gegeo & G. White (Eds.) Disentangling: Conflict
discourse in Pacific societies (pp. 459-89). Stanford, CA: Stanford University Press .
Duranti, A. (1992). Language and bodies in social space: Samoan ceremonial
greetings. American Anthropologist, 94, 657-91.
Duranti, A. (1994). From grammar to politics: Linguistic anthropology in a Western
Samoan village. Berkeley and Los Angeles : University of California Press.
Duranti, A., & Ochs , E. (1986) . Literacy instruction in a Samoan village. In B. B .
Schieffelin & P. Gilmore (Eds .), Acquisition of literacy : Ethnographic perspectives
(pp. 213 -32) . Norwood, NJ: Ablex .
Duranti, A., & Ochs , E. (in press) Change and tradition in literacy instruction in a
Samoan American community. In B. McLaughlin, B . McLeod , & S. Dalton (Eds.),
Teaching for success : Reforming schools for children from culturally and
linguistically diverse backgrounds.
200
Eibl-Eibesfeldt, I. (1974). Similarities and differences between cultures in expressive

movements. In S. Weitz (Ed.), Nonverbal communication (pp. 20-33) . New York:
Oxford University Press.
Gal, S. (1987) . Codeswitching and consciousness in the European periphery. American
Ethnologist, 14(4), 637-653 .
Goodwin, C., & Duranti, A. (1992) . Rethinking context: An introduction. In A .
Duranti & C. Goodwin (Eds.), Rethinking context: Language as an intera ctive
phen omenon (pp. 1-42). Cambridge, England : Cambridge University Press .
Goodwin , c., & Goodwin , M. H. (1992) . Context, activity and participation. In P .
Auer & A. d. Luzio (Eds.), The contextualization of language (pp. 77-99) .
Amsterdam: Benjarnins .
Goody , J. (1977) . The domestication of the savage mind. Cambridge , England:
Goody, J., & Watt, I. (1968). The consequences of literacy . In 1. Goody (Ed.), Literacy
in traditional society (pp. 27-68). Cambridge, England: Cambridge University
Press .
Grillo, R. D. (1989) Dominant languages: Language and hierarchy in Britain and
France. Cambridge, England: Cambridge University Press.
Hanks , W. F. (1986). Authenticity and ambivalence in the text: A colonial Maya case .
American Ethnologist , 13(4) , 721-44.
Hanks , W. (1987) . Discourse genres in a theory of practice . American Ethnologist, 14
(4), 668-692.
Hanks, W. F. (1990). Referential practice: Language and lived space among the Maya .
Chicago: University of Chicago Press.
Heath , S. B. (1983) . Ways with words: Language, life and work in communities and
classrooms . Cambridge, England : Cambridge University Press .
Herskovits, M. J. (1937) African gods and Catholic saints in New World Negro belief.
American Anthropologist 39: 635-43 .
Herskovits , M. J. (1952) Man and his works. New York: Knopf.
Herskovits , M. J. (1966) The New World Negro: Selected papers in AfroAmerican
studies . Bloomington: Indiana University Press .
Hill , J., & Hill, K. C. (1986) . Speaking Mexicano: Dynamics of a syncretic language
in Central Mexico. Tucson : University of Arizona Press .
Hymes , D. (1962) . The ethnography of speaking . In T. Gladwin & W. C. Sturtevant
(Eds.), Anthropology and human behavior (pp. 13-53). Washington, DC:
Anthropological Society of Washington. (Reprinted in J. A. Fishman, Readings in
the sociology of language, pp. 99-138. The Hague : Mouton, 1968).
Hymes , D. (1974). Ways of speaking. In R. Bauman & J. Sherzer (Eds.), Explorations
in the ethnography of speaking (pp. 433-51). Cambridge, England : Cambridge
University Press.
Keating, E. (1994) . Power sharing: Language, space, and rank in Pohnpei.
Unpublished doctoral dissertation, University of California, Los Angeles .
Kendon, A. (1992) . The negotiation of context in face-to-face interaction. In A .
Duranti & C. Goodwin (Eds.), Rethinking context: Language as an interactive
phenomenon (pp. 323-334). Cambridge, England: Cambridge University Press .
Kroskrity, P. V. (1993). Language, history, and identity: Ethnolinguistic studies of
the Arizona Tewa. Tucson : University of Arizona Press .
SyncreticLiteracy in a Samoan American Family
201
Kulick, D. (1992) . Language shift and cultural reproduction: Socialization, self, and
syncretism in a Papua New Guinean village. Cambridge , England: Cambridge
University Press.
Lave, J. (1988) . Cognition in practice. Cambridge , England: Cambridge University
Press .
Lave, J., & Wenger , E., (1991) . Situated learning: Legitimate peripheral participation.
Cambridge, England : Cambridge University Press.
Leichter, H. J. (1984). Families as environments for literacy. In H. Goe1man & A.
Oberg (Eds.), Awakening to literacy (pp. 38-50) . London : Heinemann.
Macpherson, C. (1991) . The changing contours of Samoan ethnicity. In P. Spoonley ,
D. Pearson, & C. Macpherson (Eds.), Nga take: Ethnic relations and racism in
AorearoalNew Zealand (pp. 67-84). Palmerson North, New Zealand: Dunmore Press .
Mannheim, B. (1991) . The language of the Inka since the European invasion. Austin:
University of Texas Press .
McDermott, R. P ., Goldman , S. V., & Varenne, H. (1984) . When school goes home:
Some problems in the organization of homework. Teachers College Record, 85(3) ,
391-400.
Miller, P. (1986) . Teasing as language socialization and verbal play in a White
working-class community. In B. Schieffelin & E. Ochs (Eds.), Language
socialization across cultures (pp. 199-212). Cambridge , England: Cambridge
University Press.
Morgan , M. (1994) . The African-American speech community: Reality and
sociolinguistics. In M. Morgan (Eds.), Language and the social construction of
identity (pp. 121-48). Los Angeles: UCLA, Center for Afro-American Studies.
Ochs, E. (1982) . Talking to children in Western Samoa. Language in Society , 11, 77 104 .
Ochs, E. (1988) . Culture and language development : Language acquisition and
language socialization in a Samoan village. Cambridge, England: Cambridge
University Press .
Ochs, E., & Schieffelin, B. B. (1984) . Language acquisition and socialization: Three
developmental stories. In R. A. Shweder & R. A. LeVine (Eds.), Culture theory:
Essays on mind, self, and emotion (pp. 276-320). Cambridge , England : Cambridge
University Press .
Ochs, E., Smith, R., & Taylor, C. (1989) . Dinner narratives as detective stories .
Cultural Dynamics, 2, 238-57 .
Ochs, E., & Taylor, C. (1992a) . Family narrative as political activity . Discourse and
Society, 3(3), 301-340.
Ochs, E., & Taylor, C. (1992b). Science at dinner . In C. Kramsch (Eds.), Text and
context: Cross-disciplinary perspectives on language study (pp. 29-45). Lexington,
MA: D.C. Heath.
Ochs, E., Taylor , C., Rudolph, D., & Smith, R. (1992) . Story-telling as a theorybuilding activity . Discourse Processes, 15(1), 37-72 .
Rogoff, B. (1990) . Apprenticeship in thinking. New York: Oxford University Press .
Romaine, S. (1994) . Language standardization and linguistic fragmentation in Tok
Pisin. In M. Morgan (Ed.), Language in Creole situations: The social construction of
identity (pp. 19-41). Los Angeles: Center for Afro-American Studies .
Rosaldo, R. (1989). Culture & truth: The remaking of social analysis. Boston: Beacon
Press .
202
Sacks, H., Schegloff, E. A., & Jefferson, G. (1974) . A simplest systematics for the
organization of turn-taking for conversation. Language, 50, 696-735.
Schieffelin, B. B. (1990) . The give and take of everyday life: Language socialization
of Kaluli children. Cambridge, England: Cambridge University Press .
Schieffelin, B. B., & Ochs, E. (1986). Language socialization across cultures.
Cambridge, England : Cambridge University Press .
Scribner, S., & Cole, M. (1981). Psychology of literacy. Cambridge, MA: Harvard
University Press .
Shore , B. (1982) . Sala'ilua: A Samoan mystery. New York: Columbia University
Press.
Vo1oshinov, V. N. (1973) . Marxism and the philosophy of language (L. Matejka &
I.R. Titunik, Trans .). New York: Seminar Press .
Vygotsky, L. (1986) . Thought and language (A. Kozulin, Ed.). Cambridge, MA: MIT
Press .
Watson -Gegeo , K., & Gegeo, D. (1986) . Calling out and repeating routines in the
language socialization of Basotho children . In B. Schieffelin & E. Ochs (Eds .),
Language socialization across cultures (pp. 17-50). Cambridge, England : Cambridge
University Press .
Zentella, A. C. (1990) . Returned migration, language, and identity: Puerto Rican
bilinguals in dos worlds/two mundos . International Journal of the Sociology of
Language, 84, 81-100 .
Chapter 8
Other Voices, Other Minds:

The Use of Reported Speech
in Group Therapy Talk
Alessandra Fasulo
Dipartimento di Psicologia dei Processi di Sviluppo e Socializzazione
Universita degli Studi di Roma "La Sapienza," Italy
Voglio un ruolo importante in questo gioco di parole.
[I want an important role in this word-game]
Scusa i mancati giomi.

[Posthumous diary of a heroin addict]
Daniele Leandri
Abstract
This study analyzes the different types and uses of direct reported speech in the
context of psychotherapeutic discussions . The data are drawn from four sessions of
group therapy . The participants were 6 men, doing a program of recovery from
drug addiction, and a psychologist. The entire sessions have been videotaped and
transcribed .
Direct reported speech occurrences have been classified with regard to author,
content, and time location of the quote. An overarching distinction concerns the
fictional versus realistic character of the quotes. Data show that the therapist ' s
talk is strongly characterized by fictional quotes, whereas the p atients make a
more narrative use of reported speech . Specific function s of reported speech are
also examined in relation to therapeutic work, prop osing a view on identity and
self-understanding as based on intertextuality.
204
Alessandra Fasulo
Intertextuality , or heteroglossia, is, in the work of Bakhtin, the very feature of

discourse that makes it animate, alive. All words are hybrids; every utterance
borrows its flavor and takes its resonances from having been used within other
discourses. Thus, Bakhtin (193411991) brought to our attention the significance
of our use of the words of others:
The transmission and assessment of the speech of others, the discourse of another,
is one of the most widespread and fundamental topics of human speech . In all areas
of life and ideological activity, our speech is filled to overflowing with other
people 's words, which are transmitted with highly varied degrees of accuracy and
impartiality. (p. 337)
In his early discussion of reported speech, Voloshinov (192911973) , a student of

Bakhtin and probably a pseudonym for Bakhtin himself, treated it as a document
of the way another speaker's speech is received, which is assumed to be dependant
on social tendencies "crystallized into language forms" (p. 116). The reporting is
also said to be constrained by the context and aims of transmission and by the
receiver of this transmission. The forms used thus reflect patterns of
interrelationships among speakers, which are seen, at this stage of Bakhtin
theorization, as representative of the communicative ideology of a given historical
period.
Bakhtin's insights have proved useful and inspiring for subsequent studies
focusing on the relations between formal structures of language and its pragmatic
functions. Such an approach constitutes a twist in linguistic work that, especially
after Saussure, has analyzed language's structural features independently from its
enactment in concrete situations (cf. Lee, 1994).
As concerns reported speech, a distinction was classically drawn between the
direct and indirect form. The first type was defined as the introduction into one's
discourse of words that were authored by someone else, exactly reproducing them
as they were uttered originally. The second type consists of the report of the
content of someone else's statement, without any assumption of fidelity in form,
with transformations in tense, grammatical persons, deictic elements and verbal
modes of the original utterance according to the narrative context in which the
report takes place.
Leech (1978, quoted by Hill & Irvine, 1993) argued that, in both cases, the
reporter is responsible only for fidelity (of form in the direct style, and of content
in the indirect one). In this view, author and reporter of the reported speech remain
separate, but such a view does not always hold in light of actual examples .
Voloshinov (192911973), drawing on literary examples, presents a variety of ways
in which, both in direct and indirect reporting, the boundaries between an author's
and a narrator's discourses are blurred, and one meshes into the other, adding to
each other evaluating tones, comic effects, and so on. Hill and Irvine (1993),
discussing reported speech in interactive talk, maintained that, "because of
leakage, we prefer to say that reported speech . . . 'distribute[s)' responsibility,
Reported Speechin Group Therapy Talk
205
thinning out and socializing its central focus, rather than absolutely relocating it
at a distancefrom the animator" (p. 13).
It appears fruitful then to consider reported speech as a linguistic structural
resource co-occurring in the determination of meaning, particularly with regard to
ascribable intentionality anddegree of accountability of the spoken words. Rather
than assuming reported speech as detached from the reporter, we want to think of
it as a manipulation of the "otherness" of discourse, containing cues for the
interpretation of the speaker's stance toward the quote.
Direct Reported Speech

Directreported speech consists of a recognizable change of the speaker's footing
(Goffman, 1981), indicating that the words now coming out from the "sounding
box" (= the speaker) belong to an I that does not correspond to the I who is
talking. The shift can be signaled by a variety of indicators, possibly cooccurring. The most explicit can be an introductory sentence such as "He said,"
but there can also be code-switching (Gumperz, 1982) and voice alterations
(Macauley, 1987). Whendirect quotesslip into a context of indirect reporting (as
in Voloshinov's preset direct discourse), they can be indicated by use of
characteristic features of first persontalk,such as exclamations, discourse markers
(e.g., oh, well; Macauley, 1987), deixis andimperative forms, or by use of parody
(Bakhtin, 1991)and mimicry (Macauley, 1987).
Directreportedspeechallows operations in talk that would be dispreferred when
appearing in plain speaking. Self-praise and negative evaluations of others can be
accomplished by presenting them as coming from other people's mouths
(Macauley, 1987); affect canbe displayed covertly through intonational means and
rhetorical style of the quote(Besnier, 1993).
Reported Speech and Identity
Self-quotation can also occur. One's thoughts and speech acts, when framed in
reported speech, makefor a weakened form of reflexivity, that is, discourse about
a self that does not entirely overlap with the present one, incorporating a certain
degree of otherness.
In autobiographical narratives, splitting the omniscient narrator of the present
from the figure actingin the past can create suspense in the telling (Hill & Irvine,
1993), which sustains listeners' involvement (Tannen, 1988). In fact, if, on the
dimension of accountability, reported speech increases distance, it also creates a
kind of co-presence of quoting and quoted discourse, transporting speakers and
listeners in a different here-and-now(Blanche-Benveniste, 1991).
206
Alessandra Fasulo
The management of identity through reported speech is a central focus of this

chapter.
Data and Methods

The data used for this study are conversations occurring in a psychotherapy
setting. The participants are six men (19 to 29 years of age) living in a
community home in southern Italy where they were recovering from heroin
addiction,I plus their therapist, who had done his training in transactional
psychology. All the participants including the therapist were born in the area.
They were in a weekly group therapy program: I videorecorded 8 consecutive
sessions starting from the first time they met as a new group that had moved to
the mountain site where the home is located. The recordings have been transcribed
following Jefferson's (1985) method.
The setting had originally been chosen for the study of autobiography in
context (Fasulo, 1994a, 1994b). From a discourse analysis point of view, I was
interested in observing how life stories were presented and the work that was done
on them, within an activity that had talk about personal issues as its main focus,
and in a context socially (and physically) constructed as one of recovery and,'
'presumably, of identity reshaping. Finally, the visual angle of these persons on
their own lives was extremely peculiar. They were at that time in a sort of liminal
space, as described by Turner (1974); secluded from the rest of society, where they
expected to return with a different membership status, these people had abandoned
their life as heroin addicts, but their present everyday routine was, by definition,
temporary, and the future had uncertain contours . Therefore, my aim was to look
at how, in their interaction with the therapist and with each other, the participants
would recast (or simply redecorate) meaning and values of their experience.
The organization of participation was flexible. The standard pattern was of
separate interviews of the therapist and one participant, during which another who
wished to intervene was expected to ask for permission. Between one interview
and the next, or before the start, a participant could self-select for questions or
open up a new topic. Non-participating members were often called in by the
therapist while interviewing another member. Rounds of opinions on some issues
were also frequent, within or outside the individual interview.
Direct reported speech appeared to be abundant in this talk, used in various ways
ranging from repetition of parts of adjacent turns to long dialogue sequences. In
All subjects were there voluntarily, except for one for whom the community was
legally an alternative to jail. The others could leave when they wished, but a one-year
stay was considered the minimal time for a successful program. The members of the
group were the only residents of the community, without any control authority .
Reported Speech in Group Therapy Talk
207
the next section, I list the different types that were identified, trying to keep a
record of the discourse context in which they took place and the different "owners"
of the quoted words.
Initial letters refer to the fictionalized names of the participants : the therapist,
Antonio (T); Mauro (M); Luca (L); Nino (N); Francesco (F); Andrea (A); Daniele
(D).
Quotations are indicated by asterisks. The numbers in square brackets stand for
the number of the therapeutic session and the page of the transcript. The turn
where the quote begins is indicated by an arrow. The translation does not mirror
dialect aspects and jargon; intonation marks are reproduced partially to permit
comprehensibility.
Direct Reported Speech Types

(1) Simple repetition: Incomplete parts of a preceding turn, generally coming
unmarked in the turn immediately following the original utterance.
Excerpt 1 [2:3]
T
T-4
eh inizia da 'na cosa faciIe=faciIe va .

let's begin with something easy
quanti anni c'ha :i,
how old are you
cliciannove . Ii devo fa' a febbra :io diciannove.
nineteen. in february I 'll be nineteen .
ancora cliciotto cenn'ho.
I am still eighteen
*diciannove* . (1 .0) di dove sei?
*nineteen * . (1 . 0) where do you come from?
This form appears mainly to have a topic-tying function, but it can assume
evaluative nuances, such as, in this case, where the young age of Mauro is
repeated and followed by a significant pause.
(2) Self-quotation of therapy talk: A participant refers back to something he said
in the current or in a former session. Usually it is anticipated by introductory talk
that specifies the sense of its reappearance.
208
Alessandra Fasulo
Excerpt 2 [4 :21]
[The therapist turns to the other participants with a question asked him by
Nino]
T
N
dai va iniziamo il giro.

come on let's start the round.
ogmmo mi dica che ce sta di buono della comurii t.a
evezybody told me what is the good of the communi ty
no:: *che ce sta di buono* non hai capito
not *what is the good* you didn't understand
*di che cosa dobbiamo essere contenti*
*what we should be happy about*
[no non ho capito
no I didn't understand
[di stare in comunita* [O:Ltre- oltre al fatto
being in the community* beside- beside the fact
[bra :vo mgglio ancora
[good even better
Here Nino specifies the exact wording of his question, which had just been
challenged through a procedure of group judgment. Metapragmatic work of
explanation and redefinition is common in self-quotation occurrences. When this
is the case, they can be considered as articulated forms of self-repair (Schegloff,
Jefferson, & Sacks, 1977).
(3) Other-quotation of therapy talk: Words previously uttered by another
participant are again made to enter the circuit of discussion. They can be taken up
in the immediately following turn (as in the example) or reprised at a different
point of the discussion .
Excerpt 3 [3:12]
D
T
D
(1 .0) come quando uno pensa non 10 so de se ne i'=

like when one thinks- I don't know- of leaving=
=hai capi . chess' e chell'altro secondo me e sempre
you see- this and that- to me is always
il bambino che : che parla capi=comprime l'adu :lto.
the child who : who speaks y'see=presses the adult
allo :ra,
(then, )
"p a r che [e debo Le?
because he's weak
[bra : vo : Dan.i.e. questa me piace.
good Danie. I like this one .
209
*come quando uno se ne vuole andare*. piu 0 meno

*like when one wants to leave* . more or less
io l21!re cre :do siamo in una situazione ..i:mile.
I too believe we are in a similar situation
vale a dire( . . . )
that is to say ( .. . )
(4) "Doing the other": The speaker produces words on behalf of another. The
person to whom the words are meant to belong is the immediately preceding
speaker or the next one, after turn allocation.' The sequence positioning of this
kind of utterance makes it possible to read them as reformulations, anticipations,
or free interpretations of the other's thoughts or feelings. Often the quote is
produced in a parodic style and has a mocking spirit.
Excerpt 4 [4 :32]
[The therapist looks at Nino, who is sitting back on the sofa with his arms
folded]
Nino s'e chiuso ha ditt' *augh a me n' me fate

Nino has closed himself off he said *augh. me I won't
parla manco se me torturate*
talk not even if y ou torture me
(See also Excerpt 9, last line)
After the conclusion of an interview with one particip ant, the next speaker is
selected by the therapist. The direction of his look is a first indicator of the
choice, and a change of posture by the called person is common . The immobility
of Nino was thus a visible absence that got voiced by the therapist.
(5) Candidate utterances for another: This is a more complex way of enacting the
role of another participant. It is the offer of one or more propositions of which the
grammatical first person corresponds to the addressee, whereas the content comes
from the point of view of the speaker. I analyze this kind of reported speech in
more detail in the next section.
2When not otherwise indicated , all references to turn-taking organization are based on
Sacks, Schegloff, & Jefferson , 1974.
210
Alessandra Fasulo
Excerpt 5 [2 :15]
T
quello ehe rna:nea Luea e il diseorso del valore=eioe
what is missing Luca is the discourse of value=that is

~
*per me- per me io Luea cd.oe
*to me to me I Luca I mean

ehe valore ha anda a rubare?*
w.hat 's the value of robbing?*

L
non ha nessun valo:re .
it has no value.
Here the therapist offers Luca a question that, in his view, Luca should ask
himself. This excerpt also shows a systematic feature of candidate utterances,
namely, that they are a translation in reported speech of an immediately preceding
assertive statement by the same speaker (Fasulo, 1994b).
(6) Candidate utterances for himself: The quote belongs to the speaker himself,
but what he is presenting is an utterance that he is now constructing as a possible
one. This is made clear by the conditional mode of the introductory verb.
Excerpt 6 [4 :35]
T
ehO. io ho detto ehe con te, se non ti metti
in un atteggiamento-se non ~ntri in eomunita:

an at ti tude- if you don't come in the conrnuni ty:
allora pur'io mo me pigliavo la rnasehera=dieo *Anto
I said that with you, if you don 't take
well I too could take the mask=and say *Anto

sto lavora :ndo sto faeendo un bel programma
I am working I am doing a nice program

dentro di me sto eambia :ndo.
inside me. I am changing .

L
rna do l'hai vi :sti dentro eea? ste
but where did you see inside here? these ( ... )

ehe se mette Ie rnasehere .
w.ho pu t on masks .
The reported speech is framed here as the equivalent of putting on a mask. Because
Nino said he could do that too, Luca reacts, asking to whom he is referring. It
could be argued that Nino is reporting a sample speech of other members of the
group, but, because the reported words are accompanied by a negative evaluation,
he makes them up as a candidate utterance for himself.
211
(7) Fictional dialogues : One or more "turns'? are produced in a story-telling

fashion, but the depicted situations are hypothetical.
Excerpt 7 [2 :26]
ah nel sen [so quando
lYi
you mean when he)

[Andrea te 10 vi en e a ch i ed er e nella pausa
Andrea asks f or that during the break
dopa pranzo fino alle tre. invece de anda a dormi
after lunch 'til three. instead o f going to sleep
dice : [*Danie
he says : *Danie
[aspe
wait
invece d ' anda a dormi per'cha non me fai veda
instead o f going to sleep why don 't you show me
un disegno come se fa*
a drawing h ow I can do it *
(ub
D
T
The therapist here is testing Daniele's availability for helping the others. He does
so by a short hypothetical narrative (the setting is the after-lunch pause in the
community activities) in which a member of the group advances a request. Given
the exemplar character of the situation, Daniele's response represents an instance
of his general attitude.
Direct reported speech in these cases performs in the present of hypothetical
event, yet maintaining the abstractness of the argument.
(8) Anim ation of psychic entities: The child, the parent, or the adult are internal
unconscious instances, according to the psychol ogical theory adopted by the
therapist. Here they are represented as internal voices.
Excerpt 8 [4 :26]
il genitore e quello ch e ci dice *questo si fa questa

the parent is the on e who tells us *on e does this, on e
non si fa. questo 10 QUoi fare questo non 10 pu oi
does that. yo u can do this you cannot do that
fare =questo s e i in grado di farlo questa non s ei in
you are able to do t his y ou are n o t able t o do that*
grade di farl o.*
A turn is intended here not as the whole contribution of the current speaker, but as the
reported utterance of a single author or principal.
212
Alessandra Fasulo
(9) Narrative self-quotation and (10) Narrative other-quotation are the canonical
forms of quotation, where an utterance or a dialogue is presented by replaying
episodes that happened in the past. These types come within a specified temporal
frame but often are instantiations of prototypical episodes; therefore, it is not
fidelity with the original utterances that is central but the transmission of the
affect connected with the situation . As in the following excerpt, the verbal tense
is the Italian imperfetto (translated here with conjunctive), which is an iterative
form. Both types are illustrated in the following recalled dialogue.
Excerpt 9 [4:3]
L
no all'inizio : in comunita :
no in the beginning in communi ty- when I had
[appena entrato in comunita : me dicev'
just come in the conmunf ry they'd say
T
[eh in comunita si .
yes in the communi ty
N -7
*fai gygs:to fai quell'altro me la pijavo proprio:
*do this and do that* I was really bothered=
-7
dicevo *oh rna tu che vuoi da me rna tu :
I 'd say *oh what do you want but you :
perche aggi 'a fa' ~ta co :sa* capl*[~
why should I do this * y'see
[*che rni paghi *
T
*are you paying me? *
This is part of a typical narrative of the patients, which 1 have called the "I am
changed" narrative (Fasulo, 1994a). This implies the contraposition of two
moments that, when displayed in conjunction, show a change in the patient's
attitudes or feelings. Here Luca reports his initial discomfort (second arrow) with
the obligations of the community life, as represented by a generic older resident
giving him instructions (first arrow). The temporal reference is not just a neutral
landmark but also sensitive to the point of the story," in that it places the story
itself just after Luca's entrance in the community, which is also the beginning of
the therapeutic path.' The continuation of this narrative is presented in the next
example.
4Por a discussion on temporal and spatial references in conversation, see Sacks (1992)
and Schegloff (1972) .
5Luca had had four months of community life. Before moving into the home where I
found them, some of the members were staying in a different place run by the same
organization, although they had another therapist following them.
213
(11) Reported thought: This is a form of self-quotation referring to a mental

event. Packaging thoughts in a direct reported speech form permits one to lend to
the thought quoted the meaning of a punctual action.
Excerpt 10 [4:4]
T
L
[eh . e poi' ch'e carnbiato che-,

(eb. and then what changed thate carnbiato che : me so'fatto un resoconto dico: (0.3)
it changed that I have made myself a report I say
vedevo che: gli altri che lavoravano no?
I saw that the others who were working y'see?
dico *io mica posso sta : cosi senza fa' nignte .
I say *I can't stay this way doing nothing
lore non possono lavorare e io no perche :
It can't be tha t they work and I don't because:
l'aggia fa'* e penso che sto fatto cca
I have to do it*. I think this fact too
pure t'aiuta a crescere=t'aiuta: t'aiutera penso . (1.0)
helps you growing=helps you: it will help you I think .
This instance of reported thought testifies to a change that is narrated as a

punctual mental action (a realization), and a following resolution that led to a
behavioral change.
Comparing this excerpt with the last one, we can observe how the relation of
reported speech to the outside context whence it is supposed to come is
maximally dependant on the procedure by which it gets introduced, and specifically by the punctuative versus the iterative tense of the introducing verb. What
we have here is that a real dialogue has an undetermined duration or, actually, an
indefinite number of occurrences, whereas a thought is presented as a specific
event.
(12) "Generic voice" of a member of the group or a drug-addict: These are
instances of words expressing the point of view of an unspecified person, whose
only recognizable feature is that of being within the category of "member of the
community" or "heroin-addict."
214
Alessandra Fasulo
Excerpt 11 [4: 27]
F
N
Anto rna tu ti : hai trovato qualAnto but you- did you ever findcioe in percentuale no? gente che ha detto
I mean in percentage no ? people who said
[*no*no[una persona che vuole fa la comuru ta hehe
a person who wants to do the communi tj! hehe
aggio a anda vi :a- m'aggio fa due anni*
I must go I must do two years*
cioe gia dai primi quattro cinque mesi .
I mean from the very first four five months .
o generalmente: siamo tutti quanti titubanti
or generally: we are all doubtful about leaving
per uscire dice- siamo tutti quanti cosi vero?
one says- we are all like that isn 't that right?
Here the therapist is asked whether, in his experience, he has found some people
who were determined to complete the program from the very beginning of their
stay in the community. The operation is that of a normative confrontation (the
information requested is, in fact, a statistic) on the dimension of will as expressed
in public statements.
The cases that I have presented reveal that reported speech is used in the
psychotherapy situation in a highly situated fashion. Its pragmatic implicatures
and aspects of performance make it a useful tool for conveying affect, past and
present attitudes , and envisaged possibilities. Let us see how the different forms
are distributed in the talk of the patients and of the therapist. Asterisks mark the
types that are consideredjictional, that is, those that are not meant to refer back to
something that had already been uttered (or thought).
6"Doing the community" is a conventional saying that expresses well what is both an
ideology and an observation: Leaving the addiction behind is an act ive
accomplishment and not a simple change of residence.
215
Distribution of Direct Reported Speech Types (%)

Patients
* 1.
Therapist
Simple repetition
13,9
* 1.
Simple repetition
6,1
2.
Self-quotation within
therapy
2,4
2. Self-quotation within
therapy
3.
Other-quotation within
therapy
5,9
3. Other-quotation within
therapy
Doing the other
5,9
5.
Other candidate utterance
1,1
5. Other candidate utterance
6.
Self candidate utterance
2,4
6. Self candidate utterance
*4 .
33 ,7
*7. Fictional dialogues
* 8.
9.
Animation of psychic
entities
Narrative self-quotation
0,9
14,4
* 4.
2,2
Doing the other
14,9
8,8
38,7
12,7
*7 . Fictional dialogues
*8 . Animation of psychic
entities
4,9
9. Narrative self-quotation
0 ,5
10 . Narrative otherquotation
19,3
10. Narrative otherquotation
4 ,4
1 1. Reported thoughts
10,6
11 . Reported thoughts
0 ,5
1 2 . Generic member
6, 1
5,1
1 2 . Generic member
Realistic
quotations
Fictional
quotations
(*)
66,6
Realistic
quotations
33,4
Fictional
quotations
100
28,7
(*) 71,3
100
All the types of direct reported speech are available resources for participants in
both roles, showing that the two roles are performed not via thoroughly
differentiated verbal modes but via a difference in their rate of use. Patients and
therapist have a mirroring pattern in their use of direct reported speech, with regard
to fictional versus realistic types. As is predictable on the basis of the roledefining traits, the patients make much use of narrative quotations, including
reported thoughts, whereas the therapist shows a higher degree of offered candidate
utterances and other-quotation of therapy talk. The fictional type of reported
speech most often present in the participants' talk is that of fictional dialogues;
that is, they actualize possible problematic situations to be submitted under the
therapist's scrutiny. The simple repetition is also higher in the patients' talk .
Presumably, given that most of their dialogues are between one of them and the
therapist, it is the therapist's words that get embodied in their following
utterances and used as a point of departure of the following discussion.
216
Alessandra Fasulo
Exercises on Future
In his analysis of psychotherapy in radio call-in transmissions, Gaik (1992)
argues that the use of irrealis is an index of therapeutic activity, by which he
means a particular type of conversational move of the doctor. Through this mode,
a proposal is conveyed concerning the patient's internal states that the listenerpatient can decide to acknowledge or reject. In my view, the interesting feature of
therapists' utterances in the irrealis mode is that it occupies an intermediate zone
between the interactants: the joint outcome of the patient's telling of his story and
the therapist's interpretive tools. I think this is the domain of candidate utterances.
Among the therapist's candidate utterances are positive and negative ones, by
which we mean that they are presented to be accepted or disregarded, plus some
that take on a "Hamletic" form. The last ones are interrogative sentences in which
an alternative is posed, and they never appear in patients' talk.
Therapist's positive candidate utterance:
Excerpt 12 [4: 11]
T
~
sul futuro tu hai pote :re tu puoi farcela=

on the future you have power you can make it .
puoi decidere *voglio avere una famigliayou can decide *I want to have a familyvoglio avere una famiglia tranqui:lla
I want to have a quiet family
in cui si sta be:ne=che i miei figli non vivano
where one feels good=that 1l!Y sons won't live
illlftllo che ho vissuto io* questa 10 puo:i fare .
what I have lived* you can do that.
non e facile ,
it's not easy
speriamo
I hope so
Here is a candidate utterance that represents an act of will, introduced by "you can
decide": Note that this therapeutic action is not addressed to outline a possible
state of the world, because he is not saying "you can have a normal family" but
"you can decide *1 want to have a normal family* ." The action is messed
primarily to communicate to the patient his psychological capability of making
decisions and pursuing his own will.
Therapist's "Hamletic" candidate utterances :
217
Excerpt 13 [4:39]
ognuno lsi conosce I!lQltQ ~ dopa La comunita
everyone knows himself very well after the corrrnuni ty
[vabbe questa si
T
M
a.k . that 's right

T
[sa *io sto scivolando un ' altra vo:lta sto
knows *I am slipping again, I am doing

M
[questo si
that 's right

T
a fa' Ia stronzata . Ia vado a fa' Ia stronzata,

shit . shall I go and do it, or shall I go to
o m- 0 me ne vado a par.l,a co padre Domenico?
talk with father Domeni.cov' or with 1l!Y friends, or with
o cogli amici, 0 con chi,-* capito Mauro?
si questa e vero=so' ~:rdo

yes tha t 's true . I agree
who, - you see Mauro?
The same type of action in the previous excerpt is carried on here, but we see that
two options are presented, and the patient, Mauro, is provided with the illustration
of what it is like after the community. He is only in the third week of his stay, so
these words also have the meaning of a definition of the therapeutic work in terms
of what it does to the members. And again, what it does is not to rule out the
possibility that a person, once out, can feel like using drugs again . The focus here
is on the mental activity of pondering different options and acting according to a
decision. Reported speech again helps in describing internal processes as actions
to be performed at a given moment (when realizing one is "slipping" again and
just before doing it).
Patient's negative candidate utterances referred to others:
Excerpt 14 [3 :15]
F
penza a esse autonomo da te stess-non chiedere rnai
think of being autonomous by yoursel- don't ever ask

niente cioe non puoi dire
for anything you can't say

*io faccio Ia comunita tu rni compri Ia rnacchina*
*I go into the community you buy me a car*

A
no [s(h)e: :
no
A priest who is the founder of the community and to whom requests for help and
counseling are often addressed.
218
Alessandra Fasulo
Here one patient is describing to another what he deems a wrong pact with parents
and also a wrong approach to the community, namely, entering it under the
condition of some material reward.
Patient's negative candidate utterances for oneself: see Excerpt 6.
Types of Candidate Utterances (%)

Therapist
Patients
Se lf
Other
Posit ive
37,8
20,7
16,6
N eg ati ve
27,1
46,6
16,6
Hamletic
35,1
100
100
The different types of candidate utterances are homogeneously distributed in the

therapist's talk, but the patients show a higher frequence of negative self-referring
ones.
A possible interpretation of such a result can be in terms of altercasting
(Hewitt, 1976). This is a process of contrastive self-description through the
presentation of some other person's characteristic. Here, as well, through
unacceptable self-candidate utterances (that is, words that they are not saying), the
patients make clear what kind of person they are not, obtaining by contrast a selfimage profile that is acknowledged. Through such a procedure, though, the only
aspects that undergo judgment are those that are rejected, whereas the implicated
ones remain safely covert.
More generally, the regular exchange of candidate utterances (from therapist to
the patients and from patients to other patients and to themselves) reminds me of
the technologies of the self described by Foucault (1992) and particularly of the
stoical askesis, that is, "exercises that permit the subject to verify whether he is
able to cope with the events and to make use of the [teachers'] discourses that he
had memorized" (p. 32, my translation). The task is described as that of thinking
up the right words or arguments that one would use in challenging situations
from an ethical point of view. Negative events were a consistent part of what was
imagined, and it was crucial to think of them as "something actual and in course
of its happening. The procedure is that of keeping in consciousness, one beside
the other, the future and the present" (Foucault, 1992, p. 33). The stress is on
both the discursive nature of such technologies of self-domain and the necessity to
make up the imagined event as a "live" one. Candidate utterances (and in this
sense also fictional dialogues) are precisely discursive events that receive actuality
and closeness from the reported speech form, with the difference, in our case, that
the guide is present and participates in the exercise.
219
Trying on Identity
The continous word-sharing in the therapy situation also appears with regard to a
particular type of utterances that have the aspect of claims of identity. The
following is one of the most relevant, given what the group is doing :
(1) Therapist provides candidate utterance:
Excerpt 15 [4:43]
M
T
io quello- quello sto a di :,=

what I - wha t I am saying
=Mauro sai che significa=significa che qua :ndo esci
Mauro you know what it means=it means that when you
dalla comunita il- 10 rinnovi in ogni cosa che fai.
leave the community the- you renew it in anything you
*io non sono piu un tossicodipendente* e come se ogni
do *I am no longer a drug-addict *. It's as if every
volta dic- *io sono un ex-tossicodipendente* .
time you'd s- *I am an ex drug-addict * .
questa e vero che uno quando fa 'na cosa , se prima a
that's true that when one does something if before
facevi : (O .2) superficia :le=no? io so sempre state una
you did it: (0 .2) superficial=y' see? I have always
persona molto superficiale pero adesso posso cambia
been a very superficial person but now I can change
(
... )
(2) Participant recalls a past dialogue:
Excerpt 16 [4 :41]
M
T
N
allora noi rna qua non siamo usciti dalla ro :ba Anto,
then now we here are not out of the junk Anto,
no .
Daniele e io ieri stavamo a parla proprio di questa
Daniele and I yesterday were just talking about it
qua . *dopo due me:si sono gx tossicodipendente?*
*after two months am I an ex drug-addict?*
no Mauro . diciamocelo chi aro e tondo Ma' .
no Mauro . let 's be frank about it Ma'
220
Alessandra Fasulo
(3) Participant provides candidate utterance to the addressee:

Excerpt 17 [2:7]
F
si della situazione di una persona che : quando uno

yes about the situation of a person who: when one
dice eh: *quello- quello lA eh quello :
says uh: *that one- that one eh that one:
e drogato * *ero*.
is a junker* *I was*.
si: rna tu tu *ero* si rna tu non devi fa proprio
yes but you you *I was* yes but you don't have to
u discorso, ( .. . )
talk to them at all ( ... )
In Excerpt 15, the therapist stresses the need for adapting behavior with the claim
of being an ex-drug-addict, once out of the institution. In Excerpt 16, a participant
wonders about how long it takes to become entitled to utter the claim, and, in
Excerpt 17, a problem of labeling is dealt with by suggesting that they counter
the label by saying "I was."
The hot topic of being or not being a drug-addict is dealt with in an interesting
fashion: It appears as if the movement were from the utterance to the actor. The
utterance is "there" and circulates, but the actor has to work on himself in order to
become entitled to pick it up and use it when characterizing himself.
Identity claims can be tried on in a relatively safe environment until they get
attached to the person who "wears" them and are felt to belong to that person .
Conclusion
The analysis of reported speech in a group-therapy context shows that others'
words enter speaker talk at various levels of complexity, from repetition of some
words of an immediately preceding utterance to the alleged quotation that serves as
a point of departure for further discussion (where the utterance can be changed in
the process) and to the temporary substitution of one speaker with another.
In the microculture of six persons living together, sharing a past of drug
addiction and interacting with their regular therapist, some utterances appear
consistently throughout the therapy sessions, with a changing actor behind the I
who speaks. The fact seems to account for the primacy of discourse on single
individuals , the contemporary presence of different voices within one's repertoire,
and a conceptualization of psychotherapeutic work as the handing over from
221
therapist to patient a set of different voices that, entering into a dialogue with the
preexisting ones, can help patients on their way to change.
Certainly reported speech must be considered a socialized mode of expression
and a resource for pulling back to earth abstract concepts, especially in
interactional contexts where a gap in education is present among participants .
Nevertheless, other forms of reported speech are used without any problems in the
context I have observed. Therefore, I would not be satisfied with an explanation
invoking simplicity and rapidity at the basis of quoted speech usage. This appears
to be another piece of evidence in support of a view of the social person as a
crossing of possible discourses, none of which is truer than the other; on the other
hand, some can be felt as beyond one's grasp, and some have a limited horizon .
Such a theory of self works against the idea of monolithic individuals as much as
against an image of the person as the expression of gross sociological categories,
but it does not claim that we can say what we want: Doing "being normal," as we
have seen, can be no trivial undertaking.
Acknowledgments
The work presented here has benefited from discussion with the participants at the
workshop and especially from suggestions of further readings by Clotilde
Pontecorvo and Alessandro Duranti. Roger Saljo's revisions were also of great
help on both formal and content aspects. The remaining faults are due only to the
author's stubbornness.
222
Alessandra Fasulo
References
Bakhtin, M. M. (1991). Discourse in the novel. In M. Holquist (Ed.), The Dialogic
Imagination (pp. 259-422). Austin : University of Texas Press. (Original work
published 1934)
Besnier, N. (1993). Reported speech and affect on Nukulaelae Atoll . In J. H. Hill & J.
T. Irvine (Eds.), Responsibility and evidence in oral discourse (pp. 161-181).
Blanche-Benveniste, C. (1991) . Le citazioni nell' orale e nello scritto [Quotations in
oral and written texts]. In M. Orsolini & C. Pontecorvo (Eds.), La costruzione del
testo scritto nei bambini [The construction of written text in children] (pp. 259273). Florence, Italy: La Nuova Italia.
Fasulo, A. (1994a) . L'autobiografia. Modelli narrativi, contesti evocati e interazione
sociale nel processo di costruzione della storia personale [Autobiography. Narrative
models, evoked contexts and social interaction in the process of construction of
life-stories] . Unpublished doctoral dissertation, University of Rome "La Sapienza."
Fasulo, A. (1994b). La psicoterapia come traduzione tra linguaggi dell'esperienza
[Psychotherapy as translation among languages of experiences] . Rassegna di
Psicologia, 3(9) , 123-142.
Foucault, M. (1992) . Tecnologie del se [Technologies of the self]. In L. H. Martin, H .
Gutman, & H. Hutton (Eds.), Un seminario con Michel Foucault. Tecnologie del se
[Technologies of the self: A seminar with Michel Foucault] (pp. 11-47). Turin , Italy :
Bollati Boringhieri.
Gaik, F. (1992). Radio talk-show therapy and the pragmatics of possible worlds. In A.
Duranti & C. Goodwin (Eds.), Rethinking context (pp. 271-289). Cambridge,
Goffman, E. (1981). Forms of Talk. OXford, England : Basil Blackwell.
Gumperz, J. J. (1982). Discourse strategies. Cambridge, England : Cambridge
University Press .
Hewitt, J. P. (1976). Self and society. A symbolic interactionist social psychology.
Newton , MA: Allyn and Bacon.
Hill , J. H., & Irvine, J. T. (1993) . Introduction. In J. H. Hill & J. T. Irvine (Eds .),
Responsibility and evidence in oral discourse (pp. 1-23). Cambridge, England :
Jefferson, G. (1985) . On the interactional unpackaging of a gloss. Language &
Society , 14, 435-466.
Lee, B. (1994). Metapragmatics and subjectivities. In J. A. Lucy (Ed.), Reflexive
Language. Reported speech and metapragmatics (pp. 365 -391) . Cambridge,
England : Cambridge University Press.
Leech, G. (1978) . Natural language as metalanguage. Transaction of the Philological
Society of London. 1976-7, D, pp.l-31.
Macauley, R. K. S. (1987) . Polyphonic monologues: Quoted direct speech in oral
narratives. IPRA Papers in Pragmatics, 1(2), 1-34
Sacks, H. (1992). Story organization; Tellability; coincidence, etc. In G. Jefferson
(Ed.), Lectures on Conversation (Vol. II, pp. 229-241). Oxford, England : Blackwell.
Sacks, H., Schegloff, E.A., & Jefferson, G. (1974) . A simplest systematics for the
organization of turn taking for conversation. Language, 50, 696-735.
223
Schegloff, E. A. (1972) . Notes on a conversational practice: Formulating place . In D.

Sudnow (Ed.), Studies in social interaction (pp. 75-119) . New York: The Free Press.
Schegloff, E. A., Jefferson, G., & Sacks, H. (1977) . The preference for self correction
in the organization of repair in conversation. Language, 53, 361-382.
Tannen, D. (1988). Hearing voices in conversation, fiction and mixed genres. In D.
Tannen (Ed.), Linguistics in context: Connecting observations and understanding
(pp. 89-113) . Advances in Discourse Processes. Norwood, NJ: Ablex,
Turner, V. (1974) . Dramas, fields and metaphor: Symbolic action and human society.
Urbana: University of Illinois Press .
Voloshinov, V. N. (1973) . Marxism and the philosopy of language. New York:
Seminar Press. (Original work published 1929)
Chapter 9
Situational Effects in
Computer-Based Problem Solving
Paul Light' and Karen Littleton'
'Department of Psychology, University of Southampton, UK
2Faculty of Social Sciences, Psychology , The Open University, UK
Abstract
Possible bases of peer facilitation of children's learning are explored , drawing
upon concepts from developmental psychology. Against this background,
studies of social aspects of children 's computer use in school are reviewed . A
series of experimental studies is described in which pairs of children are found to
perform better than individual children when faced with novel computer-based
problem-solving tasks . Peer benefits are associ ated with patt erns of verbal
interaction in some degree, but facilitation by the presence of one or more
partners even occurs in the absence of direct verbal inter action. Gender
differences in performance are shown to be highly software dependent and also
prove to be sensitive to the gender of the partner, even (in fact, especi ally) when
there are no opportunities for inter action . The contextual sensitivit y of
performance is discussd in terms of situational constraints on cogniti ve
performance. It is argued that the important dimensions of context extend
beyond direct interpersonal interactions to encompass wider social processes.
One of the most salient aspects of the content of most human activities is the
presence or absence of other people. In the case of children' s use of computers in
learning , the presence of other children affords opportunities for collaborative
modes of learning . Computers lend themselves well to collaborative use. But the
Situational Effects in Computer-Based Problem Solving
225
presence of others in a learning situation impacts on that situation in other,

more complex ways, some of which are explored in this chapter.
The possibilities inherent in children's collaborative modes of working with
computers have in recent years attracted the interest of researchers working
within many different disciplines. Researchers with different backgrounds and
motivations inevitably bring different perspectives and methods to bear on the
issue. Educational technologists and software developers tend to be interested in
the practicalities and payoffs of collaborative modes of working and learning,
whereas psychologists are typically more interested in the processes involved.
But different types of psychologists also look for process variables of different
kinds. Developmental psychologists influenced by Piaget and Vygotsky, such as
ourselves, tend to focus on the process of overt (and typically verbal) interaction
among children working together at a computer. Piagetian influences tend to
favor examination of processes of conflict and conflict-resolution among equals,
whereas Vygotskian influences tend to direct attention toward the guidance that
the more able or experienced child or children can offer to others. Thus, these
approaches draw attention to certain aspects of the learning situation, but they
also tend to neglect others. In particular, these approaches have little to say
about motivational or affective dimensions of the learning experience or about
the significance of such processes as social comparison. In this chapter, we use a
series of experimental studies that we have conducted over a number of years
with various colleagues to argue that this wider spectrum of social-contextual
factors is highly relevant to learning outcomes in such situations.
Piagetian cognitive-developmental theory has often been accused of taking an
excessively individualistic stance. Piaget was not much concerned with the
specific influences exerted on development by specific kinds of experience, social
or otherwise . Moreover, he was often dismissive of any question about the
causal significance of social factors in children's cognitive development. Because
of the power relations inherent in adult-child relationships, adult interventions in
the learning process tended to be seen as irrelevant or even harmful, fostering
only nonconstructive processes such as imitation . By contrast, a more
constructive role was envisaged for symmetrical child-child interactions. Here,
the equality of status allowed conflicts of viewpoint to be confronted and
resolved constructively: "Criticism is born of discussion and discussion is only
possible amongst equals" (Piaget, 1932, p . 409) . Particularly in Piaget's early
writings, peer interaction of this kind was seen as a major factor in overcoming
egocentrism and achieving operational thought.
Much more recently, in the 1970s, researchers in Geneva embarked on a
program of experimental work designed to evaluate the impact of peer interaction
on the transition to operational modes of thinking in 5- to 7-year-olds (e.g.,
Daise, Mugny, & Perret-Clermont, 1975). Typically, in the first stage of these
experiments, children were pretested individually on some Piagetian task: for
example, a spatial co-ordination task or a conservation task. Then, on the basis
of these pretests, children in the experimental condition were placed in small
226
Paul Light and Karen Littleton
groups to work on similar tasks, whereas children in the control condition

worked on the same tasks alone. In a third stage, all children were given an
individual posttest.
Using this kind of design, Doise and his colleagues were able to show that
children of slightly different pretest levels, working together in dyads or triads,
tended to do better than children working alone . Moreover, this benefit carried
over to the children's individual posttest performances. Even children of exactly
the same pretest levels benefited from working together, provided they were
given a different perspective by virtue of their positions relative to the task array
(e.g., Doise & Mugny, 1984) . These benefits were interpreted as products of
sociocognitive conflict; the children's individual responses to the task were
conflicting, whereas the social situation provided the impetus for finding some
resolution of the conflict. The resolution of their partial or centered solutions
could only be found by adopting a higher level, more decentered solution, thus
ensuring cognitive progress.
One problem with this line of work is that the effects were almost too good to
be true . With conservation, for example, it seemed from some of the studies of
Perret-Clermont (1980) that a mere ten minutes or so of interaction with a peer
around a conservation problem was enough to shift a child from non-conserving
to conserving responses. It seems surprising, to say the least, that so brief an
encounter could overcome what, in Piagetian terms, is one of the major hurdles
in the way of achieving operational thought. Another aspect of the problem is
that children's responses to Piagetian tests, even when only a single child is
involved, turn out to be highly sensitive to logically irrelevant aspects of the
test situation (Donaldson, 1978). To take examples from our own work , it turns
out that rendering the transformation of materials apparently incidental to the
proceedings, or setting the task within the context of a game, can have a
substantially facilitative effect on children's conserving judgements (e.g. , Light,
Buckingham, & Robbins, 1979; Light & Gilmour, 1983; Light, Gorsuch, &
Newman, 1987). Sensitivity to rather subtle discursively created expectations
appears to be a hallmark of the later pre-school and early school years, leading to
real difficulties in establishing the validity of measures of the cognitive
competences of an individual (Light, 1986) .
An awareness of these considerations led Light and Perret-Clermont (1989) to
argue that, in the Genevan experiments on peer facilitation of cognitive
development, the presence of other children might have its effects not so much
by exposing children to conflicting points of view in need of resolution as by
changing the received meaning of the task. For instance, in the case of
conservation, social norms such as fairness of distribution are typically
introduced in a way that they are not in the standard individual task. This is an
aspect of what Doise (1990) has called socialmarking, and it serves to highlight
the fact that social processes may enter into learning situations at a variety of
levels.
227
Problem-solving situations may be social in the sense of involving direct,

face-to-face interactions between partners. However, even in the absence of such
direct interpersonal interaction, such situations may still be social in other
senses . The situation itself is set up by the experimenter, whose apparent wishes
and expectations may structure the child's activity. Other participants may be
implicitly present insofar as the child feels that his or her performance is being
compared with that of others who came before or after. Wider social norms or
conventions of etiquette may have bearing on the child's response. Moreover, the
actual presence of a partner may alter the implicitly social features of the
situation in a host of ways. Developmental psychological research on peer
interaction in learning has tended to focus on the direct cognitive consequences of
the availability of an interacting partner, rather than on less direct social and
motivational concomitants of the partner's presence.
At this juncture, we shift discussion away from research on children's
performance on cognitive developmental tasks in the Piagetian tradition to
consider more recent research on children's reasoning and problem solving when
working collaboratively with computers. Within this very different research
literature, we can see the same tendency to attend to the overt, explicit (and
especially verbal) dimensions of interaction while ignoring the subtler social
dimensions of the situations we create.
Children, Computers, and Collaboration

The potential of new information technologies to support learning has been
widely recognized, and educational institutions at all levels have invested heavily
in realizing such potential. Thus, systematic research on the factors governing
the effectiveness of computer-based learning is of considerable significance.
Moreover, because school children's computer-based work often takes place in
relative isolation from ongoing classroom or other activities, it is possible to
study such situations experimentally, without excessive distortion or
artificiality . Despite a good deal of early interest in using computers to
individualize learning, one of the clearest findings of recent research in this field
is that learning is often positively affected by the availability of working
partners . Both fmc-grained observational and case study reports (e.g., Griffin,
Belyaeva, & Soldatova, 1992; Hoyles, Healy, & Pozzi, 1992) and experimental
studies (e.g., Howe, Tolmie, Anderson, & Mackenzie, 1992) show clearly that
computers can and often do provide excellent environments for collaborative
modes of learning.
A general review of the social dimensions of computer-based learning is
provided in Light and Blaye (1990), and the recent European research literature is
well reflected in two recent journal special issues (Blaye, Light , & Rubtsov,
228
1992; Mevarech & Light, 1992). It is apparent from this research that working
with others at the computer can confer benefits, both in terms of group
performance and individual learning, although it is equally clear that it does not
always do so. In looking for explanations, researchers tend to concern themselves
with the factors that influence how effectively the participants are engaging with
one another and with the task. Interpretations are characteristically framed within
cognitive-developmental perspectives traceable to the influences of Piaget and
Vygotsky . Although Vygotskian concepts such as scaffolding have for the most
part been developed in relation to markedly asymmetrical (usually adult-child)
exchanges, they have also been applied in relation to peer learning interchanges
(e.g., Mercer, Phillips, & Somekh, 1991). The capacities of the computer for
reshaping the interaction among participants have also been discussed within a
Vygotskian perspective (e.g., Crook, 1994; Jones & Mercer, 1993).
Whether they are approached from a Piagetian or a Vygotskian standpoint,
studies in this area typically construe the benefits of the presence of a partner at
the computer in terms of overt processes of interaction, most frequently
involving discussion and negotiation. Evidence of a role for cognitive conflict
has been provided by, for example, Teasley (this volume), whereas Rubtsov
(1992) analyzes his findings in terms of the progressive coordination of joint
actions.
With various colleagues, we have been researching school children's computerbased problem solving for almost a decade. Our first studies of computer-based
problem solving by pairs (as against individuals) involved some rather standard
tasks, such as the Tower of Hanoi and the Balance Beam. Here, turn-taking was
the predominant pattern, and gains from having a partner present were only
apparent when rather artificial steps were taken to force the joint engagement of
the partners with each turn (Light & Foot, 1987; Light, Foot , Colbourn, &
McClelland, 1987).
Influenced by some of the adventure game software that was appearing in
primary school classrooms in the late 1980s, we designed some more elaborate
problem-solving tasks that arguably made better use of the distinctive potential
of the computer. Hypermedia authoring tools lend themselves particularly well
to the creation of rich nonlinear information bases in relation to which planning
problems can be posed. Such software also facilitates relatively straightforward
development of very user-friendly and highly motivating programs. Using
HyperCard on Macintosh computers, we produced a planning task couched
within an adventure-game format. In the first version (Blaye, Light, Joiner, &
Sheldon, 1991), the scenario involved a quest to rescue a king's crown. The
basic screen was a map showing various towns on a mainland, some sea (replete
with pirates), and various islands. Air, sea, and land routes were marked. All the
place names and other labels on the map were buttons. The only interface device
used was a single mouse. If this was used to move the cursor to one of the
buttons and clicked, additional screens of information were revealed. In this way,
children could discover the whereabouts of all the available characters and modes
229
of transport, together with the various constraints, on a workable route. The

pirates prevented the most direct route being used, although this was not
immediately obvious. This blocking of the direct route necessitated a detour
involving the use of an airplane. The children could access an action screen to
make moves, and the software, updated accordingly, kept a record of all
interactions between the user(s) and the system.
In our first study using this software (Blaye, Light, Joiner, & Sheldon, 1991),
we took either same-sex pairs or individuals from a class of eleven-year-olds and
gave them two sessions working on the task, followed by an individual posttest
with a slight variant of the task a week later. Assignment to the paired versus
individual condition was at random. We were interested in answering three related
questions. First, on a task of this type that makes heavy demands on children's
information-handling and planning skills, would we find that children in pairs
performed better than children working individually? Second, to what extent
would any advantage shown by the pairs carry over to subsequent individual
performance? Third, what psychological processes might be involved in any such
peer facilitation of children's performance in this type of task?
In the first session, very few children in either the individual or the paired
conditions actually solved the problem, but those in the pairs made significantly
greater progress toward task solution. In the second session, the pairs were
significantly more likely to solve the problem, and pair members carried. this
advantage over to the individual posttest session, even though this involved a
slightly different version of the task. At this stage, our interpretation was that
the paired-advantage might be associated with the adoption within the pairs of
distinct roles, which could be labeled driver and navigator. More generally, our
assumption was that the factors contributing to productive interaction would lie
in the observable patterns of interaction between the pair members.
So, with this in mind, we designed a second study (Barbieri & Light, 1992,
1993) with the aim of identifying important aspects of verbal interaction that
might underpin the pair advantage. Particular attention was paid to verbal indices
of the co-construction of knowledge, a focus that reflected our interest in
establishing whether such verbal interaction was associated with the performance
outcome of the pairs, either in terms of the joint achievements of the pairs or the
subsequent posttest performance of the individual ex-pair members. All the
children in the study worked in pairs for one session, then individually for a
second session. Once again, the second session involved a slight variant of the
task in which the initial dispositions of characters/transports were changed.
Interactions within the pairs on the first session were videotaped, transcribed, and
analyzed in terms of verbally explicit planning, negotiation, and so on.
Correlations between these measures and task performance, either in the paired
session itself or at posttest, were statistically significant but modest, accounting
for 10% to 20% of the variance. We also observed that boys out-performed girls
very substantially.
230
That the verbal interaction measures did not account for much of the variance
in performance or learning may simply reflect the inadequacy of the measures
used. On the other hand, it could be seen in rather different terms. In these
studies, unfamiliar adults were taking children from their classrooms to a less
familiar part of the school and presenting them with a very novel computer
(many had never used a mouse before) and a novel task. Given only scant
introduction, they were supposed to find out what had to be done and to work out
how to do it. It could be that the pair advantage in this situation reflects not so
much that the pairs were doing particularly well as that the individuals were
doing particularly badly under these conditions.
To address this, we conducted a further study (Joiner, Messer , Light, &
Littleton, 1992) in which we simply brought children to a room in the school
for a single session on the task, working on their own at a computer. The only
contrast was between children who (a) came on their own from the classroom and
worked alone (apart from an experimenter/observer) in the room and (b) children
who came to the room in groups of four and each worked individually at one of
four machines. In neither case was there any verbal interaction, and the children
could not see one another's screens. Nevertheless, the children in the latter
condition, with peers present, still did significantly better.
In subsequent studies, we have directly contrasted this parallel working (or
coaction) condition with a paired (or interaction) condition. Under these
conditions, we have found only a marginal advantage for the latter, which
disappears at individual posttest (Littleton, Light , Joiner, Messer, & Barnes,
1992). Thus, it seems that at least some of the benefits that we attributed in the
first instance to cognitively productive forms of overt interaction between
children working at the computer together may not, in fact, depend on such
interaction . Rather, they may depend, at least in part, on more covert processes
arising from participants' perceptions of the social situations created within the
experimental situation.
Meanwhile, we also had to consider a worryingly large gender difference in
response. One obvious possibility in relation to this is that superficial
characteristics of the software (namely, the scenario within which the task was
set) might make a difference. We recast the software in another version,
isomorphic in all respects with the first, but with a very different storyline . Here
some bears had gone for a picnic but had forgotten the honey, so they had to go
back for it. There were honeymonsters in the river, however, which necessitated
a detour by hot air balloon. The substitution of this version for the King and
Crown version made a dramatic difference. From being markedly inferior on the
King and Crown, the girls showed themselves as good or better than the boys on
the Honeybears (Littleton et al., 1992; Littleton, Light , Barnes, Messer, &
Joiner, 1993). We have made direct comparisons of boys' and girls' performances
on the King and Crown and Honeybears with children (N=48) all working one to
a machine in a coaction condition. Figure 1 shows the resulting pattern of
performance, which generated a significant gender by software interaction . We
231
Situational Effects in Computer- Based Problem Solving
repeated the study with a new sample of children and with an even closer
isomorphism of the two versions (the new King and Crown here being called
Pirates).
6
5
4
Girls
Boys
King and Crown
Software Type
Honeybears
Fig. 9.1. Levels of performance for boys and girls using either the King and Crown
or the Honeybears version of the task.
e
"3
:;j
'c.5
Girls
Boys
c
3 2
Pirates
Honeybears
Software Type
Fig. 9.2. Levels of performance for boys and girls using either the Pirates or the
Honeybears version of the task.
232
The results, as can be seen in Figure 9.2, show a close replication of the
previous findings. A not dissimilar pattern of responses also emerges from a
recently completed study (Littleton & Light, 1994) in which we looked at the
performance of girl-girl and boy-boy pairs working for one half-hour session on
either the Pirates or the Honeybears software. Here again, the girl-girl pairs
showed markedly superior performance on the Honeybears, whereas the boy-boy
pairs responded similarly to the two software types.
The reason for the girls' greater responsiveness to the Honeybears is difficult
to articulate precisely . It was apparent that many of them identified with the
characters, and afterwards they spontaneously talked about which bear was their
favorite. More than one talked about taking particular bears on their journey to
get the honey because "they wouldn't want to be left behind." This kind of
identification with the characters was not apparent among the boys or with the
King and Crown . It seems that these software hooks served to engage the girls
with the Honeybears' software. In themselves, these kinds of identification and
projection will not help to solve the task. Indeed, they could potentially hinder
solution. It is apparent, however, that, given an initial motivation to engage
with the task, the girls were every bit as capable of handling the interface and
thinking their way through the problem as were the boys. Whatever the precise
nature of the mechanisms underpinning the lability of the girls' response, the
results as they stand serve as a caution to all who develop software for children,
whether for use in the home or the school. Gender differences in response can be
massive, but they are far from immutable. The imagery or metaphors used in the
presentation of the task can have an influence out of all proportion to their
significance to the designer. Here, as elsewhere (e.g., Donaldson, 1978; Light &
Butterworth, 1992), context effects exert a critical influence on cognitive
performance and can affect not just the absolute difficulty of a task, but also its
relative difficulty for different groups of children.
In addition to the question of gender differences in response to computers, there
is the question of patterns of interaction between boys and girls while working
with computers . The issue of whether girls are disadvantaged by working with
computers in the presence of boys is one that has attracted a good deal of
attention in educational circles recently, and there is a significant move in the
United Kingdom toward segregating the sexes for computer-related activities,
especially out-of-classtime computer activities (Culley, 1993). Existing
empirical studies, however, show rather conflicting results . For example,
although Underwood, McCaffrey, and Underwood (1990) found evidence that
girls were disadvantaged by working with boys, Hughes, Brackenridge, Bibby,
and Greenhough (1988) found the opposite, with the girls actually having an
advantage in one study and with no differences in several subsequent studies.
Using the Honeybears, we have, in collaboration with David Messer and
Richard Joiner, recently undertaken a large study (N=120 eleven-year-olds) in
which we had children working either in coactive or interactive pairs, in girl-girl ,
girl-boy, or boy-boy pairings. The children were matched on the basis of an
233
individual on-task pretest , both within the particular pairs and across the
conditions . Overall, while they were actually working together, the children in
the interaction condition did slightly better than the children in the coaction
condition. This advantage disappeared, however, when the children were
individually posttested a week later on a slight variant of the task.
6
::IE
..~
<:
:3
::IE
Single Gender
Mixed Gender
Boys
Girls
Fig. 9.3. Levels of performance for mixed and single gender pairs for the
interacting pairs
Single Gender
Mixed Gender
Boys
Girls
Gender
Fig. 9.4. Levels of performance for mixed and single gender pairs for the coacting
pairs
234
In the interaction condition, the boys and girls did about equally well, and it
made no apparent difference to their performance whether they were working in
same or mixed gender pairings. In the coaction condition, however, where the
children did not have any opportunity to interact with one another, the mixed
gender pairings produced a marked and statistically significant polarization of
performance, with the girls in mixed pairs doing worse at posttest than girls in
other conditions and the boys in mixed pairs doing much better than boys in
other conditions. This result is illustrated in Figure 9.3 and Figure 9.4.
The absence of gender polarization of performance in mixed pairs in the
interaction condition was consistent with our previous findings using this
software. However, the discovery of such polarization in the coaction condition
was very unexpected. As a check on these results, we ran a further study, with
Stuart Bales, the results of which were reported briefly by Light (1994) . Using
the Honeybears software again, we compared the performance of pairs of elevenyear-old children working on the task during a single session under coaction
conditions. One third of the pairings were boy-boy, one third boy-girl, and one
third girl-girl. Once again , exactly as in the previous study, we obtained a gender
x pair-type interaction, with boys performing markedly better than girls only in
the mixed gender pairings. The effect that a number of educationalists have been
concerned about (namely, the fostering of gender differences in mixed gender
groupings) occurs quite markedly but only when the children are working
alongside one another without interaction. When they are actually collaborating,
the same polarization does not seem to happen .
We had not predicted these results, and explaining them is not straightforward.
We know from attitude questionnaires and interviews that the children subscribe
to the view that, in general, boys are more interested in and more capable with
computers than girls. But we also know that, with this software, the girls are
actually at least as good as the boys, and, in the first study , we had paired the
children so that (although they did not know it) the partners always had the same
pretest scores. We might speculate, then, that the opportunity for interaction
provides some feedback to the children on how they are doing relative to their
partner and that this feedback will tend to undermine their gender stereotypes. In
the case of the coactive pairs , however, these stereotypes may be evoked by the
situation of working alongside a partner of the opposite sex and cannot be
ameliorated by feedback. In the first of these studies, the two children did not
come from the same school class, so their knowledge of one another's abilities
was limited. Although the situation was not explicitly competitive, one could
imagine that the children would be anxious about their own performance relative
to their partner's, and that the threat of losing face might be particularly acute for
the boys in the mixed gender pairs. The absence of any feedback during the
session could only serve to heighten this process of social comparison.
These results obviously need to be explored further. But, for now, the message
for teachers might be that, instead of segregating boys and girls when working
with computers, they should foster as much interaction between the genders as
235
possible. The message for us as psychologists must be that the effectiveness of

peer interaction in learning situations, with or without computers, may be at
least as much a matter of social comparison as it is of social interaction. The
focus of both Piagetian and Vygotskian work in this field has been on
identifying the cognitively constructive elements of interaction. The results
discussed here suggest that a more social-psychological approach, focusing on
children's perceptions of their own ability relative to the task and to their
partner, may be at least as illuminating.
Interaction, Social Comparison, and the Negotiated

Meaning of Situations
The experimental studies described here are clearly very limited essays in
understanding what is going on when children work together on a problem using
a computer. However, experiments can sometimes serve to confront researchers
with the inadequacy of their initial conceptualization of a problem. The notion of
the individual child as a control condition for testing the effects of peer
facilitation is clearly a case in point. The individuals in our earlier studies
actually were not working alone but were working in the presence of an adult
experimenter. Even if they had been alone in the room, they would have been
there on the experimenter's terms, and the situation would have been rich with
socially induced expectation, anxieties, and so on. Similarly, the situations in
which the children are present are infinitely richer in meaning for the participants
than is suggested by our dry characterization of the experimental conditions in
terms of interaction, coaction, and so on.
Many of the important dimensions of the situation antedate any interaction
with the partner or with the problem on the computer. For example, despite our
injunctions to the contrary, we know that the children in our studies talked to
one another about the software outside the experimental sessions. This was
apparent because, later in the testing sequence, children not infrequently showed
knowledge of task information that they had not actually elicited from the
computer in the session. Perhaps not unconnected with this was the observation
that, in our studies with the King and Crown task, the gender asymmetry in the
mixed pairs was apparent, even in the seating positions that the children chose at
the beginning of the sess ion, before any introduction to or interaction with the
software. Boys tended to sit in the chair that gave them best (right-handed) access
to the mouse, which was placed directly in front of the screen.
Given this kind of observation and the rather consistent picture of gender
stereotyping in relation to computers among these children, the complete
elimination (and even reversal) of gender differences in performances by reversioning the software came as something of a surprise to us . We still have
236
some way to go to establish at what stages and in what ways the change in
software impacts on performance.
The most recent finding, that gender polarization in the mixed pairs is much
more marked in non-interactive than in interactive pairs, is also interesting, not
least in the light of current debates about the best way to redress gender
differences in the context of educational computing. The finding also serves to
highlight the centrality of processes of social comparison in these kinds of
learning situations. The contemporary work of Monteil and colleagues in
Clermont-Ferrand (e.g., Monteil, 1993; Huguet, Chambres , & Blaye, 1994),
which stems from social rather than from developmental psychology, illustrates
the extent to which children's expectations of their own success relative to their
classmates can impact on their learning. Huguet et al. (1994) have established
recently that an artificially raised expectation of success on our Honeybears task
can lead to enhancement of children's performance.
Developmental psychologists working in a number of fields are gradually
coming to appreciate the all-pervading nature of contextual effects on cognition
(e.g., Forman, Minick, & Stone, 1993; Light & Butterworth, 1992). One aspect
of this belated shift involves the recognition that the social context of a
cognitive task embraces not only direct interpersonal interactions but also the
social norms, expectations, representations, and comparisons that condition such
interactions. In the end, taking this wider sense of social into account in our
research may turn out to demand richer and more diverse research methods than
those used in the research reported here. Nonetheless, our rather tightly
circumscribed experimental approaches have perhaps justified themselves, at least
to the extent that they have served to highlight their own limitations!
Acknowledgments
The work reported in this chapter was supported in part by the Leverhulme Trust
and in part by the United Kingdom Economic and Social Research Council. We
would like to thank the Year 7 staff and pupils of Rickley County Middle
School, Bradwell Village Middle School, and Alex Campbell County Middle
School , all in Milton Keynes, England, for their support and cooperation.
237
References
Barbieri, M. S., & Light, P. H. (1992) . Interaction, gender and performance on a
computer-based problem solving task. Learning and Instruction, 2(3), 199-213 .
Barbieri, M. S., & Light, P. (1993) . Risolvere insieme un problema al computer:
Differenze di genere e del tipo di interazione [Problem solving together on the
computer: Differences according to gender and type of interaction]. In C .
Pontecorvo (Ed.), La condivisione della conoscenza [The sharing of knowledge]
(pp. 305-326) . Firenze : La Nuova Italia .
Blaye, A, Light, P., Joiner , R., & Sheldon, S. (1991). Collaboration as a facilitator
of planning and problem solving on a computer-based task. British Journal of
Developmental Psychology, 9, 471 -483.
Blaye, A., Light, P., & Rubtsov, V. (Eds.). (1992). Interactional learning situations
with computers. European Journal of Psychology of Education, 7(4), 257-372.
Crook, C. K. (1994). Computers and the collaborative experience of learning.
London : Routledge.
Culley, L. (1993) . Gender equity and computing in secondary schools: Issues and
strategies for teachers. In J. Beynon & M. Mackay (Eds.), Computers into
classrooms: More questions than answers. London : Falmer Press.
Doise, W. (1990). The development of individual competencies through social
interaction. In H. Foot, M. Morgan, & R. Shute (Eds.), Children helping children
(pp. 43-64) . Chichester, England : Wiley .
.
Doise, W., & Mugny, G. (1984) . The social development of the intellect. Oxford,
England : Pergamon Press .
Doise, W., Mugny, G., & Perret-Clermont, A-N. (1975) . Social interaction and the
development of cognitive operations. European Journal of Social Psychology, 5 ,
367-383 .
Donaldson, M. (1978). Children 's minds. Glasgow, Scotland : Collins .
Forman, E. A., Minick, N., & Stone, C. A (1993) . Contexts for learning:
Sociocultural dynamics in children's development. Oxford, England: Oxford
University Press .
Griffin, P., Belyaeva, A, & Soldatova, G. (1992) . Socio-historical concepts applied
to observations of computer use. European Journal of Psychology of Education, 7,
269-286 .
Howe, C., Tolmie, A , Anderson, A, & Mackenzie, M. (1992) . Conceptual
knowledge in physics: The role of group interaction in computer-supported
teaching . Learning and Instruction, 2, 161-183.
.
Hoyles, C., Healy, L., & Pozzi, S. (1992) . Interdependence and autonomy : Aspects of
group work with computers . Learning and Instruction, 2, 239-258.
Hughes , M., Brackenridge, A, Bibby, A, & Greenhough , P. (1988) . Girls, boys and
turtles. In C. Hoyles (Ed.), Girls and computers (pp. 31-39). London: University of
London : Institute of Education . (Bedford Way Papers No. 34).
Huguet, P., Chambres, P., & Blaye, A. (1994) . Interactive learning: Does social
presence explain the results? In H. C. Foot, C. 1. Howe, A. Anderson, A. K.
Tolmie , & D. A. Warden (Eds.), Group and interactive learning (pp. 445-451).
Boston : Computational Mechanics Publications
238
Joiner , R., Messer, R., Light, P., & Littleton, K. (1992 ). Peer presen ce and computerbased learning. Hatfield Polytechnic Research Report, PSY/92-10JOI. Hatfield,
England .
Jones , A., & Mercer, N. (1993) . Theories of learning and information technology. In
P. Scrimshaw (Ed.), Language, classrooms and computers (pp. 11-26) . London :
Routledge.
Light, P. (1986). Context, conservation and conversation. In M. Richards & P. Light
(Eds.) , Children of social worlds (pp. 170-190). Cambridge, England : Polity Press .
Ligh t, P. (1994, September). Social comparison and computer-based learning. Paper
presented at the International Conference on Group and Interactive Learning,
University of Strathclyde, Scotland
Light, P., & Blaye, A. (1990). Computer-based learning: The social dimensions. In
H. Foot, M. Morgan , & R. Shute (Eds.), Children helping children (pp . 135-137) .
Chichester, England: Wiley.
Light, P., Buckingham, N., & Robbins, H. (1979) . The conservation test considered
as an interactional setting. British Journal of Educational Psychology, 49 , 304310 .
Light, P., & Butterworth, G. (1992) . Context and cognition. London: Harvester
Press .
Light, P., & Foot, T. (1987). Peer interaction and micro-computer use. Rassegna di
Psicologia, 2/3, 93-104.
Light, P., Foot, T., Colbourn, C., & McClelland, I. (1987). Collaborative
interactions at the microcomputer keyboard. Educational Psychology, 7, 13-21.
Light, P., & Gilmour , A. (1983). Conservation or conversation? Journal of
Experimental Child Psychology , 54, 356-363 .
Light, P., Gorsuch, C., & Newman, J. (1987). Context and communication in the
conservation task. European Journal of Psychology of Education, 2, 73-82.
Light, P., & Perret-Clermont, A.-N. (1989) . Social context effects in learning and
testing. In A. Gellatly, D. Rogers , & J. Sloboda (Eds.), Cognition and soc ial
worlds (pp. 99-112). Oxford , England : Clarendon.
Littleton, K., & Light, P. (1994 , September). Gender and software effects i n
children's computer-based problem solving: Analy sing interaction . Paper
presented at the International Conference on Group and Interactive Learnin g,
University of Strathclyde, Scotland.
Littleton, K., Light, P., Barnes , P., Messer , D., & Joiner, R., (1993, March). Gender
and software effects on children's computer-based problem solving . Poster
presented at the annual meeting of the Society for Research in Child Development.
New Orleans, LA.
Littleton, K., Light, P., Joiner, R., Messer, D., & Barnes, P. (1992). Pairing and
gender effects on children 's computer-based learning . European Journal of
Psychology of Education , 7, 311-324.
Mercer, N., Phillips, T., & Somekh , B. (1991) . Spoken language and new
technology: The SLANT Project. Journal of Computer Assisted Learning, 7, 195202 .
Mevarech, Z ., & Light, P. (Eds.). (1992) . Cooperative learn ing with computers.
Learning and Instruction, 2 (3), 155-285.
Monteil, J. M. (1993) . Soi & le contexte . [The self and the context] . Paris : Armand
Colin .
239
Perret-Clermont, A.-N. (1980) . Social interaction and cognitive development in

children. London : Academic Press.
Piaget, J. (1932). The moral judgment of the child. London : Routledge .
Rubtsov, V. (1992) . Group work with the computer: The organisation of joint action.
European Journal of Psychology of Education, 7, 287-293.
Underwood , G., McCaffrey, M., & Underwood, J. (1990). Gender differences in a
cooper ative computer-based language task. Educational Research, 32, 16-21.
Part Three
Learning in Practice: How People and Tools

Shape One Another
Chapter 10
Discourse and Development:

Notes from the Field
Joseph Glick
City University of New York, Graduate Center, USA
Abstract
This chapter develops the view that workplace learning and, more generally,
development can be conceived of as comprising two different processes :
acquisition and coordination. Although many studies of learning and development
focus on the problem of acquisition, we develop the view that coordination is the
more central problem. Knowledge systems can be acquired, but, if they are not
coordinated with existing knowledge systems, there is little development - only
an accretion of more knowledge. Little is known about the conditions under which
such coordination can be optimized.
Empirically, this chapter examines some linguistic markers of conceptual
coordination, under conditions that we have identified as fostering coordination
of knowledge systems . In particular, we examine the discriminatory values of
several discourse-related measures , such as intrasentential code-switching. We
show that coordinated knowledge states show up in the way language is used and,
therefore, that linguistic markers not only are useful for producing cognitive
measures but also can be used to examine the means by which membership in
communities of change can be interactively produced .
244
Joseph Glick
From Principles to Practices

An emerging common conception unites many of the contributions to this
volume. This shared conception differs greatly from the conceptualizations of the
field , its problems and methods which have characterized mainstream approaches .
Although it is natural for authors of chapters in edited volumes to unite around a
common theme because that is generally the premise for their invitation to begin
with, what is unusual in this case is that the distance between mainstream and
current consensus is vast. Without heralding the beginning of a brave new world,
the conceptual distance between where we are here and now and where some of us
have been (and where the mainstream was and is) deserves, if not extensive
analysis, at least principled commentary.
Some key transformations or conceptual shifts serve to identify the distance
that has been traveled :
-
A shift from theorizing in terms of general principles to (1) an attempt to

identify particular practices and (2) a more specific focus on the what and
the where of such practices.
A shift from abstracted characterizations to descriptions and analyses of
people in settings and in groups.
Greater understanding and focus on the problems of how particular details
of setting (e.g., artifact structure, organization) constitute necessary
features for any analysis .
A shift from an individual activity to an interactional perspective.
And, most important,

-
A redrawing of the boundaries of what counts as a basic analytic unit. Do

we talk, for example, of three analytic units, (i.e., individuals, groups, and
settings), or are we better off talking in terms of a broader unit of
individuals in groups in settings.
A growing understanding that the features of the analytic unit, no matter
how broadly or narrowly conceived, are features that do not pre date
analysis but are emergent within the setting, and are interactionally
achieved in the course of functioning and hence must be recovered in the
analysis.
In short, we are more in the business of looking for self-constituting universes

and less in ,the business of looking for independent, intervening, and dependent
variables.
Perhaps this is only a personal fantasy of tran sformation, a dream that I and
some of my colleagues have lived that may not be common . So, instead of
speaking for all, I want to layout the natural history of a research program,
Discourse and Development
245
stretching from the mid-1960s to the present, with special focus on a research
program begun by Mike Cole, John Gay, and me in the late 1960s, and continued
and extended by Sylvia Scribner in the 1970s and 1980s, and currently being
extended by Cole and his colleagues and by this author and several students (Lia
DiBello, Elena Zazanis, and Jessica Kindred) who had started work with Sylvia
and have continued the work with me since she died. The theme of this chapter
then is "how we got here," and how we got from a traditional starting point to
share in the common conception just laid out. I take this approach with the belief
that our laboratory(s) have taken a journey that, in many ways, is characteristic of
some of the shifts in topic and perspective that the larger field within which we
work is going through.
Our general project had its origins in an attempt to understand the workings of
culture in human cognition. In the 1960s and early 1970s, we became involved in
a project that formulated this effort in terms of comparative studies: looking at
and comparing, for example, the performances on cognitive tasks of people in
Liberia with performances on cognitive tasks of people in the United States or
Mexico. Our hope was that comparative studies would enable us to make
orthogonal certain classes of variation that are inevitably confounded in
intracultural studies of development. For example, in a book authored by Cole,
Gay, Glick, and Sharp (1971) called The Cultural Context of Learning and
Thinking, we made a great deal of being able to disentangle the influence of
schooling and age, by working in an area where schooling was not yet universal
and where people could opt in to schooling at almost any age, because the school
system (as understood in Western educational terms) was relatively new. Thus, we
could compare first graders, for example, who were 6 with those who were 40.
Although this approach was clever, it brought more problems than solutions.
Working cross-culturally allowed us to get a handle on some independent
variables (and even here we had to put on some pretty restrictive blinders), but we
quickly became disenchanted with our experimental strategy which , at its heart,
involved the assumption that one could compare different groups on common
dependent variables.
The problem was that the tasks we used seemed to have different meanings in
the different cultures. We found that slight modifications of the comparative task,
which were made to make the task more culturally relevant, often led to large
changes in measured performance . Recognizing this as a problem, we began to
systematically vary tasks, either in materials used or in format of the experiment,
and we found that performance was not stable across the variations; people could
perform well or poorly depending on how we arranged the task. The classic
experimental formulation of the issues, independent variables whose effects were
measured by dependent variables, suggesting some intervening variables, seemed
inadequate to the problems we were facing. If we could make some independent
variables nicely independent, we tripped over dependent variables that were not
stable enough across their possible embodiments. And, if we found stable
dependent variables, this stability was often due to the operation of some
246
Joseph Glick
presupposed and hence unseen independent variable that was producing the
stability, for example, tasks whose performance demands mirrored the expected
performances that might come from within a cultural group for whom schooling
and the discourse practices and activities of the schooled were common points of
reference. For reasons such as these, we found that schooling seemed to be a
major variable in determining task performance. This turned out not to be because
of the influence of school as some abstract independent variable, but because of
the influence of schooling on the interpretation of the task itself.
The most trenchant discovery from this period of our research effort was that
generalities about cultural differences or the influence of such variables as
schooling were too loosely specified. The world did not neatly arrange itself into
variables whose impact could be traced in a general form . Indeed, this is a mild
way of saying that we discovered that what we wanted to discover was
undiscoverable in the terms in which we set out to discover it.
It soon became clear that we had to seek new terms of engagement. This
direction took us to the point of realizing that, rather than looking for variables to
provide the bird's-eye view that would allow us to compare, we had to find ways
of getting a more actor's-eye view that would get us closer to the concrete. Our
focus was shifted from variables to more microscopically analyzed situations and
performances within situations. Although comparison and grand explanations in
terms of variables was a lofty goal, it was a goal that had to await more engaged
under standings of particularities. We had to get closer to have even the chance of
getting to an overview of the big picture.
A critical way station on this march was embodied in work on the impact of
literacy on cognitive functions (Scribner & Cole, 1981). Rather than finding
generalized cognitive effects of being literate, Scribner and Cole found highly
specific effects related directly to uses of literacy by the person studied. They were
able to define different kinds of literacy and different uses for the different kinds.
Rather than having a general cognitive impact, the measurable effect of being
literate was directly related to the uses to which literate practice was put. It
seemed that, rather than looking for a generalized cultural/developmental variable,
one had to look at much more specific cases: that is, at what people actually did
and what it took to do those things.
In other words, the work had shifted from looking for general developmental or
cognitive rules, structures, or variables to looking for very specific grounded
cognitions or, more properly, to looking at cognition that was situated in real
activities (Lave, 1988; and see McDermott, 1993, for an update of one of the
seminal documents of this period, a paper by Cole, Hood, and McDermott entitled
"Ecological niche picking.")
247
Practicing Practices - What are the Principles?

There are deeper issues underlying the transformation of approach that I have just
outlined. We had begun our search as more or less traditional, developmental
psychologists. We had problematized the issues in child development terms,
seeking to account for development by using cross-cultural research as a way of
disentangling developmental variables. As we began to engage the phenomena of
situated and grounded cognition, however, it became clear that this perspective
was unduly narrow.
Indeed, I will argue that the developmental model and the situated cognition
model are vastly different models of psychological functioning, and that each
model has its place in psychological analysis. They should not be conflated,
however, and every effort must be made to examine the assumptions underlying
each model, and every effort must be made to understand the proper domain of
application of the models.
I have tried to set out the differences between developmental and situated
cognition models in the following.
ASSUMPTIONS ABOUT KNOWLEDGE: THE COGNITIVE

DEVELOPMENT VERSUS SITUATED COGNITION MODELS
Cognitive
development
Situated
cognition
Knowledge as general and

abstract
Knowledge as applied to the

particular and grounded in practice
Knowledge is carried
with the person
Knowledge is applied within an

environment and shaped to and by
that environment
Knowledge must be general

because the future is unknown
Knowledge must be specific

because the problem is specific
Notion of multi formed future

- Understanding that the
tasks of school/family
are preparatory for some
unknown future destination
Notion of rationalized future

- Progress will be organizationally
defmed
Defined by flexibility
- Generalized skills
- Transferable
Defined by efficiency
- Specialized skills
- Maximally adapted to local conditions
248
Joseph Glick
As you can see, the basic assumptions of a child development-centered approach

to cognitive development privilege a theory of a particular kind of knowledge, the
kind of knowledge that is the object of a basically idealistic epistemology. What
counts as knowledge is the sort of abstract knowledge systems that might be
considered as, in principle, at a distance from any local knowledge or "tricks of
the trade." In contrast, the assumptions of the situated cognition model reverse the
emphasis. Focus is placed on how knowledge is applied in a given setting and
how knowledge is tailored to that setting . In contrasting the underlying
assumptions, it becomes immediately clear that the boundaries between individual
and setting are quite different in the different models . Developmental theories tend
to be atomic, whereas practice theories become molecular. The atom in
developmental theory is the model of knowledge as abstract and generalizable in
principle and, hence, recombinable and useable in any number of different
settings. The molecule in practice theories reflects the idea that the basic unit of
analysis must be one that includes the setting as an essential part of the definition
of knowledge. The atom can be treated as disengaged from its surroundings,
whereas the molecule gains its essential properties from the surround within
which it is embedded. Indeed, within the situated cognition model, knowledge can
be considered to be as much a part of the environment as it is of the individual.
Scribner's dairy studies were an early attempt to exercise the new research
paradigm. In those studies, Scribner (1984, 1986) focused on the area of
mathematical operations and job activities that would necessarily involve some
sort of mathematical operations. For example, she looked at the ways in which
dairy delivery men filled their trucks for an early morning delivery run . By careful
observational analysis, she found that many shortcuts were used that involved the
use of mathematical operations mediated by artifacts. The operations were used
very differently from the way in which mathematical problems might be solved in
a paper-and-pencil, school-like, mathematical practice. Workers took advantage of
the fact that dairy products are routinely packaged in cases of a given size. The
amount that a case would hold allowed for short-cut perceptual strategies that
supplanted using numbers . For example, if a case holds 12 bottles, arranged in 4
"rows" and 3 "columns," it is easy to see when there is half a case (2 rows filled
and 2 rows empty). If the delivery man had to prepare an order with 6 items , it is
easier to fill half a case than to count 6 items, and so forth. Similarly carpenters
simplify the Pythagorean theorem into the "3-4-5 rule," which expresses the
Pythagorean theorem without the squaring and reduces it to a specific formulation
adapted to an expectable environment. Although it may not then be a general and
abstract rule, this rule is fitted to carpentry practice. It seemed that, although it
took a lot of work to find the engaged cognitive operations, the operations found
obeyed a least effort principle, with head being substituted for hand. But, the
substitution of the intellectual (head) for the manual (hand) required, in turn, an
environment with organized artifacts that could support the substitution. Thus,
the intellectual knowledge that was studied depended heavily on details of setting
for its possibilities of actualization .
249
The dairy studies led Scribner to expand her research effort to looking at other
work sites for similar embeddedpractices and cognitive principles. The research
strategy employed was one that began with ethnograph ic observations of selected
work episodes, which yielded candidate practices that seemed to embody some
cognitive rule. The observational data would then be followed by quasiexperimental tests that simulated the work practice but with certain experimental
controls to test the limits of the rule that was thought to have been found.
As the engagement with various work sites expanded, Scribner turned her
attention to an increasingly common phenomenon in the workplace: The
introduction of technologies that transform the nature of work itself. What would
happen to the local knowledge that skilled workers had when a new form of
knowledge was demanded within the workplace? What are the traceable
relationships that exist between expert knowledge (of the workers) and expert
systems (of the machines)?
When she died, Scribner had been working on this problem in two factories that
were undergoing such technological transformation: a primary site, here called
Kemps, and a comparison site, here called Intek. A cornucopia of topics was
under study, including the following: the manner in which practical knowledge of
stockroom practices is communicated by more expert workers to less expert
workers, according to what principles and involving what operations; did such
cognitive operations as thinking in terms of ratios , which was involved in the
common stockroom practice of "counting by weighing," change when new kinds
of scales got introduced in the workplace? ; and, generally , what was the specific
orchestration of skills that was necessary to the performance of particular jobs,
and how were these related to the structure of the physical artifacts and conditions
that applied in the workplace?
The general thrust of these studies was an extension of the earlier work on dairy
workers, which might best be described as compensatory. The studies of
cognition in practice represented an attempt to identify, in people who are often
given the social status of low-level workers in work settings, sets of important
cognitive abilities involved in doing the job , which, once identified, might
contribute to a rethinking of such fundamental social/occupational classifications
as the division of labor between mental and manual components and the
accompanying class/occupational structure that went along with it (e.g., the
division of employees into hourly and salaried and assumptions about what kinds
of prerequisite training employees should have for different job classifications).
These were all matters not only of great psychological interest but also with
enormous social and educational implications .
These issues assume even greater importance in periods of great social and
organizational change. And, if it was nothing else, the time period within which
the research was being conducted was a period of great social and technological
change . As the research unfolded , we found ourselves being drawn more and more
to the issue of change. During the mid-1980s, studies began to examine the
impact of changing local technology, such as the impact of the move from
250
Joseph Glick
mechanical scales to electronic scales (Beach & Zazanis, in preparation, Zazanis

& Glick; 1993) from hand and eye machining to Computer Numerical Controlled
(CNC) machining (Martin & 'Beach; 1992).
It became increasingly clear that the simple search for embedded cognitive
processes or situated cognition was telling only a small part of the cognitive
story . The moment that one shifts from looking at situated cognition in a stable
environment, to looking at situated cognition in an environment of change, new
features needed by an analysis begin to emerge. In the remainder of this chapter, I
would like to give you a taste of the kind of work that we are currently doing,
which focuses particularly on the theoretical implications of research that looks at
embedded cognitions in an environment of change.
From Practices to Communities of Practice

After several years of working at our primary field site (Kemps), we began to
shift attention from the local technologies embedded in the performance of
particular jobs (e.g., counting by weighing or pulling orders in the stockroom),
to some of the larger scale changes that began to characterize overall operations in
the plant.
In keeping with changes that have swept American industry under the new
conditions of global competition, Kemps had implemented a production control
system that sought to rationalize work practices and, in particular, to rationalize
the use and stocking of materials by use of a Materials Requirement Planning
(MRP) system. The MRP system is inherently a management tool that utilizes
the number crunching capacities of computers to plan in great detail, and, to a
level heretofore unattainable, the details of production. Systems of this sort have
come into increasing use in American industry as manufacturers have had to shift
from a mass production logic to small batch production logic to meet
increasingly differentiated consumer desires (Sabel, 1991). This shifted the logic
of production from mass quantities to special orders, which, in turn, require close
control over materials inventory .
The requirement for such close control extended the hand and logic of
management down to the shop floor level. As Sylvia described it in her last
paper, orally delivered at the Activity Theory conference in Lahti, Finland,
MRP is a theory of the management of manufacturing and the pl anning ,

organizing and control of production - precisely those functions whose
separation from the doing of production was the historical basis for the division
of so-called intellectual from so-called manual labor. Therefore, in addition to this
being a program which indeed takes on very important functions with respect to
251
production, it is an ideology. Although management theories were formalized in

some respects before the advent of computer systems, they were integrated in to
management practice, if at all, using technical means that severely limited the
capacity of theory to influence practice . Empirical knowledge systems dominated
management practice. MRP and similar systems allow theory to regulate practice
more directly. (Scribner, 1990)
Four principles underlie this description, all of which are critical to our analysis.
They are set forth in the form of the following assumptions:
(1) That the logic of management-guided production and the logic of production
systems as they exist in practice are different;
(2) That management systems had previously been unable to control production
systems , but that, with the increasing availability of computers integrated into
manufacturing, the control aspect of management practice is increasingly
possible ;
(3) That the embedded rationalities of production will have to meet, at some
point, with the ideological rationalities of management control;
(4) That progress in this environment of change will require either the total
supplanting of one system by the other, or that a point of integration must be
found.
Much would depend on the degree of fit between the system of production as it
existed and the new demands of the rationalized production system.
In essence, the logic of the MRP system is quite different from the logic of
production on a number of key points. Rather than thinking of manufacturing
production as a process that moves forward in time - from available materials,
upon which labor and machine processes are applied to yield a finished product MRP thinks of manufacturing as a process that moves backward in time from
customer orders. Thus, the system plans backwards. In order to do so, it needs
several key pieces of information. It needs to know the structure of manufactured
items so that it knows the constituent parts of finished items and hence can plan
manufacturing down to a relevant level of detail. It must also know the process of
part assembly so that it can plot what comes before what. This is represented in
an indented bill of materials that outlines the relations among the parts in terms
of a family structure, that is, of parent-child relations among raw parts and the
sub-assemblies into which they go. Once the system knows the item structure
and the parent-child relations, it can then introduce, into planning, the concept of
lead times for the acquisition of materials, so that the various parts come together
in an assembly process just as they become needed .
The logic of the MRP system introduces a new language for describing things.
Thus, a finished item, which we are likely to call "the thing that is made," is
redescribed as a "0 level item" that is composed of "levels" beneath the "0 level"
252
Joseph Glick
that represent the necessary parent-child relations of assemblies to sub-assemblies

and to constituent parts of sub-assemblies.
Once the system is armed with (1) its bills of materials; (2) level labeling
processes; (3) parent-child relations (which, for those of you who like the arcane,
can be "pegged up and down" to trace family relations of the parts involved; and
(4) lead times for parts acquisition, it can then apply a fairly rigid logic to yield a
number of "planned order recommendations," which are recommendations for what
to order when. But these recommendations are often unrealistic and fatally
unconnected to the reality of factory life. Indeed, the recommendations are
theoretical in both the best and worst senses of the word.
In order to gain analytic power, certain assumptions are left out of the materials
ordering process . For example, the machine has no knowledge of the outside
world. It does not know whether there is bad weather and/or bad relations between
the part supplier and the plant. It does not know whether there might be a strike.
MRP assumes that materials are the only things to be planned and can therefore
make recommendations that are unrelated to either machine capacity or
availability of labor hours. It assumes a work-force that can be hired and fired in
relation to fluctuating demand and a parts supply relationship that is on demand
(given lead times). Its cycles for making decisions are not long term enough so
that they could be rationally connected to business flow. The computer-generated
recommendation can fluctuate on a day-by-day, or order-by-order basis, assuming
that the work force is a completely malleable item. This outcome of the strict
logic of computation could prove fatally unrealistic in a workplace that used, for
example, skilled employees who would quickly seek other employment if they
could not know when they would be working, or who faced alternating cycles of
no work and 24-hour days. The system could not work where there was a civil
service workforce who could not be made dispensable. In short, only fools would
slavishly follow the computer-generated recommendations. Human judgement
must be inserted into the equation to evaluate the recommendations against
everything else that is known about the actual practices of real business with real
labor forces in a real world. MRP-II is an extension of the MRP system, which
attempts to include capacity planning into the program, but it too has its own
blind spots that require the same sorts of human evaluative judgment.
In order to evaluate properly the recommendations of the computer program, a
user of the system must fully understand the logic of the system in order to see
where it can lead to unrealistic results . The informed user, must , as well,
understand in detail the realities of the workplace about which the computer
program is making predictions. This process of integration between workplace
logic and c~mputer logic is greatly aided if the MRP system is tailored, in its
fundamentals, to those aspects of production that can be smoothly integrated.
Specifically, this would mean, at least, (1) that indented bills of materials follow
the logic of parts assembly, (2) that the calculation of lead times for supply
acquisition is historically accurate, and (3) that there is sufficient forecasting in
the system to buffer the day-to-day calculation of demand so that planning can
253
occur over longer historical cycles and thus allow for non-catastrophic
adjustments of work flow.
Typically, people engaged in production know the reality of their businesses.
They have real relations with real suppliers in a real industry in a real plant. What
they do not know is the constraint on the logic of the computer program and the
many ways in which it can be unreal. Thus, it is not surprising that such systems
work to a greater or lesser degree across implementations. In some cases the
computer "fouls things up" and becomes progressively less used, or, more
precisely, ignored, while in some cases it helps a manufacturing plant to become
more smoothly efficient in its operations.
Thus, the MRP system, although a powerful analytic tool, must be harnessed
and adapted to each work environment into which it is placed. According to
industry manuals (e.g., Wallace, 1990), the key to a good implementation of
MRP is a rather long period of implementation that operates under the following
major constraint:
Implementing MRP-II properly requires a great deal of time and effort on the part
of many people throughout the company : Data must be made more accurate ,
people must be educated and trained, new software must be acquired and installed,
new policies and procedures must be developed and made operational....
Successful implementations are done internally. In other words, virtually all of
the work involved must be done by the company's own people. The
responsibility can't be turned over to outsiders, such as consultants or software
suppliers. That's been tried repeatedly , and hasn 't worked well at all. (Wallace,
1990, p. 13)
The composition of an internal implementation team is a critical and not well

understood aspect of the implementation process. The team is not only critical for
the technical issues of system implementation, but it is critical from the point of
view of the organization of the workplace as well. As Scribner (1990) pointed
out, the MRP system instantiates a philosophy of management control. With the
added ability of the implementation team to write "new policies and procedures,"
the impact of the MRP system on the organization of the workplace can be
massive. And, indeed, the representation of the work process, as instantiated in
the details of system implementation , can, under many circumstances,
misrepresent what people actually do when they do their jobs . This can have
disastrous consequences for the value of the system and perhaps even for the
perceived value of the job.
For such reasons, it is sometimes understood that the implementation team
should represent a spectrum of those who will interface with the system.
Although there is considerable variation in how broad this spectrum turns out to
be, the process of implementation has the potential for introducing a new
dynamic into the workplace by placing people in new relationships to one
254
Joseph Glick
another. It is often observed that the implementation process creates a new

cultural entity within the workplace. For want of an agreed-on term, we can call
this new entity a community of innovation.
A community of innovation, which grows around the introduction of a new
technological system into the workplace, can serve a critical role in reorganizing
workplace social relations and the workplace itself. The process of technological
implementation provides new opportunities for status and advancement that might
not otherwise be available. It also provides the possibility of interacting across
levels of a workplace hierarchy as the member of a team. Thus, it can flatten
hierarchies. But, beyond these social and status gains, critical aspects of job
design are also at stake.
Workers recognize both the real and the semiotic power of computers to
influence their lives within a work organization . They recognize, too, that most
implemented computer systems reflect, to a greater or lesser extent, a
management philosophy and a managerial view of rational operation. They know
that they can easily become the "objects" to which these rationalities apply . They
know, as well, that they are often seen as the source of problems that the
computer system is intended to fix. Thus, in many workplaces , the major
question gets framed as it did in one workplace that we have intensively studied:
"I know that I am supposed to listen to the computer. But will it listen to me?"
Implementation and Membership

The complex dynamics of implementation of computer-aided decision support
systems, such as MRP, and their implication for workplace organization were
highlighted by our earlier studies of MRP at two workplaces that followed very
different paths in the way that the MRP system was introduced and implemented.
We began to study the way that the MRP system interfaced with production
processes at our primary field site, which we identify as Kemps, which , at the
time of our studies, was a small company providing electrical parts to the defense
industry. At about this time, we picked up a secondary field site, Intek, a hardware
manufacturer of approximately equal size. Both companies had implemented the
MRP system about the same time. By the time we got to study these systems
and their interface with the situated cognitions of various workers, the systems
had been in place for approximately five years.
Interestingly , by all accounts, the system was failing at our primary field site
and was thriving at our secondary site. Intek had managed to use the MRP system
to effect materials savings of several million dollars a year. Kemps had not. Intek
had extremely low employee turnover rates and widespread indices of job
satisfaction, whereas Kemps was experiencing difficulty with employees and high
turnover. But, in demographics and level of skill, the employee population was
quite similar .
255
Ethnographic observations at the two sites supported the hard business data of
differential success. At Intek, we found that a "culture of secondary artifacts" had
been created by employees to elaborate the read-outs of the computer screens.
Thus , computer terminals were surrounded by print-outs of Lotus spread sheets
and the like. Nothing like that was happening at Kemps. And Intek employees
were even able to use the output of the system to make savvy predictions about
production processes. On one tour of the factory floor at Intek, an employee
glanced at a computer screen and said, ''That machine is about to go down,"
pointing to a machine on the floor. When we asked, "How do you know?" , she
explained that the output of the machine was progressively deviating ("look there,
see those figures") from its expected output, and therefore it was failing. Knowing
this, the machine was pulled from production for fixing, and a secondary machine
was put on line. A costly breakdown in production was avoided. Although Kemps
had similar screens (but no secondary artifacts), the information on those screens
was not used the same way. Indeed, because the MRP system had not generated
materials savings at Kemps, it was not clear that the information on their screens
was being used at all.
Conceptual Integration and Code Switching

Ethnographic observations were supported by the results gathered from using
quasi-experimental probe batteries to test the limits of the kind of production
and/or MRP knowledge that workers in the two plants had attained. Our measures
indicated an important aspect of the differences between the plants and the
knowledge of MRP and production held by workers in the plants. Although
workers at Kemps could display quite high levels of sophisticated knowledge of
the MRP system and its concepts, that knowledge seemed to be specific to the
system as it existed in the computer or the front office. They were unable to link
the system knowledge with production processes . They knew the system and how
it did what it did, but there was little linkage with production. At Intek, although
formal knowledge of MRP was comparable and, in some cases, even a little
below MRP knowledge at Kemps, workers displayed, in a number of ways, that
they had integrated their knowledge of production and MRP . What became clear
was that in one setting knowledge remained theoretical, whereas in the other
setting, it had become integrated and practical. In short, it seemed that at one
plant (Intek) MRP had become a tool for regulating production process, but at the
other (Kemps), it was just an additional source of information with little practical
consequence.
One of our team, Elena Zazanis, had particular interest in language issues . She
examined the way in which workers at the two plants responded to our probe
batteries. Analogizing the knowledge system of production and the knowledge
system of MRP to "languages," Elena had the bright idea of analogizing the
integration of these knowledge systems to the linguistic phenomena of code
256
Joseph Glick
switching in bilinguals. She analyzed (Zazanis & Glick, 1993) responses to probe
batteries in terms of whether a stretch of talk reflected the structure and semantics
of production language or of MRP language . Finding that these languages can be
analytically distinguished, she then examined their relation to one another in the
workers' talk. In particular, she focused on whether a worker's language
productions kept the two knowledge systems separate or blended them in
compound utterances.
There is some suggestion in the literature that the phenomenon of intrapropositional code switching can be used as an index of the degree of balance
between language systems in bilinguals. Intrapropositional switching, then,
might be an index of a greater degree of balance than switching that occurs across
propositional boundaries. When examined in this way, workers at Intek showed
greater evidence of intrapropositional code switching, as is shown in Tables 1 and
2.
Table 10.1. Intrapropositional and Interpropositional Code Switching Analysis
# OF PROPOSmONS
S1REAMS
1N1RA-
PROPOSmONAL
SWITCHES
Intek
Kemps
Production
MRP
125
156
42
58
Production
30
27
MRP
9
18
9
2
The code-switching measure related strongly to other measures of the acquisition

of an integrated knowledge of the MRP and the production system . For example,
one of our quasi-experimental probes for MRP and Production knowledge
involved a card sorting task. Subjects were presented with cards representing the
kinds of tags that one might find on objects in any manufacturing environment.
Our cards over-specified an object by providing alternative information systems;
some of the information provided corresponded to MRP nomenclature and logic
(e.g., level 0; levell); and some of the information on the same card corresponded
to a production logic and nomenclature (e.g., "brown tack goes on tackboard"). In
fact, the cards employed four overlapping but not coordinate alternatives. Subjects
were asked to sort the cards in any way that made sense to them. After the first
sort, people were given the opportunity to re-sort. Because any given sort was
going to be either tracking the MRP system information or production
knowledge, or something that was neither, our measure of conceptual integration
of the systems was related to whether subjects sorted flexibly (i.e., sorting one
Discourseand Development
257
way on the first sort and another on a second sort) or not (i.e., sorting according
to one system on both sorts). The results of this analysis are indicated in Table 3.
Table 10.2. Mean Proportions of MRPlProduction Streams and Switches
at Kemps and Intek
Intek
Kemps
(N=12)
(N=14)
41.1
9 .1
39 .5
10.8
58.9
9.1
61.8
12 .2
66.1
18.0
57 .5
Proportion of
MRP Streams
Mean
SD
Proportion of
Production Streams
Mean
SD
Proportion of
Inter-Stream Switches
Mean
SD
21.9
Proportion of
Intrapropositional
Mean
SD
5 .7*
4.4
* produced by 11/12 subjects ** produced by 7/14 subjects
0 .9**
1.0
258
Joseph Glick
Table 10.3. Card Sort Flexibility by Plant and Code Switching
SORTTYPE
Flexible
Inflexible
BYPLANT
Intek
Kemps
9 (86%)
2(14%)
5 (36%)
9 (64%)
BY SWITCHING
Flexible
Inflexible
Intrapropositional
No intrapropositional
13 (76%)
1 (12%)
4 (24%)
7 (88%)
As is plainly seen, there is a difference between the two plants in terms of

flexibility of card sort behavior, indicating, we believe, different levels of
knowledge integration at the two sites. This finding is buttressed by another
finding, also shown in the table, of a clear relation between intrapropositional
code switching and flexible sorting.
One of the questions that you might be asking now is the question that we
asked ourselves: What could possibly account for the different levels of knowledge
integration found in the two plants . They are different, but why?
Procedural Versus Constructive Organizations
One of our research team, Jessie Kindred (Scribner, DiBello, Kindred, & Zazanis
1991) had been working most heavily on an ethnographic analysis of the work
sites. She noticed some differences in the way that jobs were performed at the two
plants. One way of specifying what a job entails is to look carefully at the kinds
of activities that are dominant in the performance of that job . We call this a
leading activity analysis. The leading activity of a job is not fully specified by a
job title. You have to look carefully at what people do and then try to make sense
of it.
Looking at the issue that way, we could distinguish between two types of
organizations for any job. One we call procedural; the other we call constructive.
A procedural activity is one in which there is a clear specification of both the goal
of an activity and the means by which the goal is to be accomplished. A
constructive activity may have the same goal as a procedural activity, but there is
a less clear specification of the means. For example, I might specify how to get
from point x to point y by providing a path statement: "First, you turn left; then
after 10 feet, you turn right , and after 15 feet, you turn left again, etc." Or, I
259
might provide you with a map and let you figure out how to get from here to
there. Jobs in manufacturing environments can be path-like or map-like.
Kindred's analysis looked at jobs in this way, carefully specifying activities and
carefully specifying the degree to which that activity was organized by a clear and
constraining specification of the means to accomplish it.
Several illuminating findings emerged from this analysis. Most obvious, it
became clear that job title did not completely specify the degree of constructive
latitude accorded to a particular worker. Job titles may be seen in this sense as a
statement of the goals implicit in a job. The procedural-constructive distinction
assumes roughly common goals but identifies different relationships between
goals and means . That having been said, the two plants differed overall in the
degree to which their workers were given such constructive latitude in achieving
job goals . Kemps used more specified procedures than did Intek. In fact, all of
those artifacts that festooned the area around MRP screens at Intek had been
constructed by the workers, sometimes for amusement, sometimes as a secondary
tool, and sometimes as a display of their own constructive latitude. In many
plants, such employee-created artifacts are considered undesirable, allowing for too
much individual variation in the workplace. Here we found that the constructive
engagements that such artifacts betokened were key elements of a success story.
Lia DiBello and I (DiBello & Glick, 1992) examined some of these data from
the point of view of plotting the relationship between the degree of constructive
activities on the job in each of the plants to one of our measures of conceptual
integration, namely, the card sorting measure. Without going into excessive
detail, a few things are clear. First, we found considerable variation within job
title within both plants. This suggests that there is some variation that relates to
individuals within organizations . Second, the overall level of constructive
opportunity is higher at Intek (mean of 70% constructive activities) than at
Kemps (59.8%). Third, the variability in constructive opportunity is higher at
Kemps. There is an interaction between variability attributable to individuals and
overall organizational ethos . Fourth, and finally, constructive opportunity and
conceptual integration covary. When one is high, the other is likely to be high,
too. This shows up in a Pearson r of .67 between amount of constructive
opportunity and our measure of cognitive integration. We plotted a similar
correlation for educational level and conceptual integration and found no
correlation at all.
The MRP system had come to these two plants in quite different ways. Intek is
a family-owned business, and the owner has taken a more or less hands-off
attitude, allowing his children and managers the freedom to invent new procedures
to make the business run more smoothly. Kemps is more traditionally organized.
This history seems to have some current consequences. At Intek, the system is
seen as a part of the factory as a whole, and all people are given some level of
training in its use (even the receptionist) . At Kemps, the system is part of a
management control system, and training in its use is reserved for the managerial
260
Joseph Glick
ranks . The system shows up in employees ' lives as a set of orders from the front
office rather than as a tool for all to use.
We might wonder whether this history has been responsible for the different
patterns of implementation success that we have observed, both in business terms
and in terms of cognitive and linguistic measures. This question brings us up
against one of the primary limitations of the project that has been described so
far. We started studying the integration of management control systems with
production after the systems had been in place for a number of years. Our late
entry did not allow us to look at the processes by means of which these patterns
had become established.
We were left with serious questions about how it came to be this way, which
we cannot answer with the current data set. We are now working in an area where
some of these questions can be answered.
From Requisite Skills to Embedded Cognitive Practices

Our current research site is the Transit Authority (TA) of a large east coast city.
This TA is now realizing that some of its operations are in direct competition
with the private sector. But, because it is civil service and there is a strong union,
there is little chance of competing on price, because labor costs are not as flexible
as they might be in the private sector. Upper-level management has seen the
chance of saving a great deal by controlling materials costs and has convinced
some union leaders that this will provide joint benefit. To do this, the TA is
implementing an MRP system in its remanufacturing plants, where subway car
components such as compressors and electric switches are repaired. Our research
team has been there from the beginning, documenting the manner in which the
system is put into place. We are working, as well, in another division of the TA
where a somewhat different system, but with a logical structure similar to that of
MRP, is being implemented in order to more coherently organize repairs on
vehicles in the transit fleet.
Being a part of the scene while the workplace organization is dynamic and in
the process of rearranging itself to cope with the problem of structure formation
in relation to a new technology allows us to examine some of the issues that we
have already discussed. We are pursuing a hypothesis that flows from the data
analyses we have already presented. In any workplace in flux and rearrangement, a
central issue involves who will be rearranged and who will do the rearranging . We
have shown, in our previous work at Kemps and Intek, that this issue relates to a
most important cognitive issue: whether the engagement with the system will be
constructive or procedural. Who will be given the opportunity to constructively
engage the system and who will be merely expected to follow its orders.
261
Knowledge Dynamics in an Environment of Change

We suspect that part of the answer to the question of who gets opportunity and
who does not will relate to an emerging modem conception of the way in which
realities are constructed, a conception that is clearly topical in this volume. In
order to see the issues more clearly, let us consider the different emphases that
underlie the cognitive development model of knowledge and the situated cognition
model of knowledge from a slightly different perspective .
COGNITIVE DEVELOPMENT AND SITUATED COGNITION IN AN

ENVIRONMENT OF CHANGE
COGNITIVE DEVELOPMENT
SITUATED COGNITION
School/family as
- Prerequisite skills
Job as site of preparation

- Postrequisite skills, expertise.
agent of preparation
Guidance toward
progress
- Teaching, scaffolding
- Piggybacked demands
- Development as constraint
Guidance toward efficiency

of function
- Apprenticeship
- On-job training
- Opportunity as constraint
Focus on
Focus on legitimacy
as one who knows
- Displayable competen ce
knowledge
- Testable competence
- Younger to older
Notions of efficiency
alongside of progress
- Novice to expert
Focus on individuals
- Knowledge is at issue
Focus on organizations
- Legitimate participation at issue
Notions of progress
As can be seen from the contrast just shown, the cognitive development model
focuses on priorandprerequisite krwwledge states, whereas the situated cognition

model focuses more on opportunity to learn within a current situation (Lave &
Wenger, 1991). The cognitive development model leads inevitably to discussing
such issues as the relationship that should be constructed between educational
institutions and the needs of the workplace, whereas the situated cognition model
looks to the workplace and the dynamics of opportunity and participation within
it as the focus point.
262
Joseph Glick
Our current research activities are exploring issues within the situated cognition
model. For our group, the issue has become one of identifying the means by
which people accomplish or fail to accomplish legitimacy of participation within
a culture of innovation . Clearly, some of the variance in this process will be
related to management philosophy and institutional ethos, but we also expect that
some of the variance will be explained by a closer understanding of how people
act within their work environments. Although there may be institutional agendas,
people nonetheless seem to be involved forever in carving out a niche within the
constraints of the agenda. Technological innovation can be a powerful force in
providing more latitude for niche finding, even though the underlying motivation
for the technological innovation (as is the case with MRP) is to institute more
top-down control.
We suspect that our finding that knowledge states show up in ways of
speaking, not just in terms of the content of propositions but also in the more
fine-grained structure of discourse, may be an important analytic tool for
investigating processes of niche construction. Discourse may not only reflect the
results of constructive engagement, but it may also be the means by which
constructive opportunity is earned. We are suggesting and already have some
evidence to support the notion that some people begin speaking knowledgeably
long before they are knowledgeable. It could be that display of linguistic
competence earns workers constructive latitude in their jobs .
If these suspicions bear out in findings, we may be in for some serious
rethinking of what the prerequisite skills are for functioning in a modern
manufacturing environment, where the kinds of changes that we have begun to
see with MRP will more likely be the rule than the exception. It could be that
conceptions of competence or required competence for various job classifications
will need to be shifted from the cognitive development inspired intrapsychic
notions, such as abilities or skills, to more situated cognition inspired
interpsychic notions, such as discourse. When it comes to cultural practice, what
improves the chances of a person's being a good practitioner more than being able
to enter into the interpsychic community of practice as a full member.
What I am trying to suggest here, perhaps too elliptically, is that the path that
I described earlier, from cross-cultural studies of general cognitive operations to
cultural practice notions of embedded and situated cognition, will be taken again
as we shift our focus from the skills embedded in work to issues of legitimated
participation in the workplace. It could be that if the opportunity is provided, the
rest will take care of itself. The problem then is to understand both how and why
opportunity is provided and how it can be earned.
I do not think that the issue will turn out to be simply an issue of expertise.
To account for differences in participation in terms of expertise is circular,
because what is often at issue is the opportunity to become expert. Rather I
suspect that there are some highly specific factors blended with some very general
factors involved. To be sure, knowledge and training in the intraindividual sense
263
are involved, we must but equally consider the general issues of interpsychic
selves and organizational philosophies.
This formulation of the issue would clearly recognize that modem workplaces
can and do support a number of different coexisting levels of competence. Under
stable conditions, these differences do not make much of a difference. However,
the situation changes once the workplace undergoes change. It is perhaps the
generalized competencies of impression management (such as language use in
discourse) that may be selecting for change survivors. The issue of earning
opportunity is a serious issue and one that has major relations to the current
manufacturing environment where re-skilling (as jobs are lost) and de-skilling
(where jobs are proceduralized) seem to go apace.
Although people leave schools, and the design of the institution of school is
predicated on that leaving, the design of the institution of the workplace has
become much less clear in recent times. On the one hand, the workplace seems to
be institutionally organized in terms of people staying within it. Thus,
knowledge in the workplace can be distributed between individual cognitions and
artifact design. However, as employment prospects in a global economy become
less certain, the stability assumption no longer holds for the workplace. People
may have to leave the workplace as well. And if they stay, they may have to stay
within environments that place more of the cognitive structure in the design of
job procedures and intelligent artifacts, raising the issue for the worker of whether
it is even worth staying in an environment where the worker counts less and less.
Although there is danger here, there is also opportunity. Many of the
technological innovations only seem to place the cognition in the artifact. Deeper
analysis generally shows that what is really involved is the construction of an
opportunity for there to be new conditions of engagement between workers and
workplaces. To the extent that the new conditions of engagement afford
opportunities for constructive engagements within the structure of the job, it can
be argued that the workplace is beginning to take on some of the characteristics of
the school. There are pockets of opportunity where quite general skills are
developed in the context ofdoing highly specific jobs.

This suggests that there is considerable promise in an approach to issues of job
training and cognitive development in practice that begins to look at the issue of
workplace design in terms that more directly relate to the issue of the
development of the worker. If workplaces are not places anymore where it is safe
to assume that the worker will remain, then focus could be shifted from thinking
in terms of prerequisite knowledge and skills to thinking of the workplace as a
learning environment. Our results already suggest that the developmental
workplace (e.g., Intek) has achieved success in both business and developmental
terms. Rather than being unrelated concepts, it is likely that human development
and business efficiency can be related rather than opposed concepts.
264
Joseph Glick
References
Beach, 1(., Zazanis, E., & Glick, J. (in preparation). Ratio in the scale or in the hand?
Does black box technology disconnect the mind?
Cole, M., Gay, J., Glick, J., & Sharp, D. (1971). The cultural context of learning and
thinking. New York: Basic Books.
DiBello, L., & Glick, J. (1992, October). The relative roles of on-the-job training and
classroom education for workers learning MRP-Il effectively. Presented at APICS
Society meetings, special academic session. Montreal.
Lave, J. (1988). Cognition in practice. New York: Cambridge University Press.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation.
New York: Cambridge University Press
Martin, L. M. W., & Beach, I(. (1992). Technical and symbolic knowledge in CNC
machining: A study of technical workers of different backgrounds. (Technical
report) University of California at Berkeley, National Center for Research in
Vocational Education.
McDermott, R. P. (1993) . The acquisition of a child by a learning disability. In S.
Chaiklin & J. Lave (Eds.), Understanding practice: Perspectives on activity and
context. New York: Cambridge University Press.
Sabel, C. (1991). Moebius-strip organizations and open labor markets: Some
consequences of the reintegration of conception and execution in a volatile
economy. In P. Bourdieu & J. S. Coleman (Eds.), Social theory for a changing
society (pp. 23-61). Boulder, CO: Westview Press & Russell Sage Foundation.
Scribner, S. (1984). Studying working intelligence. In B. Rogoff & J. Lave (Eds.),
Everyday cognition : Its development in social context. Cambridge, MA: Harvard
University Press.
Scribner, S. (1986). Thinking in action: Some characteristics of practical thought. In
R. J. Sternberg & R. I(. Wagner (Eds.), Practical Intelligence (pp. 13-30).
Scribner, S. (1990, May). The character of knowledge systems in the workplace from
the perspective of activity theory. Paper presented at The Second International
Congress for Research on Activity Theory, Lahti , Finland .
Scribner, S., & Cole, M. (1981) . The psychology of literacy. Cambridge, MA: Harvard
University Press .
Scribner, S., DiBello, L., Kindred, J., & Zazanis, E. (1991). Coordinating two
knowledge systems: A case study. Unpublished manuscript, City University of New
York, Graduate School, Laboratory for Cognitive Studies of Activity.
Wallace, T. E. (1990). MRP II:_Making it happen: The implementers ' guide to success
with manufacturing resource planning (2nd ed.). Essex Junction, VT: Oliver Wight
Limited Publications.
Zazanis E., & Glick, J. (1993). Discursive measures of cognitive structure.
Unpublished manuscript, City University of New York, Graduate School , Laboratory
for Cognitive Studies of Activity.
Chapter 11
Interactional Perspectives on the Use

of the Computer and on the Technological
Development of a New Tool:
The Case of Word Processing
MicheleGrossen' and Luc-Olivier Pochon?
lInstitut de Psychologie , Universite de Neuchatel, Switzerland
2Institut Romand de Documentation et de Recherches Pedagogiques (IRDP),
Universite de Neuchatel, Switzerland
Abstract
This chapter focuses on the user's actual behavior when he or she uses a computer .
With reference to concepts developed in the field of ethnotechnology and the
social psychology of cognitive development, a series of observations on humanmachine interaction are reported. The analysis of these observ ations shows that
human-machine interactions result in a specific inter actional space that can be
reduced neither to the user's abilities and technological knowledge nor to the
technological qualities of the machine. This result is further illustrated by a short
case study concerning word processing . First, some historical elements
pertaining to the development of this new tool are reported ; then, some examples
show the specific effect the use of word processors has in social and institutional
contexts . The conclusion is that using a computer consists of an indirect dialogue
between the users and designers . As a dialogue, the use of a computer includes
theoretical and methodological problems that are typical of human-machine
communication in general .
266
Michele Grossen and Luc-Olivier Pochon
Introduction
With the emergence of new information technologies (from the television to the
computer), the number of studies on human-machine interaction has increased.
These studies have turned more and more to multidisciplinary approaches
pertaining to social sciences, computing, and technology. Such approaches have
already been adopted in some fields of psychology (e.g., industrial psychology)
and have been promoted in the teaching of technical disciplines (Podak, 1993).
Next, to give a broad outline of these studies, we mention some technological
fields that have extended their research to include a more multidisciplinary
approach:
- Computer-assisted design and production. In this field, users are more and
more dependent on the models implemented by software. This dependence raises
an important issue regarding the transfer of knowledge and competence from one
field to another: How can users make use of their previous knowledge and
competence in this new context? The answer to this question is very complex,
because, as several studies have shown (see, e.g., Waern, 1993), an analysis of
the characteristics of the tasks is not sufficent to predict transfer. To account for
this complexity and to improve conceptualization, it has proved necessary to draw
on other disciplines that have already tackled the same problems. But, of course,
the issue is then to examine to what extent concepts and results obtained in these
disciplines are applicable to the field of human-machine interaction.
- Computer networks in firms and, in particular, computer integrated

manufacturing (elM) technology. In this field, the issue concerns the conditions
that enable collective work mediated by computer tools to be effective. In fact,
communication between network managers and users is full of misunderstandings,
not only about how the system works but also about its potential. In their
analysis of collective work (collaborative writing, for example), several
researchers (cf. Kraut, Galegher, Fish, & Chalfonte, 1992) emphasize that it is
necessary to consider not only the interactive characteristics of the media but also
their symbolic values and the user's interpretation of how the computer works. It
thus becomes imperative to turn to disciplines such as psychology or sociology.
- Research into computer-assisted learning (in the broad sense of the term). For
the first time, the machine is considered to be a cognitive partner for the user.
However, experiments carried out on computer-assisted learning have highlighted
the difficulties encountered when trying to compare general discourses with their
actual use. Researchers in the field of technology who make comparisons of this
type have to embrace broader issues, such as the user's learning processes, the
role of emotional and motivational factors, the influence of the task on the user's
behavior, and the importance of the interactional and institutional contexts of the
activity . Such issues fall within the boundaries of cognitive and educational
psychology and sociology, disciplines that are not usually attended to within the
field of technology.
Interactional Perspectives on the Use of the Computer
267
Thus, in different fields, the use of a new technological tool has revealed that
social, ideological, symbolic, and affective dimensions, which go far beyond
technology itself, must be taken into consideration.
Within this debate, our own attention has been focused on the use of computer
science in school. Following the development of this educational innovation, two
striking facts have come to our attention: The first was the confrontation between
positive and negative debates, based mainly on ideological arguments, in which
little attention was paid to the actual use of the computer in context. The
assumption was that the users (at least the "skilled" ones!) would inevitably use
the computer in the "correct" way: that is, in accordance with the designers'
expectations . The second striking fact was that most of these ideological
arguments lacked reference to previous studies and knowledge acquired in the field
of educational psychology, giving the impression that the introduction of
computers into schools had rendered irrelevant previous knowledge about learning
and teaching.
In an attempt to pool our resources and benefit from our respective backgrounds
(in psychology for the first author and in mathematics and computer science for
the second), our general aim has been (l) to clarify some of the assumptions
underlying the debate surrounding the introduction of computers into everyday life
and into the school environment; (2) to collect a series of examples from existing
scientific literature and from our own observations of users' actual behaviors; and
(3) to examine to what extent results obtained in the field of developmental
cognitive psychology (in the post-Piagetian and post-Vygotskian streams in
particular) might help construct a theoretical framework that would enable us to
account for different studies and observations. We now present this framework and
show how it can explain different observations in the field of human-machine
interaction.
This chapter is divided into four sections. In the first, some basic theoretical
concepts pertaining to ethnotechnology and the psychology of cognitive
development are presented briefly. With reference to concepts such as semiotic
tools, sociocognitive conflict, and intersubjectivity, four issues regarding the
application of these concepts in the study of human-machine interaction are
raised. In the second section, these issues are illustrated with some of our own
observations and some examples discussed in the scientific literature. Our first
conclusion is that human-machine interactions result in a specific interactional
space that can be reduced neither to the user's abilities and technological
knowledge nor to the technological qualities of the machine. In the third section,
we present a short case study of a particular use of the computer: word-processing.
Some historical elements pertaining to the development of this new tool are
reported. Then, some examples are given to show the specific effect the use of
word processors has in social and institutional contexts. In conclusion, we argue
that using a computer is a socially shared activity mediated by a technological
tool and consisting of an indirect dialogue between users and designers. As a
268
Michele Grossenand Luc-Olivier Pochon
dialogue, the use of a computerincludes theoretical and methodological problems

that are typical of human communication in general.
Human-Machine Interaction: A Research Topic at the

Crossroads of Ethnotechnology and the Social
Psychology of Cognitive Development
Our theoretical framework has two main sources: ethnotechnology and the
psychology of cognitivedevelopment, as studied along Vygotskian and Piagetian
lines. Ethnotechnology is the study of interactions between technologies and
societies (Gaudin, 1978, 1988). One of the main assumptions is that the use of
the technology as planned by the designer does not always coincide with the
user's actual use of that technology in concrete situations. Seen from this
perspective, two central questions appear: For what purpose, from the designers'
point of view, has the technical tool been constructed? What is the user's actual
use of the tool in concrete situations?
Within this field of research, Perriault (1989) has studied the way in which
individuals use communication media (e.g., the telephone, Minitel, the
computer), making a distinction between the instrumental function of a tool and
the user's actual behaviors. The latter is defined as the result of a negotiation
between the user, who plans a certain use of the tool, and the tool itself, with its
original function. Showing that certain users' actual behaviors imply a dramatic
modification of the designers' initial project, Perriault discusses the example of
the French Minitel (a public electronicnetwork). The originalfunction of Minitel
has been strongly modified by the users themselveswith the introduction of such
things as chat lines and classified ads. According to Perriault, these various users'
actual behaviors are the result of a particular rationality emerging from the
interaction between the user and the machine. Perriault calls this rationality the
logic of use. The originality of this interactional perspective on human-machine
interactions is that it considers the user an active partner in the human-machine
relationship and, consequently, sees the machine as potentially subordinatedto the
action and power of the user.
Within the field of the psychology of cognitive development, the sociocultural
approach derived from Vygotsky's work (Rogoff, 1990; Vygotsky, 1978, 1962;
Wertsch, 1991) shows that cognitive development is a culturally rooted activity
relying on cultural, social, and institutional resources. From this point of view,
cognitive development is considered to be the ability to use some semiotic tools
that have been constructed in the course of the historical and cultural development
of a social community and that are transmitted and created through social
InteractionalPerspectives on the Use of the Computer
269
interactions. Thus, a central subject matter for study is the analysis of the social
processes of creation and transmission of these semiotic tools.
The study of the role of social interaction in cognitive development and
learning is also a basic theme ofpost-Piagetian research, which we refer to here as
the socialpsychology of cognitive development. In a first generation of research
(for a more detailed account, see Perret-Clermont, Perret, & Bell, 1991), studies
relying on a three-step design (pretest, test, posttest) focused on the social
conditions enabling a peer interaction session to be efficient for subsequent
individual learning (see, e.g., Doise & Mugny, 1981/1984; Doise, Mugny, &
Perret-Clermont, 1975; Perret-Clermont, 1980). One of the main results of these
studies is that situations in which the interacting partners give diverging
solutions to a problem are likely, under certain conditions, to provoke a
sociocognitive conflict, which has the effect of leading each individual to
restructure his or her reasoning. In a second generation of research (e.g., see
Perret-Clermont & Nicolet, 1988), the emphasis was placed on the meaning of
the learning or test situationsin which children are requested to solve a problem.
Similarly, the focus shifted from observing the individual effect of a peer
interaction session to observing interactional dynamics of the partners. Under the
influence of scholars such as Ragnar Rommetveit (1979, 1992), the main purpose
of these studieshas been to examinehow the childreninvolved in a learning or in
a test situation (Grossen, 1988; Grossen & Perret-Clermont, 1994) give meaning
to the situation and understand the implicit assumptions on which the situation is
based. Attention has thus been paid to certain communication processes such as
iniersubjectivity, showing that the development of a new competence or the
learning of a given body of knowledge is intrinsically linked to the sharing of
certain states of intersubjectivity. As a consequence, the alternation of states of
sociocognitive conflict and states of intersubjectivity appears to be a condition
that enables individuals to coordinate their points of view and come to an
agreementwhile at the same time being able to benefit from their differences.
Other studies have explored the role of implicit rules, values, norms, and
expectationsin the learning process and in the construction of an answer in a test
situation, that is, analysis of what can be called an experimental or didactic
contract (Schubauer-Leoni & Grossen, 1993; Schubauer-Leoni, Perret-Clermont,
& Grossen, 1992). These studies constitute only a small proportion of a large
number of studies relying on the assumption that learning situations are basically
communication situations that are culturally rooted and whose meanings have to
be constructed intersubjectively during the interaction (e.g., see Edwards &
Mercer, 1987; Elbers, Maier, Hoekstra, & Hoogsteder, 1992; Light &
Butterworth, 1992; Nunes, Light, & Mason, 1993; Resnick, Levine, & Teasley,
1991; S1Hjo, 1992; and the work of those contributing to the present volume).
To sum up the main features of these two psychological approaches in
conjunction with ethnotechnology is to consider the individual's interpretative
activity as a fundamental part of their actions and as a determinant of the kind of
cognitive processes they use in concretesituations.
270
MicheleGrossenand Luc-OlivierPochon
Having referred to the concepts of semiotic tool, sociocognitive conflict, and

intersubjectivity, along with the concept of "logic of use," we now raise four
issues:
1. Can the computer be considered as a new semiotic tool that enables
individuals to develop a new competence?
2. Does human-machine interaction elicit sociocognitive conflict? Is it
possible to view human-machine interaction as an intersubjective construction?
3. Does the use of the computer give rise to new types of interaction between
individuals? If so, what is the specific nature of these interactions?
4. What are the users' actual use behaviors? What are the modifications? How
do they differfrom the designers' planned use?
We now examine these questions more closely and illustrate them with some
general observations.
Observations of Human-Machine Interaction

From the Computer as a Tool to the Computer as a Semiotic
Instrument
The sociocultural approach referred to earlier leads us to study more closely the
creation of new semiotic instruments and their possible impact on human thought
processes at any given time in cultural history. This issue has often been raised
by authors from diversedisciplines and backgrounds (see, among others, Belyaeva
& Cole, 1989; Papert, 1980, 1993; Pea, 1985). Moreover, each medium
promotes certain thought processes: Books, radio, and television modify
intellectual processes, each in its own particular way (e.g., memory, mental
representations of situations, how connections are made; Greenfield, 1984). Two
examples illustrate this point: The first example shows how the computer
modifies the thought processes and research practices of a scientific community.
In 1976, a mathematical theoremcalled the four colour theorem (Appel & Haken,
1977) was demonstrated through the use of a computer. Many mathematicians
were disconcerted by the result: To verify the accuracy of the demonstration, both
the method and the demonstration program had to be validated. Was this still
within the realms of mathematics? Hence, with the birth of experimental
mathematics (Cohen & Norden, 1989), the method goes beyond a mere
broadening of single cases. In mathematics, the introduction of the computer has
provided, and still provides, the means to use specific cognitive abilities:
recursive, cybernetic, systemicreason. Recursive reasoning is a good example of
this evolution (see Vitale, 1989; Pochon, 1991). In fact, having first the structure
of a mathematical proof, recursiveness has become a relatively widespread
techniquefor the construction of computerprograms (control structure).
271
The second example shows that the computer could be seen as a sort of
intellectual prosthesis, enabling problems to be solved without the necessary

cognitive capabilities being available (Pea, 1985). However, this claim seems too
general and needs to be examined in more detail in order to distinguish to what
extent the computer differs from, for instance, the abacus, an instrument that also
enables the mechanization of certain thought processes. Research carried out with
handicapped children using pocket calculators reveals the interdependence of
different kinds of tools facilitating problem solving: for example, writing figures,
symbols, formulae, and mental arithmetic. As one child put it, "It's a miracle, it
thinks for itself. . . . it works like a turbo engine and helps you to think."
(Guggisberg & Balmelli, 1992).
These examples show that the computer, whether through calculation,
programming, memory, or time factors, mediates new activities into diverse
areas: mathematical demonstrations, games, and simulations . The computer has,
for example, modified scientific methodology due to its ability to model very
abstract concepts and very complex phenomena. In keeping with Vygotsky's
psychological model, one can hypothesize that, through interaction with the
computer, specific modes of thinking are developed. The problem, however, is to
describe and characterize these modes of thinking. They could, for example, be
considered a reflection of the functioning of the computer programs themselves,
or of the actions to be carried out with the various interfaces, or a progression of
sequences on the screen. Thus, an analysis of the syntax and semantics of these
programs would provide a starting point for a description of these thought
processes. A description of this type is nonetheless far from self-evident, as
Waern (1993) shows by pointing to the difficulty of describing the nature of the
users' mental models, except in the case of very obvious errors.
In our opinion, an interactional stance on human-machine interaction makes it
difficult, if not impossible, to conceptualize to what extent the machine reflects
human thinking or vice versa. For this reason; we decided to collect examples that
illustrate how the computer is used in real situations . In other words, our
assumption is that the use of the computer is a situated activity and that studying
the situation as a whole may improve our understanding of the specific nature of
human-machine interaction.
Intersubjectivity and Sociocognitive Conflict in Human-Machine

Interaction
Does human-machine interaction provoke sociocognitive conflicts that enable
individuals to reorganize their thinking? Is it possible to assume that humanmachine interaction might involve a process of intersubjectivity?
An example of computer-assisted learning may help to answer these questions,
because it involves social interaction among several partners:
272
a. The designers. who design the program according to the technical potential of
the machine and their own representations of the body of knowledge involved;
b. The teachers, who use the program in the classroom and who, depending on
their own assumptions about the body of knowledge involved and the use of
computer in the school context, give their own interpretation of, for example, the
program, its purposes, its "correct" use;
c. The pupils, who use the program according to their own interpretation of
how to use the computer, the discipline being taught, and the definition of the
situation (e.g., should they use the program to play or to learn).
Each of these partners has specific representations of the machine, of its use in
the school context, and of the body of knowledge contained within the program.
Thus, it is highly possible that the representations of one partner do not coincide
with those of another, a situation that could give rise to misunderstandings. Thus,
when someone uses a computer, he or she is confronted not just with a piece of
technology but also with the assumptions of other individuals.
This example also shows that using a computer is far more than a simple
interaction between two partners: a machine and a user. In fact, the computer can
be considered a screen hiding the designers, the programmers, and other
individuals (the teacher, for example) who put the computer at the disposal of the
end users.
Having its own logic and sometimes taking on the role of "adversary," the
computer is not a malleable object that will bend to the will of its user. It resists
and imposes its own way of doing things on the user. Thus , it can give rise to
certain sociocognitive conflicts that oblige the users to reflect on their own
cognitive activity and to solve some of the problems they encounter (see
Mevarech & Light, 1992; Light & Littleton, this volume) .
Furthermore, programmers also have certain assumptions and expectations with
regard to possible errors and the user's reaction to them. They have conceived a
model that is supposed to match the user's expectations. Vacherand-Revel and
Bessiere (1990) describe three ways in which programmers could approach the
user : They could defme the user type; they could take into consideration the
diversity of users as a function of their needs; or, through either direct or indirect
interaction, they could consider the reciprocal adaptation of the model as
originally conceived and the users' model in practice. The latter and most recent
approach fits our hypothesis regarding the construction of intersubjectivity,
namely , the creation of a common system of reference. In fact, our hypothesis is
that the computer does not come out of the interaction altogether unscathed but
undergoes changes that turn it into an object different from that originally
conceived of either by its creator or by its users . We expand on this point later.
273
Interactions Mediated by the Computer
Does the use of a computer have a specific impact on the organization of

interindividual interactions?
Some examples suggest that the use of the computer gives rise to specific
indirectrelationshipsbetweenindividuals:
- One example is provided by users playing computer games. For example,
players of Les Voyageurs du Temps [Time Travelers] (Cuisset, 1986) have
constructed a social network that they use to communicate indirectly with each
other. Using a given game thus becomes a sign of belonging to the same group
and may result in actual encounters or communication through more classic media
(such as newspapers). The latter is called meta-communication by van der Veer
and colleagues (Gillet, 1990).
- Another example is provided by the system of interactive messages, the
modern equivalent of Turing's test, that brings about new types of cooperation
between users and new decision-making processes (Sproull & Kiesler, 1991).
Networks connected to hypertext systems also contribute to the creation of
complex interactions between the different users (Alter, 1985). These systems,
called groupware, are sometimes considered an integral part of intelligence
technologies (Levy, 1990).
These examples show that, contrary to current thinking, using a computer does
not lead to the isolation of the individual concerned but gives rise to specific
relationships that are mediated by the computer.
The Evolution of Machines and the Modification of Their Use
When entering the "logic" of the machine, the user tries to find out how it works
and to interpret how it can be used. This can lead him or her either to modify the
original design of the computer by increasing its functions and performance or to
change its use by adding elements that the designers did not originally anticipate.
What is the nature of these modifications? In what ways do they change the
functioning of the computer?
The Modification of the Original Design of the Computer and
Software. The performance of computers is constantly improving.
Programming languages are evolving at increasing speed. At each stage, N, a
language facilitates the creation of tools, which in turn enables language of N+1
orders to be translated into language N. This evolution takes place at every level:
for example, system or partner. With this in mind, we have observed how, since
the appearence of the first Dbase 2, people use the system not only to manage
data but also to manage programs.
274
Modification of Use. Modifications of use have more local consequences

than those that are modifications of the original design and software. One of the
most interesting examples of a modification of use is provided by
microprocessors, which were initially dedicated to driving the peripheral devices
of a computer but were then used as computers in their own right. This
modification of use has given rise to yet unresolved controversies about
microprocessor patents (Who Made the Micro, 1990; Faggin, 1992). It also
highlights the fact that the notion of invention is often relative. We also have
observed how children modified the use of some teaching software (Grossen &
Pochon, 1988). The task proposed by the program was to complete calculations
that appeared on the screen in the form of raindrops before the latter hit the
ground. Some pupils managed to stop the program to divide the calculations
among themselves, thus giving them all the time they needed to finish the
calculations.
The Concept of Interactive Space

Starting from two different theoretical backgrounds, ethnotechnology and the
psychology of cognitive development, our intention has been to draw attention to
a series of issues with the aim of examining the sociocognitive dimensions of
human-machine interaction.
The examples given illustrate that human-machine interactions, which at first
sight appear to be made up of two partners (the computer and the use) , in fact
involve several individuals either directly or indirectly. The designers, the
programmers, and the users having to teach, acquire, or create certain skills all
share the same space. In this space, which is both real and symbolic, they interact
with each other, and each assumes that his or her own system of representations
and technical background is shared by the other partners.
In somewhat polemic terms, we propose the notion of Added Intelligence as a
heuristically interesting way of accounting for the constant interaction between
improvements in computers and new designs, and uses that result: Even if
human-machine interactions are structured within the boundaries of computer
technology, they cannot be completely defined in terms of it. As a result of their
study of pilots and air traffic controllers;' Gras, Moricot, Poirot-Delpech, &
Scardigli (1990) developed the same idea by introducing the notion of man-
1"Henceforth, in many incidents and accidents, it has become difficult to distinguish

between human and machine error. It is, in fact, often a case of man-machine
communication breakdown . This hypothesis leads to two conclusions: The user did not
know how to use the computer; and conversely, the computer did not know how to
decode the implicit requests of the user, did not follow his familiar train of thought, did
not take into account his cultural background. " (Gras, Moricot, Poirot-Delpech, &
Scardigli, 1990, p. 159, translated from French) .
275
machine combined intelligence. Neither man nor machine is passive, and their
respective activities create an original interactive space; just as water is not found
in either hydrogen or oxygen alone, the global result of the man-machine system
is not determined by the sum of the parts.
By the same token, these observations enable us to put forward the hypothesis
that, depending on the context in which it is being used, the computer is not just
a set of determined technological characteristics . On the contrary, its good and bad
features may be deemed to result from the specific nature of the user's behavior.
In the next section, we apply this theoretical framework to a single example of
computer use: word processing. As we demonstrate, word processing offers
numerous examples of the specific impact that human-machine interaction has on
technological development.
The Creation of an Interactive Space: The Use of Word

Processing
Research into the use of word processing has often focused on the following
question : Does the introduction of this new technology enhance writing skills?
This question is linked to Engelbart's statement that "The important thing to
appreciate here is that a direct new innovation in one particular capability [namely
the process of composing text] can have far-reaching effects throughout the rest of
your capability hierarchy" (reported by Rheingold, 1992, p. 83). Despite the
numerous studies on this subject, no consistent results have been obtained, as
shown in Bangert-Drowns' recent review (1993) . According to Bangert-Drown,
this inconsistency may be explained by the fact that the situation in which users
work, their motivation, and their understanding of the task vary and are often
unclear. In our opinion , these conclusions stress the necessity to situate the use
of this tool in a broader context with the hypothesis that it is the context that
provides the users with clues about the meaning of the situation and their own
activity.
Bearing these conclusions in mind, we adopt a holistic approach to humanmachine interaction that, in our opinion, suits the complexity of the subject and
is consistent with the assumption that a computer is a cultural object used in
specific contexts.
In the remainder of this chapter, we first take a historical pespective to illustrate
the effect of human-machine interaction on the technological development of
word processors. We then give some examples showing the effect of the
introduction of word processors on institutional organization and interindividual
relationships. Finally, we report some observations concerning the actual use of
word processors in order to show the creative dimension of this interactive space.
276
A Short History of the Development of Word Processors

The historical development of word processors is not easy to retrace, because,
from the beginning, different partners (mainly computer scientists, word processor
designers, and users) interact, exchange ideas, and alter innovations to suit their
own purposes. It is thus necessary, even if difficult, to disentangle this
interactional network to examine step by step the emergence of these
technological innovations.
The development of word processors began with the editors used by
programmers when time-sharing and on-line work (instead of punch cards) were
introduced. It ends, albeit temporarily, with the now common systems that enable
us to change the setting and the organization of a text. Between these two
landmarks, the idea of using computers to produce text has emerged. With current
technological competence, the use of a computer to write a text is taken for
granted, but the historical development of word processors shows that such a
function was not thought of at all in the early 1960s. At that time, D. Engelbart
was among the few people who envisaged such a machine, and nobody seemed
able to foresee to what extent his predictions would be fulfilled: "This
hypothetical writing machine thus enables you to use a new process for
composing text. . . . If the tangled thoughts contained in the draft become too
complex, you can compile a recorded draft quickly. It would enable you to
accommodate more complex trains of thought which you might build upon in
search of the end result you wish to achieve." (Engelbart, 1963, quoted in
Rheingold, 1992, p. 83).
Examining this complex network, three threads in the history of word
processing can be seen (see Figure 1 below): The first thread takes us back to the
first word processor that was created in 1972. Until that time, nobody had
imagined that a computer could be used to write a text. Even when this idea
finally emerged, word processors were designed not for laymen but for the
designers themselves, who used them to write and edit user guides for their own
programs. Obviously, these word processors were not very flexible, because the
texts had to be filled with symbols defining the page setting . Texts could not be
read on the screen, and laymen did not have the necessary technological
competence to use this tool. Thus, the first word processors were used only for
specific types of texts and designed for technologically competent users.
The second thread pertains to another group of people : the typewriter
manufacturers who tried to replace the electric typewriter. One of the first
machines, Visiotext, was created in 1978 by IBM and was a piece of hardware
loaded with a single word processor. It reproduced the functioning of a typewriter;
for instance, the ends of the lines had to be indicated . Specific keys also had to be
added to the keyboard: end of page, end of text, print text.
The thirdthread concerns computer technologists working on the development
ofapersonal computer (PC). The first PC, Altair, was created in 1975 and was
loaded with a word processor. However, it was never distributed, and, therefore,
277
Apple II (created in 1977) can be regarded as the first commercially distributed

personal computer. No widespread use was made of the PC in business and
industry before the 1980s.
11lread Computer Science
Thread Typewriter
'Thread Microcomputer
Emergence of concepts
ter [
Processcontrol (1951)
Human machine
First interactive
device(screenand
communication (1955)
optical p!:n) (55)
Interactive computer
(1960)
Data & ressource

sharing (1961)
Eeclric t)'pewntet
1962
Word processing
(1963)
ypeWJieeler (J 970)
"Theprinter is a
file"(l972)
"What you see is

what you get" (1978)
Apple D(1978)
Wordstar(1978)
mMpc (1981)
LISA 1983)
FuU screen editor
(1984)
MacIntosh
{aeWr ite (l 984)
Visi04 (l984)
Fig. 11.1. The History of Word Processing : Some Points of Reference
From then on, word processors developed rapidly under pressure from typewriter
manufacturers who played a mediating role between computer technologists and
laymen . Subsequently, and as a result of the interactions between typewriter
278
manufacturers and computer technologists, explicit reference to typewriters was

abandoned, and multipurpose computers loaded with specific programs appeared
on the market.
Thus, a word processor for laymen, WORDSTAR, was designed by computer
technologists. This word processor had different standards from those of a
typewriter and took advantage of certain technological developments in the
computing field. For example, no symbol indicating the ends of lines had to be
given. This innovation itself was borrowed from computing: It had been
introduced in 1972 by the operating system UNIX in order to standardize the
peripherical devices of the computer. Another innovation introduced by
WORDSTAR was that the commands could be performed by combining keys
(e.g., CTRL K.B to mark the beginning of a block of text) instead of using
specific keys.
It is interesting, however, to note that, in the course of this development, many
commands had their origins in the field of technology. For example, the
command OPEN FILE, which is still the most common command creating a new
text, originates from the operation programmers had to carry out to create a space
on the disk.
Subsequently, some of the commands used by this new generation of word
processors (in particular, those of WORDSTAR) became so widespread that
computer technologists began to use them for editing their programs.
Consequently, although at the very beginning of their development, word
processors were not aimed at laymen, they first had to be used by laymen before
being adopted by computer technologists, who in turn used them to facilitate their
own work (i.e., program writing).2
Thus, after a series of interactions among computer scientists, manufacturers,
and designers producing word processors for laymen, word processors in their
present form were used by computer scientists. The creation of graphical
interfaces that followed makes it even more difficult to analyze the exchanges
among computer scientists, designers, and users.
Turning back to the theoretical framework and the issues presented in the
beginning of this chapter, the main features of this brief history may be
summerized as follows:
(1) The present design of word processors seems to be the result of successive
modifications of use: The first modification took place within the computer
technologists' group and was a result of the idea of using the computer to write
users' guides; the second modification took place within the groups of typewriter
manufacturers and computer technologists working on word processors when they
extended the use of word processors to all kinds of writing and made them
accessible to the layman; the thirdmodification was, once again, brought about
2More recently , a way to produce a help system for WINDOWS has been devised that
uses the word processor WORD .
279
by computer technologists when theyreappropriated a tool initially developed for

other uses.
(2) The historical development of word processors is one not only of
technological improvements but also of a series of complex relationships among
competing groups with different needs. To some extent, the interpersonal and
intergroup dynamics could be described in terms of sociocognitive conflicts
comparable with thosedescribed in the field of cognitive development.
(3) Two differentconceptions of the processes involved in the design of a new
tool emerge from this historical analysis: The first conception, illustrated by the
computer technologists, is one of designing a new tool without considering the
user's previous technological competence. A good example of this approach is
providedby the programLISA, which was based mostly on the competence of a
small group of researchers in computer science and was, as a consequence,
unrelated to users' everyday social practices. In the case of LISA, the
consequences were dramatic, because it never gained acceptance by the users and
was thus a commercial failure. The second conception, illustrated by the
typewriter manufacturers, was one of adapting the tool to the user's previous
technological competence, as in the case of Visiotext. Contrary to the first
conception, the second tries to take the users into consideration in order to achieve
certain states of intersubjectivity between the users and the machine. In both
cases, however, the user seems to remain an ideal rather than a real one.
The Impact of Word Processing on Social and Institutional

Organizations
Focusing on the effect of word processing on a specific competence, such as

writing skills, may lead to the neglect of its broaderimpact on the organization of
institutional contexts and on interindividual relationships. In our opinion, this
point is worth considering, because possible effects on specific skills might
actually be mediated by institutional and social factors. In this case, it would be
inaccurateto concludesimply that the tool itself did or did not have an impact on
the individual's skills.
The job of a secretary can be used to illustratethe impact of wordprocessing on
the organization and content of a professional practice. Some years ago,
secretariesspent a large proportion of their time typing texts, whereas their major
task now often involves correcting a text that is already typed on a floppy disk.
An important difference is that they do not work on real texts anymore but on
icons or other symbols representing the text. The level of abstraction of their
activity has thus increased. Moreover, the abilities required of secretaries have also
changed: With the existence of a spell check, the ability to avoid typographical
errors is not particularly important anymore, whereas the capacity to enter
modifications quickly and to improve the quality of the presentation is much
more important. On a personal and social level, word processors also have an
280
impact on the social identities of those using them, because they introduce new
social distinctions between those who use them and those who do not.
At an institutional level, we also observed that some managers see this new
tool as a means of exercising their authority. For example, a director of a school
decided to impose the use of one particular type of word processor on the
institution. Because there was no economic reason for this decision and the
availability of compatible programs made this demand unnecessary, we
hypothesized that this decision had a symbolic function comparable to the choice
of a letter head or logo. The choice of a certain word processor represents an
institutional symbol , a sort of emblem, and becomes part of the institution's
identity.
Turning now to the impact of word processing on interindividual relationships,
five different, but not mutually exclusive, kinds of effects can be identified:
(1) Effects on the individual's self-definition and group membership identity.
Some word processors, such as WORD or WORDPERFECT, have generated
competing fan clubs who support the respective qualities of "their" program. As
in the preceding example, the program gains a symbolic value and becomes a
positive sign of the individual's identity and of his or her group membership.
(2) Effects on the type of interindividual relationships. For example, in the
school context, word processors are often used to publish newspapers written by
pupils and distributed to parents and other schools . Pupils of different schools
exchange their newspapers and keep in touch through this medium.
(3) Effects on the author's relationship withhis or herwritten production. Word
processing enables two or more individuals to easily write a text together : for
example, by attaching a file to an E-Mail message or by using a "travelling
diskette," which goes from one writer to the other. Because it is rather difficult , or
at least tedious, to compare the changes occurring from one version of the text to
the next, the writers' relationship with their own written product changes: The
text no longer seems to be the "work" of a given individual but becomes common
property .
(4) Effects on the written production itself. With the use of word processing ,
the text acquires a flexibility it did not have before. Small corrections, as wen as
deep changes in the textual organization, can be carried out quickly. Texts are
regarded as being more temporary, as drafts liable to change. French writer Michel
Butor provides some testimony on this point: Questioned about his last book
(Butor, 1992) , he states that the use of a word processor has changed his
conception of a text and led him to adopt a hypertext writing influenced by the
technological possibilities of the computer (Audetat, 1993; see also Dubuisson &
Weiss, 1989).
(5) Effects on the mode of transmission of expertise. Sometimes when they
work in a group (e.g., students working in computer rooms), users do not consult
the user's guide (assuming that they have one) but prefer to ask their neighbor for
assistance . Or, if they work alone, they might ask for information in informal
discussions with friends. In fact, users frequently become involved in discussions
281
about the problems that are presented by a certain program and how they solve
them. Such discussions sometimes result in the writing of a personalized user
guide that is distributed within certain social groups (e.g., students).
The Observation of Human-Machine Interaction

We now turn to the way people actually use word processors and examine the
specific nature of this interaction . One particular issue of interest is whether the
technological qualities of the program on the one hand or the user's technological
competence on the other hand are sufficient conditions to account for the actual
use of a word processor. To discuss this point, we look at an aspect that is
usually referred to as the "user friendliness" of a program. In fact, user friendliness
is generally assumed to be one of the most important conditions for the efficient
use of a program. There is, for example, an ongoing debate regarding the
respective user-friendliness of the MacIntosh or AMIGA compared with the IBM.
Given our present knowledge about communication processes, we could
question the concept of user friendliness itself. To say that a program is or is not
user friendly implies that the user's interpretation of the messages is neglected.
The direct observation of users shows that the interpretation of a message cannot
be taken for granted and that any message contains some ambiguity. So the
question is: How does the user interpret the messages of the machine? Observing
human-machine interaction shows how misunderstandings emerge. Some
examples follow:
- In the section concerning the historical development of word processors, we
mentioned the origin of the OPEN FILE command . Observing users shows that,
despite its present-day lack of meaning, the wording of this command has become
part of computer literacy: When expert users have to use a word processor with
the command READ FILE or READ TEXT, they do not immediately understand
its meaning. So, even if this new label seems to be more accurate in describing
the kind of operation involved, it is inefficient in this context because the history
of human-machine interaction has given a specific meaning to certain wordings .
- The second example concerns the use of the SAVE command, which many
beginners find difficult to remember. With the evolution of word processors,
some protection systems have been added to remind users to save their text.
Messages such as DO YOU WANT TO SAVE THE CHANGES? appear on the
screen . However, with the latest developments, systems have integrated automatic
saving so that it is no longer necessary to save. On a technological level, the
user's task has been simplified, but actual user behavior shows that this
simplification does not seem self-evident: In the meantime, users have been
socialized to the use of the SAVE command, and some of them do not easily
accept that work no longer has to be saved manually! The use of the computer has
created a specific culture, and, from this point of view, a user-friendly interface
282
may be better defined as a state of intersubjectivity resulting from the user's

technological experience and current interaction with the machine.
- The third example of how misunderstanding occurs concerns the interpretation
of interface messages. When using the word processor WORD, some users are
puzzled by the CANCEL message that appears after the activation of a command.
Their question is exactly what will be canceled if they validate this option: their
own text or the requested operation? When faced with this ambiguity, some users
prefer to give the order OK rather than run the risk of loosing their text. But, by
settling for this option, they take another risk: that of being unable to cancel the
resulting operation, thus entering a vicious circle! Consequently, once they
became aware of this ambiguity, the designers of the program changed the
CANCEL option to CONTINUE. But users who, in the meantime, had become
used to the former label, again asked the same question: What am I to continue?
This example shows that every message contains a certain amount of irreducible
ambiguity that has to be interpreted in order to understand the message.
- The fourth example refers to the Macintosh interface, where folders can be
represented by icons. The designers intended for this graphical presentation to
provide a concrete representation reflecting everyday life. Nevertheless , as a
representation of an object (and not the object itself), it introduces an abstraction
that might make users feel uneasy: For example, a user of WORD explicitly
called her folders "folder XYZ," which would indicate that the icon itself was not
explicit enough . In a real office environment, it is rather unlikely that someone
would call his or her file "file XYZ ." In this context, a label of this sort would be
seen as a kind of pleonasm, whereas in the virtual office represented by the
interface, it is not.
- The last example illustrates the concept of modification of use. We observed a
user of WORD who used the command SUBSTITU1E (which gives the frequency
of a substitution) to count the frequencies of turn-taking in a dialogue. In this
case, a modification of use introduces a kind of technological innovation that is
not foreseen by the designer.
Some Predictions About the Future Development of Word

Processing
Having reported some elements of the historical development of word processors
and having shown that user-computer interactions create a specific interactional
space, we find the temptation to make some predictions about the future
development of word processors hard to resist. Word processors have integrated a
number of tools that were initially distinct programs: for example, spreadsheets
and calculators. Moreover, they enable us to link these programs to an
encyclopedic amount of data. Bearing this in mind alongside their historical
evolution , we believe it is feasible that word processors will once again become a
single piece of hardware with many functions , the name of which is still to be
283
decided on (see the world of Dynabook to refer to Alan Kay's visions; Kay, 1992,
p. 199) . It would include multimedia capabilities and communication
possibilities. Thus, future word processors could, like the first ones, be machines
dedicated to that purpose only.
Conclusion
Using an interactional approach, our principal aim in this chapter has been to
show that human-machine interaction creates specific interactional spaces and that
an understanding of these requires that they be observed in concrete contexts.
At the current stage of our study, three main conclusions may be drawn:
- The first conclusion is that human-machine interaction is an interindividual
relationship mediated by a technological tool and consisting of an indirect
dialogue between users and designers.
- The second conclusion is that this indirect dialogue has many of the same
characteristics as any other communication between individuals, and it requires
interpretation activity on the part of the user in order to understand the logic of
the machine. Hence, ambiguity, negotiation of meanings, and conflicts resulting
from the confrontation of differing perspectives and assumptions are fundamental
parts of this indirect dialogue . Thus, the transparency- of a system raises a more
complex issue than is generally believed . An example is provided by the fact that
the implementation of database structures or the definition of fields (type and
length) is no longer made necessary by the designer. Users nevertheless come up
against these implicit definitions when they carry out unusual operations (e.g. ,
when going over the number of permitted characters or when changing the data
type) .
- The third conclusion is that this indirect dialogue results in a specific
interactional space that arises out of the intersection between the technological
qualities of the machine and the user 's own activities and skills . From this point
of view, technological innovations might also be considered to be a modification
of use resulting from the resistance of the machine to obey the user 's will and the
user 's interpretation of the use to which the machine was originally dedicated .
This chapter represents only an initial attempt to present a comprehensive
theoretical framework capable of accounting for heterogeneous observations. Such
an attempt involves some risks, one of which is the presentation of too generic a
3Computer scientists consider a procedure to be transparent if the different operations
necessary for it to be carried out effectively by the machine are imperceptible to the
user. This particular use of the term transparent (which hides by rendering "invisible")
reflects, on the part of those who conceived of it, a certain repre sentation of the user
and a strategy to make his use as natural as possible.
284
conception of certain notions , such as user, designer, and activity. Throughout

this chapter, we have discussed some possible misunderstandings between
different partners, mainly computer scientists , designers, teachers, and users.
However, one could argue that the term user in general is unclear: What is a user?
Is it possible to give a general description of a user? This kind of debate is
reminiscent of another famous debate concerning the notion of the epistemic
subject in Piaget's work. The same point was made: The Piagetian subject was
not psychological; it would be impossible to meet her! The same can be said of
terms such as activity or text in general: what activity, what text? For what, for
whom, for what purpose? Our previous work in the field of developmental
psychology has made us very aware of this question, and we believe that it should
receive our full consideration. The next step in our research, therefore, will focus
in more detail on specific users and will attempt to identify in detail the type of
activity in which they may engage.
Acknowledgment
The authors wish to thank Anne-Marie Rifai for her translation of this chapter
from French.
Correspondance concerning this chapter should be addressed to M. Grossen,
Faculte des Sciences Sociales et Politiques, University of Lausanne, BFSH 2,
CH-1015 LAUSANNE, Switzerland.
285
References
Alter, N. (1985). La bureautique dans I'entreprise [Bureaucracy in busines s]. Paris: Les
Edition s Ouvrieres .
Appel , K. , & Haken, W. (1977). La solution du probleme des quatre couleurs [The
solution to the four-color problem]. Pour la Science, 2, 56-70.
Audetat, M. (1993) . Le voyageur Butor en zone de transit [The Butor voyager in the
transit zone]. L 'Hebdo, 4, 61.
Bangert-Drowns, R. L. (1993) . The word processor as an instructional tool: Ametaanalysis of word processing in writing instruction . Review of Educational Research,
63(1), 69-93 .
Belyaeva, AV., & Cole, M. (1989). Computer-mediated joint activity in the service of
human development: An overview . The Quarterly Newsletter of the Laboratory of
Comparative Human Cognitive, 11(3), 45-66.
Butor , M. (1992). Transit. Paris: Gallimard .
Cohen, H., & Nordon, C. (1989) . L'arithmetique assistee par la geometric et
l'ordinateur [Arithmetic assisted by geometry and the computer]. La Recherche,
20(208), 352 -358 .
Cuisset, P. (1986) . Les voyageurs du temps [Time travelers] . Delphine Software

Cinematique,
Doise, W., & Mugny, G. (1984). The social development of intellect. Oxford,
England: Oxford University Press. (Original work published 1981)
Doise, W., Mugny, G., & Perret-Clermont, A-N. (1975). Social interaction and the
development of cognitive operations. European Journal of Social Psychology, 5 ,
367-383 .
Dubuisson, C., & Weiss, J. (1989). What computers can bring to the learning of
written communication: Some promising perspectives. Neuchatel: Institut Romand
de documentation et de recherches pedagogiques ,
Edwards, D., & Mercer, N. (1987). Common knowledge: The development of
understanding in the classroom. London: Methuen.
Elbers, E., Maier, R., Hoekstra, T., & Hoogsteder, M. (1992). Internalization and
adult-child interaction. Learning and Instruction, 2, 101-118.
Faggin, F. (1992). The birth of the microprocessor. Byte, 17(3), 145-150 .
Gaudin, T. (1978). L'ecoute des silences [Listening to silences]. Paris: Union generale
d'Editions.
Gaudin, T. (1988). Les metamorphoses dufutur [Metamorphoses of the future]. Paris:
Editions Economica.
Gillet, B. (1990). De l'usager-type aux types d'usagers ou l'influence des differences
individuelles dans l'adaptation homme-ordinateur [From user-type to types of user or
the influence of individual differences on man-computer adaptation]. In Informatique
et differences individuelles (pp. 111-123). Lyon, France: Presses Universitaires de
Lyon .
Gras, A, Moricot, C., Poirot-Delpech, SL, & Scardigli, V. (1990) . Le pilote, le
contrbleur et l'automaie [The pilot, the controller and the automaton] . Paris:
Editions de l'Iris.
Greenfield, P.M. (1984) . Mind and media: The effects of television, computers and
video games. London : Fontana.
286
MicheleGrossenand Luc-OlivierPochon
Grossen, M. (1988) . L'intersubjectivite en situation de test [Intersubjectivity in a test

situation]. Cousset: Delval.
Gros sen, M., & Perret-Clermont, A.-N. (1994) . Psycho-social perspective on
cognitive development: Construction of adult-child intersubjectivity in logic tasks.
In R. Maier & W. de Graaf (Eds.). Sociogenesis reexamined (pp. 243-260) . Berlin:
Springer-Verlag.
Grossen, M., & Pochon, L.-O. (1988) . Rapport sur l'utilisation du Nano-reseau a
l'ecole primaire [Report on the use of the Nano-network in primary schools] .
Cahiers de Psycho logie, 27, 51-66.
Guggi sberg, V., & Balmelli, A. (1992) . Usage d'une calculatrice de poche dans
l' enseignement specialise [The use of the pocket calculator in specialized teaching].
A propos d'une experience tessinoise . Math-Ecole , 151, 23-28.
Kay, A. (1992) . A Kay user interface : A personal view. In L. Brenda (Ed.). The art of
human computer interface design. Reading , England: Addison Wesley.
Kraut, R., Galegher, J., Fish, R., & Chalfonte, B. (1992). Task requirements and media
choice in collaborative writing. Human-Computer Interaction, 7, 375-407 .
Levy, P. (1990) . L'Hypertexte comme technologie intellectuelle et metaphore
[Hypertext as intellectual technology and metaphor]. In Informatique et differences
individuelles (pp. 59-68) . Lyon, France : Presses Universitaires de Lyon .
Light, P., & Butterworth, G. (Eds.). (1992). Context and cognition: Ways of learning
and knowing . London : Harvester Wheatsheaf.
Mevarech, Z.R., & Light, P. (Eds.), (1992) . Cooperative learning with computers
(Special issue). Learning and Instruction, 2(3).
Nunes, T., Light, P., & Mason, J. (1993). Tools for thought: The measurement of
length and area. Learning and Instruction, 3, 39-54 .
Papert, S. (1980) . Mindstorms. children, computers, and powerful ideas. New York:
Basic Books.
Papert, S. (1993) . The children's machine: Rethinking school in the age of the
computer. New York : Basic Books.
Pea, R.D. (1985) . Integrating human and computer intelligence. In E. L. Klein (Ed.),
Children and computers (pp. 75-96). San Francisco: Jossey-Bass,
Perret-Clermont, A.-N. (1980) . Social interaction and cognitive development in
children. London : Academic Press. (Original work published 1979)
Perret-Clermont, A.-N., & Nicolet, M. (Eds.). (1988) . Interagir et connaitre
[Interacting and knowing]. Cousset, Switzerland: Delval.
Perret-Clermont, A.-N., Perret, J.-F., & Bell, N. (1991) . The social construction of
meaning and cognitive activity in elementary school children. In L. B. Resnick,
J. M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition
(pp. 41-62). Washington, DC: American Psychological Association .
Perriault, J. (1989). La logique de I 'usage. Essai sur les machines a communiquer [The
logic of use. An essay on communication machinery]. Paris : Flarnmarion .
Pochon, L.a. (1991) . Impact de l 'informatique sur la formation de concepts
mathematiques [The impact of computers on the formation of mathematical
concepts] . Neuchatel, Switzerland : Institut Romand de Recherches et de
Documentation Pedagogiques.
Podak , C. (1993) . Informatik und Gesellschaft: Defizite und Aussichten [Computer and
society : Deficits and perpectives] . Interface , 2, 2-3.
InteractionalPerspectives on the Use of the Computer
287
Resnick, L. B., Levine, J. M., & Teasley, S. D. (Eds.). (1991) . Perspectives on

socially shared cognition. Washington, DC: American Psychological Association.
Rheingold, H. (1992) . Virtual Reality. NewYork: Simon & Schuster.
Rogoff, B. (1990) . Apprenticeship in thinking: Cognitive development in social
context. New York: Oxford University Press .
Rommetveit, R. (1979) . On common codes and dynamic residuals in human
communication. In R. M. Blakar (Ed.), Studies of language, thought and verbal
communication (pp. 457-466). London : Academic Press .
Rommetveit, R. (1992) . Outlines of dialogically based social-cognitive approach to
human cognition and communication. In A. H. Wold (Ed.), The dialogical
alternative. Towards theories of language and mind (pp. 19-44). Oslo : Scandinavian
University Press .
Saljo, R. (1992). Human growth and the complex society. Cultural Dynamics, 5(1) ,
43-56.
Schubauer-Leoni, M. L., & Grossen, M. (1993). Negotiating the meaning of questions
in didactic and experimental contracts . European Journal of Psychology of
Education, 4(8), 451-471.
Schubauer-Leoni, M.L., Perret-Clermont, A.-N., & Grossen, M. (1992) . The
construction of adult-child intersubjectivity in psychological research and in
school. In M. Von Cranach, W. Doise, & G. Mugny (Eds.), Social representations
and the social bases of knowledge (pp. 69-76). Bern, Switzerland : Hans Huber.
Sproull, L., & Kiesler, S. (1991). Computers, networks and work. Scientific American,
265(3), 84-91.
Vacherand-Revel, J., & Bessiere, C. (1990) . L'interaction graphique et la prise en
compte des differences entre usagers dans la formation d'adulte assistee par
ordinateur [Graphic interaction and the consideration of user differences in computer
assisted adult education] . In Informatique et differences individuelles (pp. 21-52) .
Lyon : Presses Universitaires de Lyon.
Vitale , B. (1989) . Elusive recursion : A trip in recursive land. News Ideas in
Psychology, 7(3), 253-276.
Vygotsky, L. S. (1978) . Mind in society. The development of higher psychological
processes. Cambridge, MA: Harvard University Press.
Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: MIT Press . (Original
work published 1934)
Waern, Y. (1993) . Varieties of learning to use computer tools . Computer in Human
Behavior, 9, 323-339.
Wertsch, J.V. (1991) . Voices of the mind: A sociocultural approach to mediated
action. London: Harvester Wheatsheaf.
Who Made the Micro . (1991). Byte, 16(1),304-314.
Chapter 12
What Organizes Our Problem-Solving

Activities?
Terezinha Nunes
Institute of Education, University of London, UK
Abstract
Structuralist models of cognition emphasize the role of reasoning principles,
invariants, or rules in the organization of thinking and pay little attention to the
culturally developed systems of signs and their impact on thinking. However,
complex psychological functions , as pointed out by Vygotsky (1962) and Luria
(1979), are always carried out with the mediation of historically developed and
culturally transmitted systems of signs . Within this perspective, systems of signs
have a major role in thinking because they are central to reasoning processes .
This chapter presents research that supports the hypothesis that systems of signs
playa structuring role in problem-solving activities as they are used to support
skilled action .
Two sorts of empirical findings are reviewed . The first line of work analyzes
how cultural practices that rely on diverse systems of signs affect their users '
performance as they solve mathematical problems. The second line of work is
experimental and examines how the use of different measurement systems affects
children's problem-solving performance. The results reported here suggest that
schools need to consider carefully the systems of signs that they select for
transmission in school, in terms of both the difficulties they create for learners
and their power as tools for thought.
What Organizes Our Problem-Solving Activities?
289
Among the many current challenges to traditional theories of intelligence is the

observation of large within-individual differences in the same ability. As theories
of intelligence have developed, within-individual differences have been treated
either as manifestations of different intelligences (e.g., Thurstone, 1938; Gardner,
1993) or as fluctuations in performance to be explained by transitory factors, such
as fatigue, emotional state, and inhibition by the experimental setting . A
considerable amount of evidence indicates, however, that the same people may
consistently show radically different levels of skill in the same ability (e.g.,
logical or mathematical reasoning) in different situations' and that these
differences cannot be pushed aside under the guise of false negatives (or false
positives) stemming from motivational, affective, or physical intervening
variables. Such within-individual differences in the same ability are inconsistent
with traditional views of intelligence, which locate intelligence within the
individual.
The evidence reviewed by Lave (1988) on mathematical knowledge reveals clear
discontinuities in reasoning by the same subjects in experimental situations,
tests, and everyday settings. Lave suggests that these differences are accounted for
in a theory of cognition as situated: that is, a theory according to which cognition
always takes place in particular settings, drawing on the specific resources
available, and where particular forms of explanations and answers are valid (but
not others). She argues that objects in an everyday setting - such as a pot of
cottage cheese containing a specific amount of cottage cheese - contribute in
significant ways to determining problem solutions just as written symbols
contribute to structuring problem solutions in the laboratory or in schools.
Objects present in a situation and systems of signs used by individuals affect
problem solving through the possibilities of action that they create for the
subject. Thus, Lave argues that problem solving (or intelligence) cannot be
accounted for by models that take into account only the individual. Processes in
solving problems involve an interaction between the subject and significant
aspects in the situation (for other discussions of this issue, see Light & PerretClermont, 1989, and Nunes, 1994).
In this chapter, I explore this facet of cognition : the way in which symbols
structure thinking in relation to particular activities. There are, of course, other
significant elements in a situation that influence reasoning (such as the goals of
the activity). The very choice of thinking tools may be determined not only
within the situation but also from its relation to sets of beliefs that have a more
general character (such as the use of writing in school). However, it is possible to
1 The term situation is used with different meaning s throughout the literature . I use it
here to refer to the situations about which people are thinking. These include the social
situation in which the subjects find themselves and the imaginary situation about
which they may be thinking.
290
Terezinha Nunes
examine the structuring role of symbols on thinking without analyzing in detail

these other aspects that, nevertheless, playa role.
This chapter is organized in four sections. The first section briefly raises some
points about what may be situated in cognition and what may be more general.
The two subsequent sections each describe a line of inquiry into the role of
symbols in mathematical problem solving. The first series of studies examines
cultural practices of mathematics in and out of school and discusses their effects
on problem solving. The second set of studies considers how the offer of different
objects to children in measurement problems affects their performance. The last
section briefly presents some general conclusions from the experimental work
reviewed here.
What is Situated and Non-Situated in Cognition?

Even young children who need to count on their fingers to solve the simplest
addition problems do not think in ways that are entirely tied to context. For
example, if you ask a five-year-old to solve the problem, "Mary had 3 sweets; her
mother gave her another 2; how many does she have now?", the child may count
on her fingers in order to produce the answer. This simple action of counting
fingers to obtain an answer to a problem about sweets demonstrates the child's
understanding that whatever results apply to fingers also apply to sweets. In this
sense, the child's knowledge of addition is not situated. The child holds a general
model. Modeling in mathematics means exactly this: using methods and results
known from one domain to solve problems in another domain. However, the
same five-year-old child may not be able to answer the question "What is 3 and 2?
(or 3 plus 2; or 3 and 2 more)." Hughes (1983) found significant differences
between children's ability to solve hypothetical problems about putting bricks
into a box or buying sweets at a shop and their ability to answer addition
questions that did not refer to specific situations. One could say that young
children's difficulty in dealing with questions about numbers outside a commonsense context derives from the situated nature of their knowledge; that is to say,
they only understand addition in contexts that make sense to them. Yet this
explanation ignores the patent modeling ability demonstrated by young children's
use of fingers to solve problems about other situations.
This very simple example captures in my view one of the major difficulties in
the study of cognition: the fact that knowledge can be both general and specific at
the same time. The general aspects have traditionally been described in terms of
the structure of knowledge or the principles underlying problem-solving
strategies . The specific aspects have correspondingly been described in terms of
content or object of knowledge in which the principles of reasoning are embedded.
291
The separation between the structure and content of thinking is helpful in many
ways. It is a powerful heuristic for the analysis of cognition in many
circumstances . It is particularly useful when subjects cannot articulate the
reasoning principles they are implicitly using (e.g., in the case of young children)
or when the very articulation of the principles may alter our observation (by
making the principles of reasoning into the object of thinking). It is also a
powerful way of demonstrating the similarity of competencies underlying
markedly different performances. Elsewhere (Carraher, 1985) I have argued, for
example, that Brazilian working-class and middle-class children in first and second
grade, who show markedly different error rates in spelling , appear to rely on the
same phonological principle to generate the orthography of words. Although they
use the same phonological principle to generate spelling, when they apply this
principle to their knowledge of the language (here, the content to which the
principle is applied), they come up with very different results . Middle-class and
working-class Brazilian children speak different varieties of the language, and the
middle-class children speak a dialect that is closer to the idealized and prestigious
form of Portuguese used as the source for orthographic representation. Thus,
middle-class children using the conception of phonological representation of words
generate fewer spelling errors than working-class children using the same
conception.
In his vast work, Piaget suggested that the logic of actions was the source of
these general structures of reasoning, whereas abstractions regarding objects
offered the knowledge of contents or empirical knowledge. The logic of actions is
not a question of particular events but of general representations of actions such as
grasping, adding, and separating. This hypothesis is perfectly consistent with the
notion of situated cognition. Its basic assumption is that subjects develop their
knowledge of situations and objects, on one hand, and their reasoning structures,
on the other, from their own organization of actions (or, more generally speaking,
their participation in situations). This hypothesis is also consistent with the
notion that knowledge may be both general and specific. When actions are
represented in their generalized form, as schemata, the specifics of the situation do
not constrain the knowledge of principles .
However, this hypothesis does not offer a complete description of knowledge. It
is necessary to examine the way in which knowledge is mediated by systems of
signs, because we cannot assume a direct interphase between the subject and the
objects of thought. The mediating systems of signs cannot be treated as byproducts of the subject's reasoning principles, because many of the mediating
systems of signs are not invented by the subject but are culturally transmitted.
Piaget (1962) argued that the child's ability to use collective systems of signs
still depends on the development of the child's thought, because children need to
construct meanings for the signs. Although this may be largely true, he failed to
acknowledge that, once the system is acquired, it also comes to structure
reasoning . Thus, a more complete description of the development of reasoning
would encompass a description of the uses of culturally transmitted systems of
292
Terezinha Nunes
signs . This is the core of the hypothesis that I explore here. More specifically, I
suggest that reasoning and problem solving are organized by two factors: the
generalizable principles related to the logic of situations and the organization of
the systems of signs used in the representation of particular situations. The latter
have sociocultural significance and not only make knowledge more situated but
also color it with connotations of affect and valorization.
In the next two sections, I review studies indicating that systems of signs play
a major role in reasoning. The first section reviews evidence from participation in
different types of activities that are characteristically associated with the use of
different systems of signs . The second section explores the effect of a single
experience with different ways of representing area and its consequence for
subsequent performance in solving problems about areas.
Different Arithmetic Practices

Brazilian working-class children are exposed to different cultural practices for
solving arithmetic problems, which I refer to as oral and written arithmetic
because of their basic support either by oral or by written numeration systems (for
further details, see Nunes, Schliemann, & Carraher, 1993). These practices have
different characteristics that are explored later.
Written practice is systematically taught in school. Teaching starts from the
assumption that children need to learn first how to write numbers and then the
computational algorithm for each operation. After learning the algorithm, children
can then apply this knowledge in problem solving (see Carraher, Carraher, &
Schliemann, 1985).
The teaching of multidigit addition and subtraction is carried out by focusing on
the written numeration system . Children first learn to write numbers by using the
"place value frame" (see Figure 1). This frame contains pockets into which sticks
are placed. Children are taught the convention that only nine units can be placed
within any pocket. If they have ten sticks, they are taught to tie the sticks into
bundles of ten and place these bundles into the next-to-the-left pocket of the
frame. They then write the number of sticks on sheets of paper similarly divided
into columns; the number of bundles of tens goes into a column headed "tens,"
and the number of singles goes into a column headed "units." The groups of tens
are later tied into bundles of ten-tens, forming one hundred, reinforcing the
principle of a maximum of nine units of any value in each position. Writing and
reading numbers is closely associated with this practice for a few lessons. Later
on, the sticks are set aside with the assumption that the children have grasped the
principles of place-value notation and can rely on these principles from then on.
Multidigit addition and subtraction are then started, one at a time . Multidigit
addition is taught through the explanation of carrying . When more than nine units
293
in an order are obtained in addition, regrouping needs to be done, and one unit
(meaning one ten, one hundred, etc.) will be carried over to the next column .
Typically, teachers refer to the rules learned for writing numbers as they
demonstrate the procedure on the chalkboard.
Fig. 12.1. The Place Value Frame (Quadro Valor de Lugar) for practicing written
numbers . The frame is composed of pockets labeled C (Centenas, meaning hundreds), D
(Dezenas, meaning tens), and U (Unidades, meaning ones) into which singles, bundles
of ten, or bundles of one hundred sticks are placed . When practicing how to write
numbers, children write the number of units in each pocket on a sheet designed in the
same way.
Although the addition facts need to be mentally produced by the child, I call these
algorithms written arithmetic because the written symbols become a medium onto
which the intermediary results of different steps are off loaded (see Hatano, in
press) . After instruction on the algorithm, children carry out large numbers of
computation exercises and word problem solving , the aim of which is to offer
further practice in the procedure. Multidigit subtraction is similarly taught with
reference to the previously learned principles of place value notation. The
borrowing procedure starts with the notion that "you can't take a larger number
from a smaller one," and thus you borrow one (one ten, one hundred) from your
neighbor. Once again, subtraction facts have to be produced by the children, and
the sequential results are off-loaded onto the written representation. As with
addition, extensive amounts of practice with the algorithm and in solving
multidigit subtraction problems follow instruction. Figure 2 exemplifies the
subtraction procedure and typical instruction provided.
Three significant characteristics of these algorithms are of interest here. First,
computation is carried out from right to left (that is, from smaller to larger). This
294
Terezinha Nunes
procedure leads to a progressive amplification in the magnitude of errors and often

to a loss of the possibility of monitoring the results. Second, in order to make the
steps in the algorithm general, the procedures set aside the distinctions between
ones, tens, hundreds, etc. as they are executed. Children carry and borrow "one,"
the value of which may be ten or one hundred. Third, checking the results, when
stimulated by teachers, is done by repetition of the same procedure. Even when
word problems are used in the practice of the algorithm, children are not asked to
use their understanding of the situation to evaluate the results, because the
effective aim of the teacher is to offer practice in the procedure. All of these
characteristics result in a divorce between written arithmetic and reality, as Reed
and Lave (1981) have suggested.
Oral procedures are not transmitted in school but rather depend on children's
participation in situations where numbers are used . Although oral procedures are
not restricted to shopping situations (e.g., Carraher, 1986, has documented oral
procedures used by foremen on construction sites without any reference to
money) , interactions involving money are often surrounded by oral arithmetic.
Oral procedures are often exemplified in transactions in street markets, where a
vendor will calculate aloud total costs of purchases for customers and count the
change up from the value of the purchase to make sure that it is the correct sum.
Counting money is, in fact, adding up the values of notes, an addition procedure
that is observed in strict connection with the manipulation of symbolic
representations of value.
In contrast to written practices, oral procedures do not set aside the distinctions
between ones, tens, and hundreds. Thus the sense of the amounts being handled is
not lost during computation. A second contrast between oral and written
arithmetic relates to the direction of calculation: Oral practice tends to operate
from larger to smaller. These characteristics, plus the central significance of the
situation, help oral arithmetic users to keep meaning in mind. An example of oral
subtraction is provided in Figure 2.
Because oral and written arithmetic can be used to solve a large variety of
problems, it is possible to study the impact of symbolic systems on problem
solving without being restricted to a particular type of problem. It is also possible
to use a variety of designs so that what may be confounded in one study is not a
source of ambiguity in the interpretation of the results of another study. Next, I
review studies using three types of design in the analysis of oral and written
arithmetic in Brazil: (1) within-subject designs, where subjects are presented with
problems that they can solve through any method they wish, but in situations
where different practices are likely to be elicited; these studies control for the
subjects' general ability but can be confounded by the possibility of interference
between the oral and the written system; (2) across-group comparisons, which
select groups with large amounts of either oral or written arithmetic practices but
are ambiguous with respect to the subjects' general level of ability ; and (3)
experimental designs, where subjects are randomly assigned to the use of either
295
oral or written methods, thereby controlling for ability through randomization (for
a review, see Nunes , Schliemann, & Carraher , 1993).
A.
The written subtraction algorithm.

21
216
129
1.
Children are taught to align the digits from right to left (as shown
above) , often initially using notation sheets such as those presented in
Figure 1. Subtraction is also carried out from right to left.
2.
When subtracting with borrowing, children are asked whether they

could take 9 away from 6. Upon obtaining a negative answer from the
children, teachers then indicate that the six could borrow one ten from
the three next to it and there would be 16 units. Children are then
taught to cross out the 3 in the tens columns and write 2 above the
crossed-out 3 (as shown above) and write a small 1 next to the six ,
indicating that it is now 16. From 16 units, one can take 9 away. The
children need to produce the result of this subtraction and write it down
in the column of units.
3.
For each column, the same procedure is repeated . Each digit is referred
to during the application of the algorithm as units; that is, the 3 in 236
is spoken of as three and not thirty, and the 2 is not referred to as two
hundred . This allows the children to use the same knowledge of
subtraction facts regardless of whether the operation is on the column
for units, tens, or hundreds .
B .
An example of oral subtraction.
The problem presented to the child involved taking 57 away from 252 .
The child (Fabio, 3rd grader) first answered "185" after speaking very
low. When the examiner asked the child to explain it aloud, the child
said: "Two hundred and fifty two and fifty seven. I take fifty away, that
is two hundred . And two and seven. Five to take away, then, one
hundred and ninety five."
Note the direction of calculation (first, fifty is subtracted and then
seven) and the maintenance of the magnitude in focus during
calculation (the child does not say "five" but "fifty").
Fig. 12.2. An example of subtraction with borrowing and typically provided
instruction (top) and an example of oral subtraction.
296
Terezinha Nunes
Two within-subject studies were carried out in which the same children solved
problems in different situations that were likely to elicit different arithmetic
practices. In the first study, the children were interviewed first in street markets as
they carried out their sales and later on in a school-like situation, answering
questions about computation exercises and word problems (Carraher, Carraher, &
Schliemann , 1985). In the second study, the children were interviewed in schools
but were placed in a pretend-shop situation and in a school-like situation
(Carraher, Carraher, & Schliemann, 1987). Both studies revealed that children
performed significantly better in the vending situations (real or pretend) than in
the computation exercises.
Qualitative analyses of both oral and written procedures were also carried out in
the second study (Carraher, Carraher, & Schliemann , 1987). Children's
verbalizations during the interviews were tape-recorded, and, together with notes
taken by the interviewers, allowed for a detailed description of the steps used by
the children during calculation.
Written procedures, whether correct or incorrect, were attempts to reproduce the
algorithms taught in school. Errors observed in these procedures were of the same
nature described in the literature (see, for example, Brown & Burton, 1978;
Lawler, 1990). In subtraction, for example, some children took the rule "You
can't subtract the larger from the smaller" to imply "Subtract the smaller from the
larger." Thus, in calculating 252 - 57, some children took 2 away from 7 and
obtained the answer 205. In multiplication, for example, when calculating 15 x 5,
some children calculated 5 x 5 = 25, correctly wrote down 5 and carried the 2 but
then added 2 onto 1 before multiplying, calculating 3 x 5 = 15, and obtained the
result 155. Thus, even incorrectly used algorithms demonstrated children's
attempts to follow the rules that they had been taught but clearly had not
mastered.
As mentioned earlier, oral arithmetic is not taught systematically in school.
Thus, the qualitative analysis of these procedures involved the search for
underlying principles rather than a comparison between what had been taught and
what the children produced. In spite of great variations in the surface of the
problem-solving procedures, addition and subtraction problems seemed to be
described mostly by one method, referred to in the literature (e.g., Plunket, 1979)
as decomposition. Examples of decomposition in addition and subtraction are
presented in Table 1.
A careful consideration of these examples shows that decomposition implicitly
uses the property of associativity of addition and subtraction as a principle. In
other words, when children use decomposition, they seem to understand that it is
possible to divide into parts a number to be added or subtracted and to carry out
the addition/subtraction in steps, by subtracting each part sequentially. Errors in
oral addition and subtraction also revealed the same underlying principle, because
they were mostly memory errors. When a number was divided into more than two
parts, sometimes one of the parts was eventually forgotten, and the final response
was incorrect. For example, when subtracting 57 from 252, one child divided 252
297
into 250 and 2 and 57 into 50 and 7. He then subtracted 50 from 250, obtaining
200 and, subsequently, 7 was subtracted from 200. His final answer was 193
because he seemed to have forgotten the 2 that he had set aside.
Table 12.1. Some examples of oral addition and subtraction.

Robson (solving 215 + 15): "Add 5, 120; all together, 130."
Fabio (solving 168 + 75):
"Sixty with seventy, one hundred and thirty.

Two hundred and thirty. Eight and five, twelve
[sic]. Two hundred and forty two."
Edilene (solving 243 - 75) : "You give me 200, I will give you 25. With the
43 that you have, 143, you will have 160, 168."
Carlos (solving 500 - 80):
"Eighty plus twenty is one hundred. Four

hundred and twenty ." (Subtraction is calculated
here by the complementary addition)
Pedro (solving 263 - 68):
"Take away 60. It makes 200. That will be 205."
Examples from Carraher, Carraher , and Schliemann, 1987 ; examples not

included in original publication.
It is noteworthy that associativity, the reasoning principle that seems to be the

basis for oral addition and subtraction, is exactly the same principle implicitly
used when we add/subtract using the written arithmetic algorithms. Thus, the
children's discrepant performance across oral and written arithmetic cannot be
explained as a failure to understand the reasoning principles required for computing
addition and subtraction. Rather, the children in this study, who were exposed to
the two practices of arithmetic, clearly mastered one system of signs as a tool for
calculation but not the other. Thus, the tool they chose to use played a role in
determining their success and the type of error they made.
Type of error was influenced by the system of signs used in two ways. The first
one was described earlier: Written arithmetic errors were attempts to use the
algorithm but the rules were permeated with bugs, whereas oral arithmetic errors
often involved memory. The second influence was related to the monitoring of
results, which is part of calculation in oral arithmetic but more difficult to carry
out in written arithmetic. In written arithmetic, the position of a digit (its place
value) is ignored during calculation, thereby leading the focus of attention away
from the magnitude of numbers. Thus, errors that deviate from a reasonable
magnitude are not detected during calculation. Moreover, written calculation
follows the smaller-to-larger direction. Errors are multiplied by ten as calculation
298
Terezinha Nunes
proceeds. Consequently, errors are of a larger magnitude in written than in oral

arithmetic.
Figure 3 presents the percentage of correct responses for oral and written
addition and subtraction, as well as the percentage of errors in different error bands.
The first error band is one in which the responses fell within 10% of the correct
response (e.g., if the correct response was 20, errors between 18 and 22 would fall
within this error band). The second error band includes errors larger than 10% of
the correct response and smaller or equal to 20%. The last error band includes
errors of a higher relative magnitude. The association between error band and type
of procedure used was statistically signiftcant.
BD
.oral calculation
D'M"ltten ca lculaIlon
III
CIl
0>
60
.sc:
...
CIl 4ll
U
CIl
11.
20
COrrect
'M"ong <10::1 I/ttong >10::1 <20:1 """ong >20:1
Response type
BO
Oral calculation
D'M"'tten calculallon
60
III
CIl
0>
.sc
...u
CIl
4Q
CIl
11.
20
correct
Vrtong <10:1 I/ttong >10:1 <20:1 'M"ong >20:1
Response type
Fig . 12.3 . Percentage of correct responses and of errors in the different error bands
for oral and written addition (top) and subtraction (bottom).

Table 12.2. Some examples of oral multiplication and division.
Edilene (solving 15 x 50):
"One hundred, two hundred, three hundred, four

hundred, five hundred. Ten cars. Then I add one
hundred, six hundred. Seven hundred and fifty ."
Jose (solving 4 x 25):
"Twenty five, twenty five, 50. Two will be fifty.

Four will be 100."
Pedro (solving 15 x 50):
"Fifty with fifty, one hundred. Four is two

hundred . Five [sic] is three hundred. Or, I think,
two are one hundred. Four are two hundred . Six
are three hundred. Eight are four hundred. (pause)
Seven hundred and fifty."
Edilene (solving 100 + 4):
"Divide by two, fifty . Divide by two again,

twenty five."
Ronaldo (solving 75 + 5 in the context of a word problem that required

dividing 75 marbles among 5 boys):
"Ten marbles each, that's fifty . Twenty five left
over. To divide for five boys . That's hard. (E:
Yes, that's a hard one). That's five more each,
that's fifteen each."
Ronaldo (solving 100
4):
''Twenty each, that is eighty. So it will be more .

Twenty five. Twenty five, fifty , one hundred .
Then it's twenty five ." (Note he attempts to
multiply the divisor by a number that
approaches the dividend without exceeding it ;
this is also the way the written division
algorithm is expected to proceed .)
Robson (solving 75 + 5 in the context of a word problem that required dividing

75 marbles among 5 boys): "Thirty each. Sixty for two. No. Thirteen each .
Twenty six for two. Forty , fifty two. No. I think
it's about 17 marbles for each one." (Note that
he adjusted the value without completing the
computation. He realized that 30 marbles for
each is too many after figuring out that two boys
will get 60. He then tried out 13 and apparently
realized that, if four boys get 52, the fifth boy
gets too many . Instead of completing the
computation, as did Ronaldo in the preceding
example , he stops with an estimate .)
Examples from Carraher, Carraher, and Schliemann, 1987; example s not
299
300
Terezinha Nunes
Qualitative analyses of oral procedures in multiplication and division were also

carried out. The problem-solving procedures used in these operations, as well as
those used in addition and subtraction , showed considerable surface variation, but
they could be described by one basic method, termed repealed groupings.
Examples of this method for multiplication and division are presented in Table 2.
Although the particular groups used by the same child across problems, as well as
the groups used by different children for the same problem, varied, the method that
generates the solutions seems to be the same.
The method of repeated groupings in multiplication and division appears to
indicate that children understand, even if only implicitly, the property of
distributivity. In other words, they understand that either factor in a multiplication
can be broken into parts to facilitate computation. Each part can then be
multiplied/divided sequentially by the other factor and the results added up in order
to obtain the final answer. It is worth pointing out that distributivity is also the
property implicitly used in written multiplication and division algorithms. Thus,
the results for multiplication and division replicate those observed for addition and
subtraction. Children seem to use implicitly a reasoning principle to generate
answers when relying on one symbolic system, oral numbers, but face difficulties
when attempting to use a different symbolic system, despite the fact that the same
reasoning principle constitutes the basis for both practices of arithmetic .
Examples of errors in written and oral multiplication are presented in Table 3 .
They illustrate, as was the case for addition and subtraction , that errors in the
written algorithms are related to bugs in the procedures, whereas errors in the oral
procedure do not reflect lack of understanding of the problem or the computation
procedure.
Studies that are based on across-group comparisons have replicated the findings
of the within-subject studies already described. Across-group comparisons involve
subjects who have been exposed to only one predominant cultural practice, either
oral or written arithmetic. These comparisons have the advantage that there can be
no interference across the cultural practices in a subject's problem-solving efforts,
although they lack the control for general ability present in within-subject
designs. Schiiemann (1984) compared the performance of students, who had most
of their experience in written arithmetic, with that of carpenters, who had more
practice with oral arithmetic, and Grando (1988) compared corresponding groups
of students and farmers. Both studies revealed that the users of written arithmetic,
namely students, made more errors than the users of oral arithmetic , that is,
carpenters and farmers with little or no schooling. Furthermore, the students'
errors were significantly more often unreasonable answers for the problem they
were solving . For example, Grando presented her subjects with problems in which
7 m of wire had to be cut up in 1.S-m pieces to build a gate. Subjects were asked
how many pieces would be obtained through this division. The farmers' answers
varied between 7 and 4 pieces, whereas the students' answers varied between 0.4
and 413 pieces of wire (for further discussion , see Nunes, Schliemann, &
Carraher, 1993).
301
Table 12.3. Some examples of errors in oral and written multiplication.

1.
E:
Carlos :
E:
Carlos :
2.
Two wrong solutions but one sensible answer.

Carlos (solving the problem of the total price of 15 toy cars at 50
cruzeiros each) initially tried it in writing and wrote the following
solution, which he then explained:
50
"Five, zero, zero. Five times
x 15
five, twenty five. Times one,
250
twenty five."
Do you think that this is how much it will cost?
No. It'd cost more.
How much would it cost?
(seems to be counting on fingers but only starts to count aloud as he
says "two hundred" and touches the forth finger): 200, 250, 300, 350,
400, 450, 500, 550, 600, 650, 700." (His answer is 700; he seems to
have made a mistake as he established the correspondence between the
fingers and the 50s .)
A wrong written computation and a correct oral solution.
Fabio (solving a computation exercise, 15 x 50, tried it in writing and said

while solving it):
50
"Five, times zero, five (pointing at the
x15
digits in the units place). Now, one
55
times five, five (pointing at the
digits in the tens place). It's 55."
Later he solved the imaginary shop problem, the price of 15 toy cars at 50
cruzeiros each) : "Ten will be 500. Five is 250. Seven hundred and fifty."
Examples from Carraher, Carraher, and Schliemann, 1987; examples not
Finally, I consider an experimental study that examined the use of written versus
oral symbols in the solution of problems with negative numbers (Nunes, 1993).
The introduction of written directed numbers in Brazil involves a break with
previous written arithmetic practices taught in school. In the first four years of
school, the signs + and - (plus and minus) indicate the operation to be carried out,
and two operations with different signs cannot be carried out simultaneously. In
the representation of directed numbers from the 5th grade on, more than one sign
can be written in the same expression, and signs no longer represent the operation
to be carried out but the direction of the number. For example, -20 and -30 are
added, not subtracted, despite the fact that they are preceded by the minus sign.
These ambiguities in the use of written numerical representations seem to affect
students' ability to solve problems with directed numbers in the written mode.
302
Terezinba Nunes
However, the logic of the situation that suggests that two negatives should be
added and yield a larger negative value is understood by the subjects . When
students were randomly assigned to either a written or an oral solution group,
significant differences were observed between the groups. The students in the oral
solution group were able to use their understanding of the situation and performed
significantly better than the students in the written solution group . The ease of
solution in the oral mode contrasted with the difficulties in the written system in
a particularly clear way when subjects assigned to the written condition of testing
reexamined their answers through oral calculation (see Figure 4 for an example) .
Problem 5:
Seu Severino (the farmer's name) started the season with 10

cruzados. He earned 10 on the beans (Subject writes down +20
at this point) and had a loss of 40 on the rice (Subject writes 40). What was his situation at the end of the season?
Antonio :
Here I can't subtract. Because the one on top is smaller than

forty. I can't add them up because they are different.
Problem 9:
This time the farmer started the season with a profit of 20

(Subject writes down 20). He had a loss of 40 (writes 40) on
his corn and a profit of 10 on his beans (writes 10). What was
his situation at the end of the year?
Antonio :
Experimenter:
Antonio :
It's a loss of ten (without recourse to the written numbers) .

Why is it a loss of ten? Can you explain it to me?
Because he has twenty of profit and ten, that's thirty (writes
30 next to the first row of numbers). And he lost forty, that i s
ten of loss (writing the other values in).
Note that, in Problem 9, the numbers were written down as the experimenter
had requested, but Antonio's solution was not worked out from the written
representation. In Problem 9, unlike Problem 5, the student did not try to use
the school form of representing negative numbers and succeeded.
Fig. 12.4. A written and an oral solution to similar directed numbers problems .
303
The attempts to follow the rules for manipulating written symbols while setting
meaning aside resulted in errors that would appear to indicate lack of
understanding. However, this apparent lack of understanding contrasts strongly
with the reasoning displayed by the same subjects when the problem was solved
in the oral mode.
Taken together, the results of these studies indicate that the subjects' knowledge
of reasoning principles and understanding of the situation is not sufficient to
account for their performance during problem solving. The systems of signs used
to solve problems have their own characteristics, which also affect the processes
involved in solving problems. The characteristics of manipulations of the written
system distance it from meaning, thereby making it difficult for subjects to
monitor results through meaning. Thus, subjects are more likely to make errors,
and the errors are of larger magnitudes. In contrast, oral arithmetic allows subjects
to monitor the magnitude of intermediary results during calculation, resulting in
higher rates of correct responses and errors of smaller magnitude.
It could be argued that all the subjects participating in the previous studies were
novices in their use of written arithmetic. Expert performance should show greater
flexibility and fewer errors. Intuitively, it seems that expert performance.is not
prone to error but shows little flexibility. After many years of schooling, we
make fewer errors in arithmetic but still use the same procedures when relying on
written symbols. This is only an intuitive suggestion with respect to written
arithmetic. However, evidence from studies with experts in the use of the abacus
indicates that flexibility does not come with large amounts of practice in the use
of abacus arithmetic. Hatano (in press) put forth this argument in a recent review
of studies on the performance of novices and experts in abacus arithmetic. The
influence of the abacus as a mediating system of signs on expert performance is
illustrated both by remarkable feats and by amazing rigidity in problem solving.
For example, experts in abacus arithmetic can accomplish results in number recall
and manipulation that are clearly beyond what novices can do. They can memorize
numbers composed of 15 or 16 digits, can say the digits forwards and backwards
with the same ease, can carry out sequences of operations more quickly than they
can write out the answers, and can calculate while answering simple questions.
These characteristics of their skill suggest that experts' numerical activities are
mediated by a visually represented mental abacus that is used following specific
procedural rules. Consistent with this hypothesis, visual stimuli were found to
interfere with experts' performance more than verbal stimuli did. However,
monitoring of the operation is suppressed in the development of this skill, just as
it is in written arithmetic. Thus, experts fail to notice that they have just added a
list of numbers when minimal changes are made on the list (such as having the
first number in one list be the last in a subsequent list) and add it all up again
when presented with the second list. They also do not rely on simplified related
number calculations when they solve large number problems (such as solving 99
x 38 by solving 100 x 38 and then subtracting 38).
304
Terezinha Nunes
The studies just described indicate how different sorts of culturally developed
signs affect calculation processes in particular ways. These studies, however , do
not show whether the use of different symbols to represent the same situation can
result in the development of different reasoning structures or conceptions of the
situation . To examine this second question, I turn to studies on the measurement
of area.
Changing Tools and Goals in Measurement Problems

Measurement tools (such as rulers) are objects that can be introduced in children's
environments. They are also symbolic tools, because they allow the user to
obtain a number that will represent the measured aspect of the object. Measuring
involves reasoning principles such as understanding units and conserving them
across displacements (Piaget, Inhelder, & Szeminska, 1960), relying on transitive
inferences for comparisons through an intermediary term (Piaget, Inhelder, &
Szeminska, 1960), and using iteration and subdivision of the unit to fully cover
the measured object (Nunes, Light, & Mason, 1993).
Performance on estimation tasks indicates that subjects use their knowledge of
measurement principles as well as their experiences with measurement units in
the process of estimation. Gay and Cole (1967), for example, observed large
within-individual differences in estimation of length tasks as a function of the unit
used . They observed that adults from the Kpelle group of Liberia were rather good
at estimating length when they were asked to produce an answer in arm and handspans: Their errors were, on average, not larger than 30% of the estimated distance
for distances between 72 and 180 inches . In contrast, when they estimated the
same distances in foot-length, a possible but uncommon measure among them,
their estimations were consistently more than 40% off, with average errors for
certain distances being as high as an 80% overestimation. Considering the withinsubject design used, it is unlikely that the subjects had different amounts of
knowledge of the reasoning principles involved in measurement operations. More
likely , these results indicate that the way in which they thought about length was
affected by the system of signs they were using in the estimation task. But, as
pointed out with respect to the studies in the last section, issues of familiarity
cannot be avoided in studies of subjects' participation in cultural practices. In
order to circumvent these problems, experimental designs are needed.
We conducted a series of three studies about children measuring area with
different tools, which gives further insight into how symbols in the situation
structure solutions to problems (Nunes, Light, & Mason, 1993; Nunes, Light,
Mason, & Allerton, 1994). In all three studies, the tools available for measuring
area were rulers and l-cm" bricks. In the U.K., where the studies were conducted,
rulers are used in the measurement of area in combination with formulae. All
305
subjects in these studies had been taught in school the formula for the area of the
rectangle . Bricks and grid-lined paper are often used in introducing area work,
when children cover figures with bricks and count them in order to obtain a
measure for the area. The goal of instruction is to move the children from the use
of bricks to the use of rulers plus formulae in the measurement of area. However,
children typically show great difficulty in taking this step, even after carefully
designed instruction (see Dickson, 1989). Our goal in this series of studies was to
analyze the reasons for this difficulty in the change of measurement tools.
A parallel can be found between the change of tools in oral and written
arithmetic and the change of tools in measuring area with bricks and with a ruler.
The measurement with the support of a ruler seems very simple from the
viewpoint of the algorithm used. The obtained measures simply need to be
combined through a formula. However, the steps in the procedure are distanced
from meaning. Rulers provide linear measures, and their connection to the
measurement of area is by no means obvious. Bricks, on the other hand, can be
viewed as units of area. Thus, their relationship to area problems is much clearer.
The question for a subject measuring an area with bricks is "How many of these
bricks are needed to cover the figure?" No parallel question can be posed when the
child starts from a linear measure.
In the studies described next, we used both a choice paradigm (children had more
than one tool available, and we verified which one was chosen and how the
structure of solution was affected by the choice) and a preassignment paradigm
(children were randomly assigned to experimental conditions in which a different
tool was available) .
The task used in all three studies was the same. The children, always working
in pairs, were told that two friends had done some painting work for a neighbor
and each one painted a wall. When they received their money, the friends wanted
to figure out whether they had done the same amount of work but found it
difficult to decide because one wall was longer than the other but the second one
was taller. The children were asked to use the pictures of the walls to figure out
whether the two friends had done the same amount of work. A series of four
problems was used involving two simpler comparisons and two more difficult
comparisons. Feedback was provided between trials by covering one area with
previously cut pieces of colored paper, which were then moved and assembled over
the other figure. Figure 5 presents schematically the series of problems and
illustrates the feedback procedure.
306
Terezinha Nunes
_rn
Item I in the task involved comparing two rectangles (A) with the same area
but different dimensions. The feedback procedure involved covering each in
sequence with the same bits of pre-cut colored paper as shown in (B) .
/
Each figure in C was used for comparison with a rectangle in the more difficult
items .
Fig. 12.5. Two rectangular areas used in the easier comparison of area
problems and a schema of the feedback procedure (top); two other figures used in
the difficult problems, where a parallelogram and a U-shaped figure were each
compared with a rectangle (bottom) .
A choice paradigm was used in the first study (for further detail, see Nunes, Light,
& Mason, 1993; the description presented here is simplified). Thus, children had
access both to rulers and to bricks during problem solving and could resort to
whichever they preferred. Bricks were numerically insufftcient to cover any of the
areas to be measured in order to avoid simple counting procedures to solve the
problem. We anticipated that the analysis of problem-solving procedures would
307
indicate whether the same underlying conception of the problem was reflected in
the children's solutions, regardless of the tool used or whether different conceptions would emerge .
An initial quantitative analysis showed that children who chose the ruler
performed significantly poorer than those who chose the bricks. Approximately
half of the children who initially chose the ruler as the measuring tool, after
measuring the sides and calculating the perimeter, realized that they had not
obtained a good evaluation of the work involved in painting the walls . They then
moved to the use of bricks. Those children who started with rulers and did not
change to bricks calculated either the perimeter or half of the perimeter instead of
calculating the area. This type of error, which is well documented in the literature
(see, for example, Dickson, 1989; Douady & Perrin-Glorian, 1989), is of
ambiguous interpretation. It has been suggested that this error reflects the
children's failure to differentiate the concept of perimeter from that of area
(Douady & Perrin-Glorian, 1989), but it can also be argued that the multiplication
of centimeters by centimeters makes no sense to children, whereas addition is
meaningful, and thus children choose to do what makes sense to them.
Brick users, in contrast, were much more likely to solve the problem correctly.
They tended to discover that the number of bricks in a row in a rectangle is fixed
and then to come up with the formula "number of bricks in a row times number
of rows." This formula corresponds to a model of multiplication that Vergnaud
(1983) has termed isomorphism of measures and can be described as involving a
one-to-many correspondence scheme (that is, if one row corresponds to 8 bricks, x
rows correspond to x times 8 bricks). In contrast, the ruler-plus formula approach
to area corresponds to Vergnaud's product of measures model, where one measure
multiplied by another yields a third measure (i.e., centimeters multiplied by
centimeters yield square centimeters, a new measure).
These results indicate that the tool provided to children for solving area
problems plays a structuring role in the conception of area that they develop in
the course of problem solving. Bricks, which are area units, are more clearly
connected to the problem than linear measures obtained in centimeters. Thus , the
use of bricks as tools for measuring area allows children to use the one-to-many
correspondence scheme that they often use in multiplication problems, whereas
the use of rulers requires a different conception of multiplication.
In a second study, a pre-assignment paradigm was used to avoid possible
confounding of the results by self-selection. The same task was used, but this
time the pairs of children were assigned to either a ruler or a brick availability
condition (Nunes, Light, Mason, & Allerton, 1994). In an attempt to discourage
the most frequent errors observed in the previous study, extra feedback was
provided during problem solving to children who calculated either the perimeter or
half of the perimeter. The children were told that they had correctly calculated
"how much it was all around the walls," and then they were asked whether they
knew how much work it was to paint the whole wall.
308
Terezinha Nunes
Children in the bricks condition were, as in the previous study , more successful
in solving the problems than those in the ruler condition. In spite of our attempts
to discourage addition of the measures obtained for the sides, children in the ruler
condition still preferred this solution. Children in the bricks condition, as in the
previous study, tended to discover the regularity of number of bricks in a row in
the rectangle and to come up with the formula "number of bricks in a row times
number of rows."
All children were seen again individually about one month later and were asked
to calculate the area of some figures . In this second session, both bricks and rulers
were available. Approximately two thirds of the children chose to work with the
same tool with which they had worked in the previous session. The children who
had worked with bricks in the first session still showed significantly better
performance than those who had worked with rulers .
In the third study, we first offered the children an equal amount of instruction in
the calculation of the area of a rectangle in either a brick or a ruler condition.
Immediately after instruction, they solved another problem that involved the
comparison of rectangular areas. In this immediate posttest, no difference was
observed between the two groups. However, in the next trials, when the children
were asked to solve problems that involved comparing a rectangular area with the
area of new figures (a parallelogram and a U-shaped figure that could be
decomposed into rectangles), the children in the bricks condition were much better
at adapting the procedure that they had learned. The formula that they had come up
with, number of bricks in a row times number of rows, allowed them to maintain
the meaning of area in focus, and they were, consequently, in a better position to
adapt their knowledge to new figures than those who had worked with rulers. Area
units symbolize area in a much clearer fashion than linear units do.
This series of studies indicates that the system of signs that the children were
offered in the learning phase influenced the reasoning principles that they used in
organizing the problem solution. The structuring of the children's action was not
independent from the tool they had at their disposal in the problem-solving
situation. Consequently, the very reasoning principles developed in the situation
were structured by the system of representation used.
Conclusion
The studies reviewed here have both theoretical and practical implications. There
is clearly a need to reexamine the concept of ability, which was shown to be
determined not solely from within (neither by talent nor by developmental level)
but also to be significantly affected by the systems of symbols used by the
subject. Different systems of symbols pose different types of difficulties for
children. We have seen how oral and written arithmetic differ and how they
309
influence children's reasoning processes as they are used as tools for thought. In
this context, it is also necessary to consider what kind of power the different tools
for thought offer to users . Within the range of numbers used in the studies
described here, oral arithmetic was powerful enough, but memory errors did crop
up . In the realm of larger numbers, oral arithmetic would probably have been a
poor choice of tool. Similarly, calculating area with the formula "number of
bricks in a row times number of rows" is a good initial approach to area
problems. The use of area units allows the subjects to keep the meaning in focus
and to perform sensible calculations. However, area units are not practical in
everyday life, where it is much simpler to measure figures linearly and to calculate
the area using a product of measures approach. Thus representational systems pose
different types of difficulties for their learners and also afford different power as
tools.
The educational implications of this analysis are also significant. Children
invest a substantial amount of effort in the acquisition of culturally developed
systems of signs such as oral and written numeration systems, measurement
systems, mathematical notations, and today, forms of inputting commands into
computers. These systems are used both for communication with others and for
personal purposes during problem solving . If symbolic systems are an integral
part of the intellectual process, schools , as privileged institutions for the
transmission of many cultural practices, need to reconsider their approach to
teaching . Greater attention must be given to how systems of signs structure
thinking and what systems are to be chosen as basic for everyone .
310
TerezinhaNunes
References
Brown, J. S., & Burton , R. R. (1978). Diagnostic models for procedural bugs in basic
mathematical skills . Cognit ive Science , 2, 155-192.
Carraher, T. N. (1985). Exploracoes sobre 0 desenvo1vimento da competencia
ortografica em Portugue s. [Exploring the development of spelling in Portuguese]
Psicologia: Teoria e Pesquisa, I , 269-285.
Carraher, T. N. (1986). From drawings to building s: Working with mathematical
scales. International Journal of Behavioral Development, 9, 527-544 .
Carraher, T. N., Carraher, D. W., & Schliemann, A D. (1985). Mathematics in the
streets and in school s. British Journal of Developmental Psychology, 3, 21-29 .
Carraher, T. N., Carraher, D. W., & Schliemann, A. D. (1987). Written and oral
mathematics . Journal for Research in Mathematics Education, 18, 83-97.
Dickson, L. (1989) . Area of a rectangle . In D. C. Johnson (Ed.), Children 's
mathematical frameworks 8-13: A study of classroom teaching (pp. 89-125).
Windsor, England : NFER-Nel son Publishing Company.
Douady, R., & Perrin-Glorian, M. J. (1989 ). Un processus d' apprenti ssage du concept
d'a ire de surface plaine [A teaching experience on the concept of the area of a plane
surface]. Educational Studies in Mathematics, 20, 387-424.
Gardner, H. (1993 ). Multiple intelligences: The theory in practice. New York: B asic
Bo ok s.
Gay, J., & Cole, M. (1967). The new mathematics and an old culture: A study of
learning among the Kpelle of Liberia. New York: Holt, Rinehart & Winston.
Grando, N. (1988): A matematica no meio rural: Pequenos agricultores, estudantes e
professores [Mathematics in the rural area: Farmers, students, and teachers] . M.A.
Thesis, Mestrado em Psicologia, Universidade Federal de Pernambuco.
Hatano, G. (in press). Learnin g arithmetic with an abacus. In T. Nunes & P. E. Br yant
(Eds.), Teaching and learning mathematics: International perspectives. Falmer,
England: Lawrence Erlbaum .
Hughe s, M. (1983). What is difficult about learning arithmetic ? In M. Donaldson, R .
Grieve, & C. Pratt (Eds.), Early childhood development and education (pp. 204-221).
Oxford, England: Blackwell.
Lave, J. (1988). Cognition in practice. Mind, mathematics and culture in everyday life.
Cambri dge, England: Cambridg e Univer sity Press .
Lawler, R. W. (1990) . Constructing knowledge from interactions. In L. P. Steffe & T.
Wood (Eds .), Transforming children's mathemati cs education. International
perspe ctives (pp. 47-51). Hillsdale, NJ: Lawrence Erlbaum .
Light, P. & Perret-Clermont, A-N. (1989) Social context effects in learning and
testin g. In A Gellatly, D. Rogers, and J. Slobod a (Eds.), Cognition and social
worlds (pp 99-112) . OXford, England: Clarendon Press .
Luria, A (1979). Curso de Psicologia Geral [Course of General Psychol ogy]. Rio de
Janeiro: Civilizacao Brasileira.
Nunes, T. (1993). Learning mathematics: Perspectives from everyday life. In R. B .
Davis & C. A. Maher (Eds.), Schools, mathematics, and the world of reality (pp .
61-78) . Needham Heights, MA: Allyn and Bacon.
311
Nunes, T. (1995). Cultural practices and the conception of individual differences .

Theoretical and empirical contradictions. New Directions for Child Development,
67, 91-103 .
Nunes, T. (1994). Street intelligence. In R. Sternberg (Ed.), Encyclopedia of
Intelligence. New York : MacMillan.
Nunes , T., Light, P., & Mason , J. (1993). Tools for thought: The measurement of
length and area. Learning and Instruction, 3, 39 -54.
Nunes, T., Light, P., Mason, J., & Allerton, M. (1994). Children's understanding of
the concept of area. Research report preparedfor the Economic and Social Research
Council (ESRC), Institute of Education, University of London.

Nunes, T., Schliemann, A. D., & Carraher, D. W. (1993) . Street mathematics and
school mathematics. New York: Cambridge University Press.
Piaget, J. (1962). Play, dreams and imitation in childhood. New York: Norton .
Piaget, J., Inhe1der, B., & Szeminska, A. (1960) . The child's conception of geometry.
London: Routledge and Kegan Paul .
Plunkett, S. (1979) . Decomposition and all that rot. Mathematics in Schools, 8,2-7.
Reed , H . 1., & Lave, J. (1981) . Arithmetic as a tool for investigating relations between
culture and cognition. In R.W. Casson (Ed.), Language, culture, and cognition:
Anthropological perspectives (pp. 437-455). New York : MacMillan.
Schliemann, A. D. (1984) . Mathematics among carpenters and apprentices. In P.
Damerow , M. W. Dunckley, B. F. Nebres, & B. Werry (Eds.), Mathematics for all
(pp. 92-95). Paris : UNESCO.
Thurstone, L. L. (1938). Primary mental abilities . Psychometric Monographs, 1
(whole no .).
Vergnaud, G. (1983). Multiplicative structures. In R. Lesh and M. Landau (Eds.)
Acquisition of mathematics concepts and processes (pp 128-175). London , England:
Academic Press .
Vygotsky, L.S. (1962) . Thought and language. Cambridge, MA: MIT Press.
Chapter 13
Understanding Symbols with Intermediate

Abstractions: An Analysis of the Collaborative
Construction of Mathematical Meaning
Baruch B. Schwarz
Hebrew University, Jerusalem , Israel
In this chapter, I investigate how children understand concepts and the meaning of
symbolic representations when working with conceptual models. The particular
focus of this study is the nature of children's conversation when collaborati ng to
solve tasks embodied in what we call an intermediate abstraction. The analysis is
based on theoretical considerations developed by Greeno and his colleagues
(Greeno, 1989; Greeno, Engle, Kerr, & Moore, 1993) and on a line of research
carried out by Resnick and Schwarz (Schwarz, Kohn, & Resnick, 1993; Schwarz
& Nathan, 1993; Schwarz, Nathan , & Resnick, 1996; Schwarz, in press) with
several computerized systems . My approach focuses theoretical attention on the
nature of conversational turn-taking between children working with an
intermediate abstraction of mathematical operators. This study shows that
conversational processes of acceptance stemmed from collaborative learning , and
that those processes enabled the construction of shared meaning resulting in the
development of the concept of a mathematical operator. Three such processes were
found. The first one is a process of creation of a shared meaning through reference
to objects of the system . This process is initiated by one participant describing a
property of the intermediate abstraction. The second participant then accepts the
meaning uttered by the first participant and refines it. The second one is a process
of convergence of a different nature that is initiated by an impasse within one
particular representation and is characterized by a change of representation. The
third turn-taki ng occurrence is a process of confirmation through a different
representatio n. At the end of this chapter, I show how individuals who interacted
Understanding Symbols with Intermediate Abstractions
313
in collaborative work on a particular intermediate abstraction, the OPERA

system, were able to solve problems in novel situations, and how providing them
with the objects of the system could help them with those problems with which
they had difficulties.
Intermediate Abstractions and Conceptual Models for the

Learning of Concepts
Several researchers have created computerized environments with objects modeling
concepts . For example, Smith, Snir, & Grosslight (1992) used several interactive
systems that simulated objects floating (or sinking) in a liquid to help students
differentiate between the concepts of weight and density. Wiser (1988) used energy
units to model the amount of heat of a body and used the crowdedness of these
units to model the temperature of the body. These environments, called conceptual
models, were thought to be superior to concrete models in the sense that they
were simple and that the designer could tailor representations and tasks to focus on
specific aspects of the concept. White (1993) refined this approach by creating
models that, if assimilated, "enable children to build on the knowledge and
reasoning forms they already possess, and to conceptualize domain phenomena at
an intermediate level of abstraction" (p. 10). These intermediate abstractions nero
to be simple and to be introduced in a coherent series of progressively more
sophisticated models. Moreover, they need to be neutral, that is, not contextbound. For example, in the ThinkerTools system , White uses a data-cross
representation of motion that can match any situation because all its objects are
generic. The Planner system (Schwarz, Kohn, & Resnick, 1993; Schwarz,
Nathan, & Resnick, 1996) is an intermediate abstraction that exemplifies positive
and negative numbers, and operations on these, in a coherent system. Several
empirical studies were done to assess the effect of each of these systems when
students work in pairs (e.g., Schwarz, in press ; White, 1993). Although very
positive results were found when students worked in collaboration, the question of
what causes collaborative learning to be beneficial when interacting with
conceptual models is open.
A first attempt to define conversational patterns while working in collaboration
on an intermediate abstraction was done by Roschelle (1992). Students worked in
pairs on the Envisioning Machine , a system for acquiring the concepts of velocity
and acceleration. The Envisioning Machine displays in parallel a ball having a
given trajectory and a particle whose position, velocity, and acceleration need to
be manipulated in order to correspond to the trajectory of the ball. Roschelle
studied the conceptual change of a pair of students operating together on the
system and talking to each other. This process was characterized by (a) the
construction of a deep featured situation at an intermediate level of abstraction
314
Baruch B. Schwarz
from the literal features of the world, in relation to (b) the interplay of physical
metaphors, through the constructive use of (c) iterative cycles of conversational
turn-taking, constrained by (d) the application of higher standards of evidence for
convergence. Therefore, according to Roschelle (1992), intermediate abstractions
trigger conceptual change because they create favorable ground for developing
metaphors, and because students collaborating together can gradually construct
shared meaning (what Roschelle calls convergence).
The approach adopted by Greeno et al, (1993) is slightly different. It recognizes
both interactional and cognitive processes as contributing to conceptual change in
collaborative work. Greeno and his colleagues focused on the "winch system" to
acquire symbolic representations. In that system, the labeled size of each spool is
equal to its circumference in inches, which equals the distance that a block moves
each time the handle is turned with the spool. This means that saying or writing a
numeral can be interpreted as referring to the size of a spool, a distance per turn,
or both . Further, if a phrase is used that specifies one of these (e.g., the four
spool), its referent can be extended to include the distance per turn of a block with
that spool if the speaker and the listener are attuned to the constraint that these
two quantities are equal. Greeno et al. (1993) generalize that learning symbols is
an activity that involves attunement to constraints and affordances about (a) the
domain that is referred to, (b) the domain of notational expressions, and (c) the
meaning relations between notations and their referents. In the referent domain,
attunement to constraints makes it possible to refer to co-constrained properties of
a situation. Constraints of the notations can be considered the syntax of the
representational system. For example, in equations, there are constraints on the
sequence of numerals, letters, and operations signs that the student learns , and
there is a simple grammar that determines a phrase structure for any algebraic
expression. In graphs, there are constraints, such as the continuity of the line that
is considered as an object, and affordances, such as considering any segment of a
line as an object.
Greeno's framework about understanding as attunement to constraints and
affordances is adopted in this study. I examine the kinds of relations between
symbolic and nonsymbolic representations developed in collaborative work with
an intermediate abstraction. The two components considered are to which degree:
(a) different symbols, concepts, or interpretations were differentiated from each
other and (b) the relations between them were specified systematically. I also
adopt the theoretical ideas given by Clark and Wilkes-Gibbs (1986) and Clark and
Schaeffer (1989) about conversational processes. In particular, referring is
considered a collaborative process, with the speakers trying to establish, by the
initiation of each new contribution, the mutual belief that the listeners have
understood what the speaker meant in the last utterance (the principle of mutual
responsibility).
315
The Problematic Case of Algebraic Symbols

Operators are fundamental objects in algebra, especially the basic operators
(addition, subtraction, multiplication, or division) and their compositions.
Examples of such operators are the operator +3, which transforms +2 into +5 or
-3 into 0, and the operator -2 * 4, which transforms +2 into 0, +3 into +4, and
+10 into +32. Traditionally such operators are represented symbolically by means
of a letter expressing a variable. For example, the two preceeding operators could
be represented as x + 3 and 4*(x - 2). Manipulations on these operators are
governed by general principles such as distributivity, associativity, and
commutativity. Several researchers have studied students' behavior on the
manipulation of algebraic expressions. These studies reveal that many students'
errors are governed by buggy rules. Matz (1980) has hypothesized that those rules
are generated as over-generalizations of correct rules . Larkin (1989) has
hypothesized that they are generated because students apprehend algebraic
expressions as string data representations . For example, the expression 4*(x - 2)
is manipulated as if it were equal to 4*x - 2. This indicates that the student parses
symbols in algebraic expressions according to their order of appearance. Several
educators constructed visual displays that show the hierarchical structure of
algebraic expressions and their manipulations (e.g., Larkin, 1989). However,
most of the displays are essentially symbolic, and there is no reason interactions
with them could be applied in novel situations. As Greeno states, "It seems likely
that the approaches to instruction in algebra advocated by Anderson and Larkin
would not be helpful in overcoming the insulation of symbolic expressions from
their intended meaning" (1989, p. 295). Similarly, having in mind that most of
the displays show symbolic objects disguised as semantically based entities,
Kirshner (1989) said: "The human mind is uniquely fashioned to learn syntax as
syntax and that current syntactic instruction fails, not because it is syntactic, but
because research has not begun to fathom the depth of complexity required of
syntactic performance in algebra. . . . Computer environments designed to
promote meaningful conceptual development result first and foremost in learning
about button pushing on computer keyboards" (p. 197).
An example illustrates the complexity of algebraic reasoning . This example
shows the relations existing between algebraic word problems and abstract
formulae modeling them. The word problem, The Merry Meadows Farm, follows:
The Merry Meadows Farm sells 12 crates of peaches to the local grocery store. To
deliver the peaches , the store has to rent a truck for $36 . Altogether, it costs the
store $210 to get the peaches and bring them to the store. How much does the
Merry Meadows Farm ask for each crate?
316
Baruch B. Schwarz
The equation is 12*x + 36 = 210. For the equation, the symbolic expression is
the formula itself. For the word problem, it is the written text (i.e ., the words,
nouns, verbs, prepositions, and adjectives included in the formulation). The text
here refers to its objects (e.g., the Merry Meadows Farm, the peaches, the truck)
and to all the events (e.g., renting a truck, bringing the peaches to the store)
described in it. The equation 12*x + 36 = 210 can refer to the quantities in the
word problem. It can also relate to abstract entities: for example, the function f(x)
= 12*x + 36. In general, symbolic expressions include written or spoken
instructions, and they refer to quantities in a concrete situation or to abstract
entities. The two symbolic systems can undergo transformations. For example,
the sentence, "The Merry Meadows Farm sells 12 crates of peaches to the local
grocery store," can be transformed into "The local grocery store has been sold 12
crates of peaches from the Merry Meadows Farm." Similarly, the equation can be
transformed into 12*x = 120 - 36 or into 12*(x + 3) = 120. Therefore, when
solving algebraic word problems, algebraic symbols refer both to mathematical
entities and to objects and events of the word problem. For example, 12*x + 36
can refer either to the function f(x) = 12*x + 36 or to the price the store pays for
having the 12 crates. In general, just a part of the operations that we perform on
symbolic expressions preserves the references of their terms in the domain in
which they appear. For example, the transformation of the equation 12*x + 36 =
120 into 12*x = 120 - 36 is consistent with the transformation of the situation
into "paying for the 12 crates without the shipment costs $120 - $36." The
transformation of the equation into 12*(x + 3) = 120 is possible because 12*(x +
3) and 12*x + 36 refer to the same function. In contrast, this transformation has
no easy mapping with a transformation of the word problem situation (something
such as "Because the overall shipment costs $36, the shipment for each crate costs
$3"). Therefore, it is not surprising that students have difficulties reasoning
algebraically (i.e., to be able to use algebraic entities) to solve word problems.
The constraints of the two symbolic systems are not considered as a coherent
whole. Subjects able to conduct formal computations correctly were often totally
unable to apply their skills in real world situations; rather, they chose ad hoc
methods triggered by the situations. Such a phenomenon stresses what Greeno
(1989) calls a "perverse semantic interpretation that maps the symbolic structures
back to the notations used to express them, rather than to a domain of real objects
and situations and abstract mathematical entities" (p. 294). This interpretation
implies that problems of translation between the two symbolic representational
systems are very difficult. For example, Clement (1982) showed that students in
various grades generally failed to solve the following problem, referred to as The
Students and the Professors:
Write an equation using the variables Sand P to represent the following statement:
There are six times as many students as professors at this university. Use S for the
number of students and P for the number of professors .
317
Clement showed that students often write the relation 6S = P (instead of the
correct relation S = 6P), because letters are not perceived as representing the
number of students, and the equal sign is not used to express an equivalence but
rather a comparison or an association. In this example, the correspondence is
syntactic , not semantic. A similar phenomenon was found by Schoenfeld (1987)
in geometry where students' knowledge was linked to "drawings" and not to
spatial relations and constraints. In contrast, it was found that intermediate
abstractions allowed students to reason about mathematical and scientific
concepts. For example, students collaborating to solve problems on the Planner
system could model and solve difficult arithmetic word problems (Schwarz, Kohn,
& Resnick, 1993; Schwarz, Nathan, & Resnick, 1996) and could reason about
negatives with different referents (e.g., temperatures, chronology, elevators). This
is because children could develop links between the actions they operated on the
objects of the system and the symbols, links that were based on affordances and
constraints of each of the representations. The symbolic-descriptive expressions
were branched to demonstrative representations, that is, as objects with properties
and relations that correspond to properties and relations of objects in the situation
or domain that the expression represents. The study described in this chapter
defines the learning situation created by the use of the system OPERA, an
intermediate abstraction for the acquisition of the concept of mathematical
operator. In particular , it traces the contribution of collaborative learning and the
processes that characterize it.
The OPERA System

OPERA is an intermediate abstraction of mathematical operators. It consists of
several external representations, each representation featuring a different aspect of
operators. Because operators operate on numbers, not only operators but also
numbers are embodied in each of the representations. Activities are defined around
OPERA, these activities being characterized by a formulation (the goal), the
representations available, and the kind of interaction chosen: pairs of children or
child-experimenter working on the computer. In this sense, a situation is created
in which participants collaborate and interact with an intermediate abstraction to
find the solution of the activity. The role of the intermediate abstraction is
important not only because collaboration and interaction are afforded around
demonstrative representations but also because the representations provide a
referential semantics for the concept of mathematical operator.
The first external representation, the machine representation, is an input-output
device that embodies operators as processes that change quantities . Simple
machines represent additions, subtractions, multiplications, or divisions (see
Figure 1).
Baruch B. Schwarz
318
(a)
(b)
(e)
(d)
Fig. 13.1. Examples of machines (a, b, c) and of a team (d).
There are also what we have called teams of machines, in which operators consist
of more than one element. Machines operate on trains, the trains having lengths
that are indicated by a "flag." A machine operates on trains by putting them into
the entry portal of the machine and "clicking" on it. Figure 2 shows how the team
[+2] [*3] runs on a +2 train. The "out-train" is a +12 train. Clicking on a train
when it is at the exit portal will run it "backwards," producing a +2 train at the
entry portal. This representation shows operators as numerical processors, that is,
the number/train is processed by operation(s)/machine(s). The trains and the
machines do not display literal features of real trains but are metaphors that
express an essential common function of trains and operators: the fact that there is
a "vehicle" that undergoes a series of transformations .
......~~~~
............... ~~~ 5t1
Fig. 13.2. Running a +2 train through a [+2][*3] team
The nonliteral metaphoric correspondence trains/machines :::> numbers/operators

is evident by another interesting function : It is possible to rotate machines in
order to exchange the input and output (thus one can "flip" machines) . The
machine obtained by this rotation is the inverse operator of the operator it
represented. For example, flipping a [+3] machine yields a [-3] machine, and
flipping [+21 [*3] yields a [/3] [-2] machine (see Figure 3). Therefore, the abstract
idea of inversion is afforded by a simple geometric transformation in accordance
with its meaning: to invert an operator is to find an operator whose inputs are the
outputs of the original operator and whose outputs are the original inputs . Of
course, the terms machines and trains here lose any literal meaning.
. . . .~ Q . .<J.. ~
........
319
......
Fig. 13.3. Flipping the [+3] machine and the [+2][*3] team
..
0.1
1
2
I
4
4
7
10
II
Fig. 13.4 . Reflecting a formula in other repre sentations
In addition to machines and trains, the three traditional representations are

available, each of them uncovering a different feature of operators. The symbolic
algebraic representation is intended to show the operator as an abstract
mathematical entity; tables display a mapping between two series of values;
graphs display variations. Moreover, each of these representations can be reflected
in the others. For example, writing a formula can be reflected in a graph or as a
team of machines (and vice versa), as shown in Figure 4. This potential for
reflecting work from one representation to the others was not always made
320
BaruchB. Schwarz
available, each of the activities with OPERA being defined by constraints

imposed on the user. The parallel between representations stresses that several
objects in the system refer to the same entity, although from different
perspectives. In the next section, I show that this property of OPERA is crucial
in collaborativelearning.
JI:=='-
2
~
Ji:=:==l._-JL
Sa
J.!:=::==.L
. __
....J:::=~_
5b
J.!::==!JL
5c
~.
.=::=JL
5d
Fig. 13.5. The stack representations for the operators [+2] (5a), [*3] (5b), [+2][*3]
[5c], and [*3][+2] (5d)
The four representations alluded to thus far are symbolic. Intermediate abstractions
always contain components from which intuitions can stem. The stack
representation plays this role. It is not symbolic, showing the changes that
quantities undergo with operators in a dynamic visual way. Figure Sa shows a
stack representation of the operator [+2] as a process. It begins (as always) with
the creation of a generic block; then, a block labeled +2 is chained on its right
321
side. The operator [*3J is shown in Figure 5b as the gradual construction of a 3high stack. Figure 5c shows the creation of the compound operator [+2J [*3].
After the creation of the stack for the [+2] operator, it is expanded as a whole to a
stack with three generic blocks and three blocks labeled +2. Switching the [*3]
and [+2] totally changes the representation, as shown also in Figure 5d.
These examples show that the stack representation makes explicit the structure
of algebraic expressions as the formation of structures. The critical aspects of
those structures are unambiguous (in contrast with the traditional algebraic
syntax). Multiplication is an expansion in the vertical direction; addition and
subtraction are translated as putting blocks on the right side (in the horizontal
direction). The importance of order for operations is made explicit by the very
different stacks that result. Of course, this representation provides the child with a
set of constraints (such as the horizontal and vertical expansions for addition and
multiplication). These constraints are often arbitrary, although easily explainable.
In conclusion, OPERA provides a system of representations that express
different aspects of operators. The tables convey the idea of mapping; the
machines express operators as numerical processors; the stacks are an embodiment
of algebraic structures. All the representations can be displayed and run in parallel.
They provide co-constrained properties in a symbolic and nonsymbolic way.
Activities with OPERA

An activity with OPERA is defined as a set of OPERA tools and actions, a
formulation containing a goal, and a configuration of interaction. In most of the
activities, this interaction occurred between two students working on the OPERA
system without external help. In the last stage of the experiment, the
configuration was different: An adult negotiated with individuals those OPERA
objects that could lead to the solution of difficult word problems. In sum,
OPERA activities defined a situation of learning, where interaction between
agents, as well as interaction between users and the intermediate abstraction,
afforded the development of the concept of mathematical operator. In this chapter,
the focus is restricted to activities that concern composition and inversion of
operators. Among the numerous ways to ask questions about composition and
inversion , the numerical processor aspect can be embodied by activities such as
In-train 7 (see Figure 6). Its formulation is "What is the in-train, if the out-train
is 25T' One way to solve this problem is to run a 25 train "backwards." Another
way is to construct a table with several values and to infer for which x-value, the
y-value is 25. It is also possible to draw a line representing the operator [*3][+7]
and to "read" the first coordinate of a point on the line whose second coordinate is
25.
Inverse 3 is also an activity that stresses the numerical processor aspect (see
Figure 7). The question here is How does the black box work?; the tools are trains
and machines; the actions are "running" and "flipping ." Running trains through
Baruch B. Schwarz
322
the team of machines leaves them invariant. For example, for a 2 in-train, the
out-train is also a 2 train. To construct a team matching the black box, the
student can first disconnect it from the team [*2] [-1] and then flip this team. But,
of course, many other strategies based on the various representations available in
OPERA can also be applied.
Inlnln1
What Is the In-train If the out -tra in Is 2S?
Fig. 13.6. A "numerical processor" activity with OPERA
Inu.n . ]
How does the black bo x w ork?
(6 ' 2) - 1)-7
.,
In
Oul
.'
3
Fig. 13.7. An activity around the inverse of the operator [*2][-1]
323
An Experiment with OPERA

Six fifth-grade children from an inner-city school participated in this study . The
students were given a pretest, administrated individually, in which they were asked
to solve algebra problems. The students then worked in three pairs solving
OPERA problems. The six students were then given a posttest, which was
identical to the pretest. Finally, an experimenter chose difficult problems from the
posttest that students had failed to solve and tried to help them by proposing
OPERA objects that could be used to model these difficult problems. More than
the comparison between the achievements in the pretest and the posttest, this last
stage was a measure of the effects of collaborative learning on individuals.
The Pretest and the Posttest

These paper-and-pencil tests consisted of formal problems and word problems. In
each problem, the student was asked to find either a numerical solution or a
"formula." The problems were of four types. The following is an example of a
formal numerical problem: Find the missing number in 2(3* _-5) = 32. An
example of a numerical word problem is The Merry Meadows Farm. An example
of a formal problem demanding the use of a formula is the request "Find how a
table works" when given some of its x- and y- values. For instance, students need
to discover that a table that matches the x-values 1, 5, 7, and 10 to the y-values 4,
17,23, and 32 is governed by the formula 3*_+2. An example of a word problem
demanding the use of a formula is The Students and the Professors problem
shown earlier.
Working in Pairs to Solve OPERA Activities

The six students worked in three pairs during six sessions. The first session aimed
at familiarizing the children with the objects and the functioning of the system.
During the five remaining sessions, the pairs solved OPERA activities such as Intrain 7 and Inverse 3. Therefore, the objects of OPERA referred neither to real
objects nor to mathematical entities while the children collaborated to solve the
OPERA problems. No help was provided during this stage.
Evaluating the Effects of Collaborative Learning

Most of the problems that students failed to solve in the pretest were solved
successfully during the posttest, although the treatment consisted only of
activities with OPERA. For the problems unsuccessfully solved during the
324
Baruch B. Schwarz
posttest, students were given clues displaying the three representations in OPERA
(see Figure 8).
CuIouboi
and nu.ben
..
I i
---j---
EJ
c. Slack Folders
Fig. 13.8 . The clues for the story problem
The machines were given as two cut-outs: one in the form of a machine, the
other in the form of a */ machine. The students had to choose the right machine
or team and to instantiate it by means of number tiles. The table was a simple
double array that the students had to fill in. The stacks were sorted in six folders:
For example, the second folder contained the piles of 2, 3, 4 . . . generic blocks;
the third folder contained the stacks consisting of piles of 2, 3, 4 . . . generic
blocks, to which one block was added to the right side; the fourth one contained
two piles of equal height (see Figure 9). After reading a story (unsuccessfully
solved), the student was invited to try to solve the problem by using one of the
three sets of clues.
Understanding Symb ols with Intermedi ate Abstractions
325
Slact.. fo'det .3
Fig. 13.9. The contents of the stack s folder s 3 and 4
Below is a short protocol of Leah using the stack representati on:
1 Exp :
2 Exp :
3
4 Leah :
5 Exp :
6 Leah :
7
8
9 Exp .
10 Leah :
11
12 Exp :
13 Leah:
14
15 Exp :
16 Leah:
17
18 Exp :
19 Leah :
Can you read t he Merry Meadows Farm problem? (Leah r eads)

Can you t ell t his story with machines, with tables, or
with stacks?
The stacks .
Which of thes e folders tells the story best?
Wi t h f older 4, and I take a stack wi th twelve things
b ecause there are twelv e crat es, and you have t o pay
the truck t o get each of t hem t o the store .
You pay how much to get them?
Thirty dollars for . . . wa i t a minute . You hav e t o
take the f older with t welve piles and the thirty dollars there
(Leah points to Folder 3 in Figure 9 ) .
Can you s olve the problem now?
OK . There are two hundred ten for all of thi s , and thirty
there, s o there are one hundred eighty in the pile.
And how much do you have in each?
There are twelve to put on one hundred eigh t y . So I divide
(She writes the division 180/12 and carries out the
calcul a t ion)
. . . so i t i s fifte en (see Figure 10)
Fifteen what ?
What' s i n the cra tes . . . f i f t een dollars f or each cr a t e .
326
Baruch B. Schwarz
In this short example , a very common phenomenon is illustrated: The students are
able to recognize the story of a problem through a stack or a team of machines;
and then, with the representation they choose, they are able to solve the problem.
I
I
.,
,
I
I
I
Fig. 13.10. Leah solving FMM with stacks
Collaborative Learning with OPERA Activities

Several types of interaction in the learning situations created around OPERA
activities are described here. All these types define cases of collaborative learning
with intermediate abstractions. Collaborative learning occurred first for very
simple activities, during which pairs observed and commented on the behavior of
OPERA objects . For example, while observing the formation of a stack for the
operator C + 2) * 3 (displayed in Figure 5c), Ron and Iris commented:
[an unlabeled block is first formed; then a block labeled 2 is placed on the right
side of the unlabeled block]
1 Ron: Now it adds the 2 . . .

2 Iris : hrrm, adds to it, and . .
[the two blocks gradually grow together into a pile with three
unlabeled blocks and three blocks labeled "2"]
3 Iris : It adds then two more of the stuff . . . .
4 Ron: Yeah, but I could not follow this with the machines.

5
6
7
8
9
10
327
I want to run it again .

[Ron selects the team [+2] [*3] and clicks on the stack icon]
Iris : It 's two more . . . and now . . .
Look ! It blinks on the multiplication by three machine.
Ron :
So it gets them to the third floor . [Ron points at the
unlabeled block and the block labeled "2" beside it]
Iris : Humm, yeah, it stretches them. . . .
Ron :
Three times.
It is striking that, although Ron and Iris were not familiar with the system, they
were quickly able to develop metaphoric explanations based on their intuitive
understanding of how the objects of OPERA behave . First, Ron identified the first
stage of the stack representation as the addition of 2 (line 1). Iris identifies the
expansion of the two blocks as adding "two more of the stuff' (line 3). Then Ron
modifies her interpretation: "it gets them to the third floor" (lines 7 and 8).
Although Iris accepts Ron's interpretation, she refines it by using the term
stretches (line 9). All these terms constitute a coherent set of explanations based
on metaphors (such as "stretches," "two more of the stuff," or "it gets them to the
third floor"). Because OPERA was designed to represent all the aspects of the
concept of mathematical operator, or what Roschelle (1992) has called the deep
structure of a concept, developing a coherent set of metaphorical explanations of
the behavior of the objects of OPERA implies developing an intuitive
understanding of the concept.
Another very salient process in this excerpt occurs when the participants take
turns in their conversation. That is the process of convergence to a shared
meaning by referring to objects of the system. For example, in line 2, Iris adds
"to it" to the explanation "it adds the 2" given by Ron . Her remark refines the
meaning of "It adds 2" in a way that is essential: In a mathematical language, an
operator such as +2 needs to be applied on numbers. This convergence of meaning
occurred here by referring to an object of the OPERA system. Similarly, in line
9, Iris uses the term stretches to qualify the meaning of "it gets them to the third
floor" uttered by Ron. At last, Ron specifies "how much" the stack stretches.
Again, in this turn-taking sequence, the second participant helps to make the
shared meaning converge. In summary, convergence is a turn-taking process in
which (a) one participant describes a property of the intermediate abstraction, (b)
the second participant accepts the meaning uttered by the first participant, and (c)
the second participant refines this meaning. Convergence often occurs when the
objects to which the participants refer are visible.
Inverse 3 was given much later than In-train 7. Two different solutions of
Inverse 3 given by two pairs are presented here. These solutions uncover different
processes of collaborative learning, processes that both stem from convergence.
The first pair, Amos and Leah, uses the representations of OPERA extensively .
328
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Baruch B. Schwarz
Amos : If I put a 2 train in t he entry [Amos creates a 2
train, puts it at the portal entry of t he t eam [*2 ] [1]
[bla::k l:ax] , am cliCks en this tE:am] , i t give:; .
two. And if I put a 3 train , it gives a thre e .
So, these machines do not change. .
[Leah takes the mouse and creates another t rain, a 5 train,
and runs it through the team]
Leah : Yeah! It does not change the trains (pause). The two
is multiplied by two, it's four , minus one, i t 's
three . So the three in the black box gives, hmm.
Amos : I do a table for all the mach i ne s together and also
for the black box separately. This one is easy , yo u
know . [Amos enters the x-values 1, 2, 3, and 4 and the
s ame y-values ]. The b l a ck box . . .
[Leah points out on t he t eam [ *2] [-1 ] [b l a ck box] and on the
first en try of t he table o f t he bla ck box (s e e Figu r e 12)]
Leah: (whispers) One by two minus one, it's still one . . .
(pause)
Amos : Yeah! It does not change . .
Leah : But for three, uh , for two . . . If two enters here
[Leah points on the t eam of machines ] it gets to four
minus one , it's three , and through the black box, it.
Amos : must get out two .
Leah: hmm!
[Leah enters 3 i n t he x -entry and 2 in the y-entry for the
table of the black box]
(Leah and Amos continue to fill in the black box table; they obtain 1,3, 5, and 7
for the x-values and 1,2,3, and 4 for the y-values).
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Leah: I know how it works . Hmm . It . . . for nine it will be

five . . . . So the black box . . . . (pause)
Amos : I want to see how the stack is . [Amos selects the team
of machines and clicks on the stack icon. Amos then
comments While observing the construction of the
stack]. It adds two , then takes off one . Now t h e
b l ack box needs to t urn i t t o t he f irs t empt y brick .
So I a dd one more ; I get only the t wo enpty br icks.
Leah: And you can reduce it i f you take off one of these blocks.
Amos: Yeah! you have to shrink it from two to one block .
Leah: So, I divide it by two .
Amos : I' l l see if it's working on the table. Nine gives
five . So nine divided by two is . . . . (pause)
Leah: No, first nine plus five is ten . And ten divided by
two is five . Yeah it's working. The black box is +1
and divided by 2 .
329
Inuene 3
How does fhe btack box work?
..
In
2
3
4
Fig. 13.11. Leah tries to construct a table for the black box
A look at this excerpt indicates that turn-taking here is often governed by

convergence, similar to the case in the first excerpt. For example, in the two turns
in lines 25 and 26, Leah and then Amos refine the definition of the operation to
be done to pass from "two empty bricks" to one: first, to reduce by taking off
(line 25, Leah); then, to shrink (line 26, Amos); and finally, to divide by two
(line 27, Leah). However, in contrast with the first example, turn-taking very
often occurs when a participant decides to pass to another representation in order
to take another perspective. The first instance of this kind of turn-taking is in line
8: Leah has difficulties in using the machine representation in order to find the
"out-train" in the black box when the "in-train" is 3. Then, in line 21, Amos
proposes to construct tables to write out several inputs and outputs . Tables have
constraints in the sense that they are only descriptive and that they do not allude
to the operations done to compute the output values. On the other hand, they
afford the recognition of patterns (as expressed by Leah in line 19). The two
instances of turn-taking instigated by Amos occurred when Amos and Leah
reached an impasse. In line 5, Leah has difficulty calculating outputs for the black
box (on line 6), and Amos decides to use them as markers; in lines 19-20, Leah
fails to express the unknown operator while observing the table, although she
seemingly recognizes a pattern in this table. Amos then decides to run the stack
representation of the team of machines.
A third type of turn-taking occurs twice at the end of the excerpt: a process of
confirmation through a different representation. One example of this turn-taking
occurs in lines 26-27, where Amos asserts that the black box "shrinks it from two
330
Baruch B. Schwarz
to one block." The referent here is, of course, the stack representation . Then Leah
confirms Amos's assertion by saying, "So, I divide it by two." It is not clear to
which representation Leah refers when using the term divide (the machine "divide
by two" or the algebraic representation). What is certain, however, is that it is
symbolic in contrast to Amos's reference to the stacks. Another confirmation
done through a second representation occurs in line 28. It is Amos's reaction to
Leah in which he tries to confirm the "divide by two" by going to the table.
During most of the sessions, Leah's and Amos's discourse was governed mainly
by convergence turn-taking, with the language used by the pair often being
metaphorical . In contrast, the other pair, Ron and Iris, moved from convergence
turn-taking to confirmations through a different representation. For example, the
interaction between Ron and Iris while solving "Inverse 3" is shown in the
following excerpt:
1 Ron: I try some trains. One gives . . . one . And a two gives
two
[Iris clicks on the 2 train backward; It yields a 2 train]
2 Iris: It's the same in the two directions. If I rotate the whole
3
stuff, it will be the same .
[Iris rotates the team of three machines and checks that the 1
and 2 trains yield the same trains . She rotates the team again
to its initial position .]
4 Ron: Yeah the two do the reverse, they annihilate each other .
5
You know, I'll cut the black box from the two others . The
6
table for the black box is . . .
[Ron enters the 0, 1, 2, and 3 x-values through the black box
and obtains the y-values 1/2, 1, 1 1/2, and 2.]
7 Ron : Let's see. If I put half in the table of the two machines,
8
I get zero . Yeah . And three gives two . It's the reverse,
9
and the black box is the rotation of the minus one and / 2 .
10 Iris: I'll try it. (Iris flips the [*2] [-1] team of machines to
11
obtain [+1] [/2]) . If I attach this to a [*2] [-1], it will
12
behave like the black box .
[Iris creates the [*2] [-1] team and attaches it to [+1] [/2] .
She then enters a 2 train and a 3 train ; they go out
unchanged. ]
13 Ron : Yeah, they do exactly the same, but in the opposite
14
direction . They kill each other . . . . I f I see that with the
15
stacks, it will return to the same thing.
[Ron selects the [*2] [-1] [+1] [/2] team and runs its stack
representation. While the running occurs, Iris gives the
following comments:]
16 Iris: OK, we get the *2, and minus one . Now the plus one eats the
17
minus one. And the /2 reduces it again to the first brick .
331
The clearest instance of confirmation through referring to a different representation

is Ron's intervention in lines 13-15. He reacts to Iris's creation of a team of
machines and the running of trains through this team. The trains go out
unchanged. Ron uses two expressions, "they kill each other" and "it will return to
the same thing," referring explicitly to the stacks. Also, Iris's turn-taking while
observing the construction of the stack of the team [*2][-1][+1][12] is the same
kind of confirmation . She sees stacks but comments on Ron's action in terms of
machines. In fact, the whole conversation between Ron and Iris in lines 2-11
refers to both machines and tables interchangeably; the two actors know that they
refer to the same entity through different representations.
In summary, the two groups intensively used metaphorical accounts to explain
the behavior of the intermediate abstraction. Three main kinds of turn-taking were
found. The first one is a process of convergence to a shared meaning through
reference to objects of the system. This process is initiated by one participant
describing a property of the intermediate abstraction, with the second participant
accepting the meaning uttered by the first participant and developing this
meaning. The second turn-taking process is also convergent. It is initiated by an
impasse with one particular representation and is characterized by a change of
representation . The third turn-taking occurrence is a confirmation through a
different representation. Leah and Amos used mainly the first two kinds of turntaking, whereas Ron and Iris's conversation was often governed by confirmation.
In this chapter, I do not provide a detailed comparison of the performance of the
six students between the pretest and the posttest. Almost all the formal tasks
(numerical as well as those demanding the finding of a formula) were solved by
the six students. There was also a substantial improvement on word problems that
demanded the finding of a formula . This improvement, I believe, reflects that
OPERA displays operators as objects by themselves, even when applied to
numbers . The interaction of the experimenter and the students on problems on
which students failed in the posttest was somehow compatible with the kinds of
interaction that took place between students and gave clues about the status of
representations during collaborative learning.
Leah's and Amos's interactions were governed by the two processes of
convergence (refinement of meaning and initiated by impasses) . Leah failed to
solve the Merry Meadows Farm problem in the posttest. The protocol of Leah's
solution when working with the experimenter shows how well she is able to
match stacks to the objects of the word problem. For example, in lines 6-7, she
says, "I take a stack with twelve things because there are twelve crates." Lines 1617 and Figure 11 show how Leah could take advantage of the stack she chose to
find the solution to the problem. It seems that she is able to find what Greeno et
al. (1993) called a morphism between the two systems , the stacks and the story .
The questions instigated by the experimenter were correctly answered by Leah.
Although she is able to map the relations, actions, and quantities of the story
onto a stack, her reasoning is always local, taking into consideration the stacks or
the story . Therefore, almost all of her moves (which are done with the stack
332
Baruch B. Schwarz
representation) are preceded by a question from the experimenter, referring to the

story . For example, to the experimenter's question, "fifteen what" in line 18,
Leah answers "what's in the crates" and then corrects her answer to "fifteen dollars
for each crate."
Ron and Iris were very successful in the posttest, and their improvement from
the pretest was more substantial than that of Leah and Amos . However, the
number of subjects in this experiment was small, and, as mentioned earlier, I have
restricted myself to discussing the link between the work in pairs and how
OPERA objects facilitate the solving of difficult problems. A problem on which
Ron failed was The Students and the Professors, for which he wrote the relation
P = 6S. When asked about this relation with the OPERA objects, he showed how
flexible these tools were:
Exp :
Ron:
What do you want to use to find the link between students

and professors, stacks, machines, or a table?
I 'll take the machines . I'll take a multiplication
machine , and I put into it a six, because they are six
times more . . . . So, if there are 600 students . Oh, no , I
have to turn it [Ron flips the *6 machine] . Yeah, it's the
opposite ; it will be a division by 6 machine .
Ron first goes wrong in choosing the multiplication machine but corrects himself
immediately by flipping it and saying, "It's the opposite; it will be a division by
6 machine." Ron uses the same object but changes its position to modify his
answer. In fact, Ron knows that the formal relations "multiply by six" and "divide
by six" are essentiality the same, depending on how the relation between the
professors and the students is put. The machines are not a way to facilitate
matching between a symbolic and a nonsymbolic system, but they are used in a
flexible way to express the relation .
Discussion
I have analyzed the construction and interpretation of symbolic representations in
terms of attunements to constraints and affordances. Learning a representational
system involves more than acquiring referential correspondences between symbols
and their referents . It requires knowledge about how to reason about the
represented system, how to construct and use symbolic expressions, and how to
create correspondences between symbols and their referents. However, the
achievement of coordination among all the representations is difficult: Multi-
333
representational systems with multiple constraints (e.g., the deliberate choices in

the design of each representation) and affordances (e.g., the actions or the
interpretations that stem from the behavior of objects in each representation) are
difficult to comprehend as a coherent whole. This study showed that
conversational processes of acceptance stemming from collaborative learning made
possible construction of shared meaning in the development of the concept of
mathematical operator. Three such processes were found. The first process is one
of convergence to a shared meaning through reference to objects of the system.
This process is initiated when one participant describes a property of the
intermediate abstraction and the second participant accepts the meaning uttered by
the first participant and specifies this meaning . The second process, also one of
convergence, is initiated by an impasse within one particular representation and is
characterized by a change of representation. The third process is one of
confirmation through a different representation.
In this study, each collaborative pair solved the OPERA activities, solved most
of the formal problems (numerical as well as demanding the finding of a formula),
and successfully solved the word problem demanding the finding of a formula. The
session with the experimenter seems to indicate that OPERA could help students
construct a semantic interpretation of symbolic representations that could be
learned more meaningfully. Students learned to construct and interpret symbolic
representations of quantitative properties of the OPERA system. These results are
consistent with studies undertaken with other intermediate abstractions in which
students who worked in pairs could apply their knowledge in novel situations (see
Schwarz, Kohn, & Resnick, 1993, for the concept of negative numbers; Schwarz,
Nathan, & Resnick, 1996, for arithmetical word problems) .
334
Baruch B. Schwarz
References
Clark, H. H., & Schaefer, E. F. (1989) . Contributing to discourse . Cognitive Science,
13, 259-294.
Clark, H. H., & Wilkes-Gibbs, D. (1986) . Referring as a collaborative process .
Cognition , 22, 1-39.
Clement, J. (1982) . Algebra word problem solutions: Thought processes underlying a
common misconception. Journal for Research in Mathematical Education, 13, 1630 .
Greeno, J. G. (1989) . Situation models, mental models, and generative knowledge. In
D. Klahr & K. Kotovsky (Eds.), Complex information processing : The impact of
Herbert Simon (pp. 285-318). Hillsdale, NJ: Erlbaum.
Greeno, J. G., Engle, R. A., Kerr, L. K., & Moore, J. L. (1993 , July). Understanding
symbols: Situativity-theory analysis of constructing mathematical meaning. In
Proceedings of the Fifteenth Annual Meeting of the Cognitive Science Society (pp .
504-509). Boulder, CO.
Kirshner, D. (1989) . Critical issues in current representation system theory . In S .
Wagner & C. Kieran (Eds.), Research issues in the learning and teaching of algebra
(pp. 195-198) . Hillsdale, NJ: Lawrence Erlbaum Associates.
Larkin , J. L. (1989) . Robust performance in algebra : The role of the problem
representation. In S. Wagner & C. Kieran (Eds.), Research issues in the learning and
teaching of algebra (pp. 120-134). Hillsdale, NJ: Lawrence Erlbaum Associates.
Matz, M. (1980). Towards a computational theory of algebraic competence. Journal of
Mathematical Behavior, 3, 93-166.
Roschelle, J. (1992) . Learning by collaboration: Convergent conceptual change . The
Journal of the Learning Sciences, 3, 235-276 .
Schoenfeld, A. H. (1987) . When good teaching leads to bad results : The disaster of
"well taught" mathematics courses . Educational Psychologist, 23, 145-166 .
Schwarz, B. B. (in press) . Why can intermediate abstractions help acquiring robust
representations? Interactive Learning Environments.
Schwarz, B. B., Kohn, A. S., & Resnick, L. B. (1993). Positives about negatives . The
Journal of the Learning Sciences, 3(1), 37-92 .
Schwarz, B. B., & Nathan, M. J. (1993, July) . Assessing conceptual understanding of
arithmetic structure and language. In Proceedings of the Fifteenth Annual Meeting of
the Cognitive Science Society (pp. 912-917) . Boulder, CO.
Schwarz, B. B., Nathan, M. J., & Resnick, L. B. (1996). Acquisition of meaning for
arithmetical structures with the Planner. In S. Vosniadou, E. De Corte, R. Glaser, &
H, Mandl (Eds.), International perspectives on the construction of technology-based
learning environments (pp. 61-80). Mahwah, NJ: Erlbaum.
Smith, C., Snir, J., & Grosslight, L. (1992). Using conceptual models to facilitate
conceptual change : The case of weight-density differentiation . Cognition and
Instruction, 9, 221-283 .
White, B. Y. (1993) . Thinkertools: Causal models, conceptual change, and science
education. Cognition and Instruction, 10, 1-100.
335
Wiser, M. (1988, April) . The differentiation of heat and temperature: An evaluation of

the effect of microcomputer models on students ' mis conceptions. Paper presented at
the meeting of the American Educational Research Association, New Orleans .
Chapter 14
Strategy-Specific Information Access in

Knowledge Acquisition from Hypertext
Wolfgang Schnotz
Department of Psychology, University of Landau, Germany
Abstract
A central issue in learning from hypertext is the nonlinearity of presenting
information. Nonlinearity, however, does not have a facilitative effect on
learning per se. The essential point is rather how learners make use of this
nonlinearity, that is, how they interact with hypertexts, which strategies they
use, and what kind of knowledge they construct. This chapter is concerned with
differences between successful and unsuccessful strategies in learning about a
complex subject matter from hypertext. In an empirical study, subjects had to
answer a series of questions with the help of a hypertext. Cognitive processing
was analyzed through the think-aloud method . A group of successful learners and a
group of unsuccessful learners were identified by their scores in a comprehension
test presented after learning . The results indicated that the two groups used
different strategies, which were reflected by different patterns of information
access. The successful learners tried to construct a mental model of the subject
domain, which was then successively elaborated . They combined model
construction, model analysis, and symbol processing in a flexible way, and tried
to solve inconsistencies by elaborating the respective model structure . In
contrast, unsuccessful learners tried to get by with a symbol processing strategy
and tackled the construction of a mental model only when the former strategy
could no longer be applied. Inconsistencies were solved by ad hoc repairs of
procedures rather than by systematic elaboration of the model structure. The
respective strategies can be related to a global and a local orientation of
knowledge acquisition. Instructional implications for the selecting and
Strategy-Specific Information in Knowledge
337
sequencing of learning tasks , as well as for the display of information in the

design of learning environments with hypertext, are pointed out.
Introduction
Knowledge acquisition is generally perceived today as an active process of
constructing complex mental representations. Learners who have become experts
in a specific field have acquired highly integrated knowledge structures of the
respective domain (Chi, Glaser, & Rees, 1982). Educationalists, however, usually
complain about the fact that learners only accumulate isolated pieces of knowledge
that are not integrated into a coherent whole (cf. Carey, 1986; diSessa, 1983;
Vosniadou & Brewer, 1990). Various reasons for this lack of mental integration
include the method of presenting information and the media used for this purpose .
The assumption is that traditional learning materials such as written texts do not
adequately take into account the active and constructive nature of human
knowledge acquisition, because learners are only in a passive role of receiving new
information. Furthermore, the linear format of such texts is said to be inadequate
for presenting complex subject matters, because the respective network-like
structure has to be transferred into a linear structure during writing by the author
and subsequently must be retransferred into a network during comprehension by
the learner. Due to these processes of linearization and delinearization, learners
frequently would not recognize the complexity of the content (Spiro, Feltovich,
Jacobson, & Coulson, 1991).
Today, the development of new technologies provides new possibilities for
presenting information in a way that is assumed to be better adapted to the active
and constructive nature of human learning and that allows the explicit display of
the network-like structure of knowledge. One kind of technology receiving special
attention in this context is the so-called hypermedia. Hypermedia are electronic
documents with a nonlinear, network-like structure consisting of nodes and links.
Nodes have the function of information containers, wherein texts, static or
animated pictures and graphics, sound, and simulations can be stored. If they
contain only text, the document is refered to as a hypertext. Links have the
function of relating the information of a node to the information of other nodes
(Gloor & Streitz, 1990; Jonassen & Mandl, 1990; McAleese, 1989; McAleese &
Green, 1990; Rizk, Streitz, & Andre, 1990).
On the one hand, hypertext systems provide new ways of individualizing
learning processes due to the possibility of a flexible access to large and
interrelated sets of information . The learner can inspect the node contents by
selecting his or her individual path to follow in a nonlinear structure in order to
fill in knowledge gaps (Yankelovich, Landow, & Cody, 1987). On the other hand,
however, such systems impose greater demands concerning the self-regulation of
338
Wolfgang Schnotz
learning . The individual needs to have his or her own criteria for information
access. Existing knowledge gaps have to be recognized as such, which is
frequently rather difficult (cf. Baker, 1985; Glenberg, Wilkinson, & Epstein,
1982). The information needed then has to be searched for in a large body of
accessible information without losing sight of the searching aim (Carmel,
Crawford, & Chen, 1992; Guthrie, Britten, & Barker, 1991; Guthrie &
Mosenthal, 1987). The more possibilities are available to access different kinds of
information, the higher the probability will be that learners become disoriented or
get lost in hyperspace (Conklin, 1987; Jonassen & Wang, 1993). Therefore,
learners need specific navigational aids, such as maps, indices, dynamic glossaries,
conditioned paths, or guided tours. However, such help also imposes additional
cognitive requirements (cf. De Young, 1990; Kobasigawa, 1993; Lucarella, 1990;
Saxer & Gloor, 1990).
Furthermore, some characteristics of human comprehension and learning can
hardly be appropriately taken into account with hypertexts . During comprehension
of a coherent text or discourse, a flow of thought ensues (Chafe, 1979): The
individual is constructing a topic-specific mental representation that is
successively elaborated from sentence to sentence as long as the respective topic is
maintained. Accordingly, it is possible to guide this flow of thought by means of
a well-designed thematic sequencing so that the learner can successively elaborate
his or her knowledge. Designing hypertexts, however, requires a modularization of
the respective knowledge so that the material displayed remains coherent and
understandable in any type of sequencing. It is very difficult in this case to
maintain a continuous flow of consciousness by thematic continuity (cf. Brown,
1990). Therefore, some authors consider hypertexts to be primarily data bases,
which serve experts in up-dating their knowledge rather than the acquisition of
new knowledge by novices (cf. Jonassen & Grabinger, 1990). Empirical findings,
however, are not clear in this respect. Samarapungavan and Beishuizen (1992), for
example, found that novices demonstrated better understanding of a complex
domain after learning from a hypertext, whereas other investigators found learners
had better understanding after learning from a traditional text (cf. McGrath, 1992).
Earlier discussion has frequently suffered from gross oversimplification of the
differences between hypertext and traditional text and from the erroneous
assumption that the new medium has inherent properties that have per se a
facilitative effect on learning. On the one hand, for example, it has been
downplayed that learning from traditional text is also an active process in which
learners construct network-like mental structures supported by linguistic devices
such as anaphora (cf. Smith & Weiss, 1988; van Dijk & Kintsch, 1983). On the
other hand, it has been overlooked that processes of linearization and
delinearization are also necessary in learning from hypertexts. The learner can
admittedly select his or her own navigation path through the material and decide
which information is to be accessed next. But this information ultimately also has
to be processed sequentially, the only difference being that the linearization is
made by the learner instead of the author. In other words, nonlinearity per se has
339
neither a positive nor a negative effect on knowledge acquisition. Rather, the

critical point is how learners make use of this nonlinearity.
This chapter aims to analyze this topic by focusing on how learners interact
with a hypertext system, which strategies they use, and which strategies lead to
which type of knowledge structure. First, a general framework for the analysis of
task-oriented knowledge acquisition is presented . Second, an empirical study aimed
at analyzing differences between successful and unsuccessful strategies in learning
from hypertext is described . Third, practical implications are discussed on how to
design appropriate learning tasks and how to design hypertext systems in order to
activate successful learning strategies.
Theoretical Framework
In investigating processes of knowledge acquisition, a central issue is the structure
and use of this knowledge. The question that should be asked concerns the nature
of the respective mental representations and the processes that operate on them .
Another central issue is the dynamic of knowledge acquisition . We need to ask
how learners realize that they have knowledge deficits and how they are induced to
successively elaborate and modify their knowledge. These questions will be dealt
with here. Because the acquisition of knowledge is usually embedded in and
stimulated by the use of knowledge, acquisition and use of knowledge will not be
considered separately but rather as two sides of an integrated entity (cf. Brown,
Collins, & Duguid, 1989).
Mental Representations and Mental Procedures

Recent cognitive theories assume that knowledge acquisition includes both the
construction of a propositional representation and, based on this representation, a
mental model of the situation described in the text (Johnson-Laird, 1983; Morrow ,
Greenspan, & Bower, 1987; van Dijk & Kintsch, 1983). A propositional
representation consists of complex symbols that are assembled from more simple
symbols acccording to specific syntactic rules and that refer to the signified object
by means of a certain semantic. In a representation of geographical facts, for
example, the proposition (EAST (is: Rio de Janeiro, of: Mexico indicates that
Rio de Janeiro is east of Mexico. A mental representation consisting of
propositions corresponds to a description of an object in a hypothetical mental
language. In contrast, a mental model is a temporary analogue mental structure:
an internal quasi-object that is an analogue to the signified object with respect to
certain structural and functional characteristics. Mental images as a specific form
of analogue mental representations can be considered to be projections of a mental
340
Wolfgang Schnotz
model from a certain perspective (Johnson-Laird, 1983; cf. Baddeley, 1992;

Kosslyn, 1981) .
In order to use knowledge as a tool of orientation about a domain , certain
procedures have to operate on the respective representations and produce the
required information (cf. Palmer, 1978). There are different types of procedures,
depending on the kind of mental representation on which it will be applied. A
propositional representation allows the production of new information by means
of symbol-processing procedures . The following procedure, for example ,
IF
(EAST (is: place A, of: place B))
TIffiN
(LATER (is: time at place A, than: time at place B))
could be applied to the earlier proposition (EAST (is: Rio de Janeiro, of: Mexico
and thus create the new proposition (LATER (is: time at Rio, than: time at
Mexico. To find out that the time is 5:00 pm in Rio, when it is 2:00 pm in
Mexico, one could apply the following procedure
IF
The time of day at place A is given and the time

of day at place B has to be determined,
TIffiN
take the time of day at place A, subtract the time

difference between place A and Greenwich, add the
time difference between place B and Greenwich, and
set the result as the time at place B.
Such symbol processing requires no reference to the subject domain. The rules can
be used even when an individual knows neither where the respective places are nor
what the terms ~ or later mean. It is sufficient to handle the symbols in a
purely syntactical way.
Mexico
Rio
GreenWich
14
15
16
17
13
19
20
Fig. 14.1. Example of a spatial configuration for a model-based determination of a

time difference.
341
Another type of information processing is performed with mental models . To

answer the question about the time of day in Rio when it is 2:00 pm in Mexico,
one can construct a model with a structure analogue to the spatial configuration
shown in Figure 1. From such a configuration , one can determine directly, by a
left-right comparison, that Rio is located east of Mexico and that the time of day
in Rio is later than the time of day in Mexico. Furthermore, one can determine
(by counting) that the difference between Rio and Mexico is 3 hours and the time
in Rio is 5:00 pm. Because this kind of information processing is totally different
from applying symbol-processing rules, it is referred to as pseudo-inferences
(Garrod, 1985). Such model-based inferencing requires procedures both of model
construction and of model analysis. Model construction entails the initial
construction of a model as well as possible model revisions. Model analysis
entails a search for certain entities and a determination of their relations to other
entities. It can be assumed that both model construction and model analysis are
driven by cognitive schemata (cf. Pinker, 1990). In model construction, these
schemata function as descriptions of prototypical facts (specified on the basis of
propositional knowledge) and, thus, can guide the process of model construction
(Johnson-Laird, 1987). In model analysis, these schemata serve to read off certain
information from the respective configuration and to add new information to the
propositional knowledge as does the proposition EAST (is: Rio, of: Mexico) in
the earlier example. In other words, propositional representations and mental
models interact through schema-driven processes of model construction and model
analysis.
Information Deficits and Search for Information

Construction of knowledge in learning from text ensues generally in view of a
certain use of knowledge (Marton & Saljo, 1976; Melton, 1978; Pichert &
Anderson, 1977). This usability aspect can be expected to be even more salient in
learning from hypertext, because learners have to apply their own criteria for
accessing information, which, in turn, affords clear objectives. Otherwise, they
would explore the available information unsystematically, do their sequencing at
random, and not reach an encompassing view of the facts (Jonassen & Grabinger,
1990).
One can assume that learners search especially for the kind of information that
allows them to compensate for obvious knowledge deficits (Garner & Reis,
1981). Knowledge deficits become obvious when the respective knowledge has to
be used. If learners use their knowledge to solve a certain cognitive task (e.g.,
answering questions requiring inferences), they apply certain cognitive procedures
to their available knowledge. The applicablity of these procedures, however,
requires the respective knowledge to have a specific structure. A mental structure
can therefore be considered deficient in a specific, task-defined sense, if the
respective structural requirements are not met. One kind of a structural deficit is
342
Wolfgang Schnotz
that a task-relevant procedure cannot be applied. If, for example, the question
concerning the time and date in Anchorage when it is 3:00 pm Thursday in Tokyo
is to be answered in a model-based way and the model contains no time zones,
then determining time differences is not possible. Another kind of a task-defined
structure deficit is that the procedures applied lead to inconsistent results. If, for
example, the model contained time zones but no dateline, learners would conclude
that the time is 8:00 pm Wednesday in Anchorage if they go to the west, and that
it is 8:00 pm Thursday in Anchorage if they go to the east. In both cases, the
model structure would have to be elaborated appropriately to avoid these
problems .
The assumption that knowledge deficits become obvious when the respective
knowledge is used implies that the existence of knowledge deficits can remain
unnoticed due to a limited, too-restricted use of this knowledge. According to this
view, the opportunity that learners have to recognize deficits in their knowledge
depends to a great extent on the demands with which they are confronted. In the
following study, learners were given the task to answer a certain sequence of
questions. The questions were designed and sequenced in such a way that the
learners were required to successively elaborate their mental representation. To
perform these elaborations, the subjects had access to a hypertext about the
respective topic. The analysis concerned the learners' interaction with the
hypertext and focused especially on the question concerning whether successful
and unsuccessful strategies could be distinguished.
Method
Subjects and Learning Task
Seventeen German university students participated in the study. They were asked
to answer seven questions step by step. Because these questions were expected to
stimulate learning activities, they are referred to as learning questions. The
questions (presented in German) were:
Ql)
Q2)
Q3)
What i s the time in Tokyo , when it is 2 :00 pm in Cairo?

(correct answer: 9:00 pm.)
What is the time in Rio, when it is 11:00 am in Cairo?
(correct answer : 6 :00 am)
What i s the time in Mexico City and what date is it
there , when it is SUnday 9 : 00 am in Frankfurt?
(correct answer : Sunday 2 : 00 am)

Q4)
Q5)
Q6)
Q7)
343
What is the time in Rio and what date is it there, when

it is Tuesday 10 :00 pm in Mexico City?
(correct answer : Wednesday 1: 00 am)
What is the time in Anchorage and what date is it
there, when it is Thursday 3 : 00 pm in Tokyo?
(correct answer : Wednesday 8 : 00 pm)
Why did Magellan's compani.ons , after sailing around the
world in a Western direction, think that their day of
arrival was Wednesday when indeed it was already
Thursday?
(exarrples of correct answers : They missed one sunrise .
They did not know about a dateline in those days)
One can orbi t around the earth several times wi thin a
couple of hours and pass the dateline several times .
Would it be possible, because the date has to be
changed there each time, to fly into the past or into
the future?
(correct answer : No, because the date also changes at
midnight in the opposite direction)
If subjects answer these questions on the basis of a model-based strategy, they

have to elaborate their mental models successively. Ql and Q2 require only a
model containing time zones of the corresponding earth sector. From Q3 onward,
a date is also to be determined so that zones of different dates can be distinguished.
Provided that individuals usually consider the geographically shorter distance
between the respective locations, Q4 requires a change of date due to the transition
from 11:00 pm of one day to 0:00 pm of the next day. In Q5, a change of date due
to crossing the dateline is to be considered, if the geographically shorter distance
is chosen .
In contrast, if subjects follow a symbol-processing strategy by answering the
questions as far as possible in a numerical-algorithmic way, they do not need such
an elaboration of their mental structure. To answer Ql through Q5, they would
need only an appropriate numerical algorithm and the availability of the
corresponding numbers . To answer Q6 and Q7, however, such purely syntactic
symbol processing would no longer suffice. In a previous pilot study, the subjects
reported unanimously that they had to imagine the facts with these questions.
Thus, one can assume that a model-based strategy is necessary here.
If all questions were to be answered through a model-based strategy, an
elaborated model covering all relevant facts would be required. The structure of
such a model is presented graphically in Figure 2. Accordingly, the earth rotates
within a time shell wherein different time conditions (time and date) are spatially
distributed. Due to the interplay between the dateline, which is locally fixed on
earth, and another (variable) line, which can be called the midnightline, the model
includes date zones of varying size.
344
Wolfgang Schnotz
12 h
dayi
6h
18 h
.~"":""""""""f---
date line
m idnight line
Oh
day i+1
Fig. 14.2. Structure of a comprehensive model required to answer learning questions

QI through Q7 .
Learning Material
To answer lhe learning questions, the subjects had a hypertext about time and date
at their disposal . The text consisted of 67 paragraphs with a total length of 65] 6
words (including headings) . Because previous studies have shown that hierarchical
structures cause less orientation problems than associative network structures, the
hypertext system in this study was given an easily comprehensible hierarchical
structure (Gray, 1990; Mohageg, 1992). The paragraphs were subsumed under four
main topics: (1) Geographical relations on the earth's surface, (2) local time and
zone time, (3) time differences on the earth, and (4) date differences on the earth.
Nine paragraphs contained information about algorithms with numerical examples
of how to compute time differences between different locations on the earth.
Another 58 paragraphs contained information about spatial and temporal relations
on the earth's surface. The subjects had free access to the different text paragraphs
through a hierarchically organized menu. Furthermore, the subjects could navigate
sequentially by means of forward and backward buttons as well as by keyword
buttons within the hypertext. The text sections that were already accessed in a
learning session were signaled graphically in the menu but remained available .
345
Procedure
The study was performed in single subject sessions that included the follow ing
phases: diagnosis of individual learning prerequisites, training to use the hypertext
system, knowledge acquisition, comprehension test, and interview.
Diagnosis of learning prerequisites. The subjects were interviewed with regard
to their knowledge about time differences on the earth: They were asked what time
zones are and what relationship exists with the earth's rotation. In addition, they
were asked whether the time in the east is earlier or later, what the dateline is, and
how to proceed in crossing it. Furthermore, they were asked in which country or
on which continent the previously mentioned cities are located. Finally, the
subjects received a German intelligence test combining verbal, nonverbal, and
spatial measures (Amthauer-Test).
Training to use the hypertext system. After diagnosing their learning
prerequisites, the subjects were given explanations on how to use the hypertext
system with other learning content. Furthermore, they were stimulated to think
aloud, that is, to verbalize what came into their minds during learning and
thinking.
Knowledge acquisition. Subjects were then presented questions Q 1 through Q7
step by step on a computer screen. In addition, the screen contained the
hierarchical hypertext menu, which allowed the subjects to access freely the
various sources of information. The topical question remained visible all the time.
After answering a question , the subjects were presented with the next question and
still could freely access the learning material in the hypertext. They received no
feedback concerning the correctness of their answers. The subjects were asked to
continually verbalize what was going on in their minds . Their verbal statements
were tape-recorded.
Comprehension test and interview. After the learning phase, the subjects had to
answer 25 relatively complex questions to enable us to check how well they had
understood the facts. For example, they were asked what day it would be in Alaska
if one started by plane on a Thursday at 9:00 pm in Japan and landed in Alaska
after an 8-hour flight. Another question asked which days exist simultaneously on
earth and how broad are the respective date sections if it is Monday 8:00 pm in
Greenwich. The items could be answered without time limit. In a final interview,
the subjects were asked about their strategies for answering the learning questions
and their information-access strategies.
Scoring
Besides assessing the subjects' intelligence scores, prior knowledge about time
and date differences on earth was evaluated with regard to the number of domainspecific concepts mentioned in answering the prior knowledge questions. The
think-aloud protocols from the learning phase were analyzed qualitatively with
346
WolfgangSchnotz
regard to the question-answering strategies adopted by the subjects. The frequency

of information access to the different hypertext nodes was recorded automatically.
The level of comprehension was determined by the number of correctly answered
items in the comprehension test. The final interview was again analyzed
qualitatively.
Results
The students at each end of the learning spectrum were selected in order to
compare the strategies of successful and unsuccessful learners. Each of these two
groups, one with the highest and one with the lowest comprehension test scores,
represented one third of the original group. The six subjects with the highest
scores are called here the successful learners. Their average comprehension score
was 18.8 (SD = 1.8). The six subjects with the lowest scores are called the
unsuccessful learners. Their average score was 13.8 (SD = 2.3). With regard to
prior knowledge or intelligence scores, there was hardly any difference between the
two groups. The averageprior knowledgescore was 7.4 for the successful learners
(SD = 1.6) and 7.1 for the unsuccessful learners (SD = 1.4). All subjects knew
that it is later in the East due to the earth's rotation. They also knew about time
zones, and nearly all knew about the dateline and its approximate position.
However, all subjects wereuncertain about how the date had to be changed there,
depending on the direction in which one would cross this line. They were all able
to localize the cities correctly, except for Anchorage. Only one subject knew that
this city is in Alaska. The successfullearners had an average intelligence score of
110 (SD = 3.7); the unsuccessful learnershad an averageintelligence score of 109
(SD = 6.0).
Question Answering
Concerning the subjects' answers to the learning questions, there were clear
differences between the two groups. Figure 3 shows the relative proportion of
correct answers within both groups across the seven learning questions. Although
most of the successful learners answered all of the questions correctly, the
unsuccessful learners showed remarkable fluctuations. The number of correct
answers was clearly reduced in Q3 (where the date was questioned for the first
time). Q4 was again answered correctly by most subjects, but in Q5 (where
crossing of the datelinewas to be considered for the first time), another clear drop
was observed. Finally, in Q7 (where both the dateline and the midnightline were
to be accounted for), yet another decrease in the proportion of correct answers
ensued with the unsuccessful learners. The test for statistical significance of the
347
respective differences for Q3, Q5, and Q7 between both groups resulted in p<.05 ,
p-c.Ol, and p = .13 (Mann-Whitney).
100 %
\
\
5 0%
\
\
/
\
0% +------,J----+-
Q1
Q2
IQ..
--+-- -J--- -+- - -+------l

Q3
Q4
Q5
Q6
Q7
Fig. 14.3 . Relative frequency of correct answers within successful and unsuccessful
learners across learning question s Ql through Q7.
Q uest io n-Answering Strategies

In both the information access data and the think-aloud protocols, there were clear
indicati ons that the subjects in the two groups had used different strategies for
answering the learning questions . In their information retrieval, the unsuccessful
learners had concentrated more on numerical-algorithmic information. They hal
spent, on average, 132 sees (SD = 73) for the analysis of examples showing how
to calculate time differences, which was more than three times longer than the
average time of 41 sees (SD = 58) of the successful learners (p<.05, MannWhitney) . Accordingly, the unsuccessful learners often operated with symbols in a
purely syntactical way without any reference to the designated facts. The
following extract provides an example of one subject's think-aloud protocol :
I haven ' t heard about Anchorage . But i t doesn't matter
where it is . I can also answer the question by means of t he
f Onmlla .
In some cases, the subjects only copied the surface structure of the respective
examples of computing time differences from the text, as expressed in the
following comment:
348
Wolfgang Schnotz
I have to add the time of day of the city mentioned after
the time delay of the first-mentioned city . Then I have to
subtract the time delay of the second-mentioned city .
Obviously, whether the respective city is west or east of the other one and what
time differences should therefore be expected play no role here. Accordingly, the
unsuccessful learners often expressed syntactical problems, such as where to fill in
numerical values in a formula and how to take into account a minus sign. The
following extract provides some typical remarks:
What do I have to insert where now?
Now I have two negative numbers; which one do I have to
subtract from the other one?
Do I have to add or subtract now?
What shall I do with the negative sign?
These kinds of remarks appeared 1.5 times (SD = 1.0), on average, with the
unsuccessful learners, but only 0.3 times (SD = 0.5) with the successful learners
(p<.05, Mann-Whitney) . In contrast, the successful learners did not restrict
themselves to using only a symbol-processing strategy but rather employed a
model-based strategy from the beginning. They obviously tried to localize the
respective cities mentally, through their geographical prior knowledge, and to
integrate the respective entities into a coherent image. The protocols of those
subjects frequently included comments such as the following:
It is disturbing to know just the time zone value of the
city, although this is, in principle, sufficient to
calculate the result . I want to be able to imagine the
facts as well.
Now, Rio is west of Cairo, Mexico is even further west,
Tokyo is east of Cairo. That means . . . The time in Tokyo
is the most advanced of all those cities .
I would like to know Where Anchorage is. I will assume now
that it is somewhere in the West of the U.S .A .
When answering questions Ql through Q5, which could in principle be answered

with either a purely numerical-algorithmic strategy or a model-based strategy, the
successful learners frequently combined model construction, model analysis, and
349
symbol processing in a flexible way. This is illustrated in the following excerpt

from one subject's protocol:
Well , Mexi co is 6 hours t o the west of Greenwi ch, and Rio

is 3 hours t o the west o f Greenwich . . . . If we put the
distance Mexico-Rio and the distance Ri o-Greenwich
together, then this corresponds to the distance between
Mexico and Greenwich . So, I have t o subtract these 3 hours
from those 6 hours . That's o. k . . . . Now, is it earlier or
later? . . . Rio is to the east of Mexico . That is
. . . it is later in Rio . . . three hours later . Three
hours after 10 :00 pm Tuesday is . . . It should be 1 :00 am
Wednesday in Rio.
The subject first started with a model construction by means of building a spatial
configuration: Mexico-Rio-Greenwich. Then a step of model analysis ensued, by
distinguishing between the whole and its parts within the configuration by
applying a part-whole schema. The result led to the decision that a symbolprocessing procedure was to be applied: In this case, the required difference
(Mexico-Rio) was a part of a whole (Mexico-Greenwich). Therefore, a subtraction
was necessary: The smaller amount (Greenwich-Rio) had to be subtracted from the
bigger amount (Mexico-Greenwich). If the required difference had been the whole
(such as in Rio-Cairo), however, the respective differences (Rio-Greenwich and
Greenwich-Cairo) would have been parts of a whole that, therefore, had to be added
in order to determine the required time difference. Finally, another model analysis
was performed testing whether Rio is west or east of Mexico. This analysis again
resulted in a decision about the next symbol-processing operation : Because Rio is
east of Mexico and it is later there, the difference determined before (3 hours ) had
to be added to the time in Mexico (10:00 pm Tuesday), which then (after taking
into account the change of date at midnight) resulted in 1:00 am Wednesday.
Model Elaboration Versus Repair of Procedures

In the answers to the last two learning questions , which according to the preceding
assumption required a model-based strategy, differences between successful and
unsuccessful learners could also be observed. This was especially true when
answering Q7 (concerning why it is impossible to fly into the past or the future
by crossing the dateline repeatedly). The subjects had difficulty associating the
familiar event, midnight, with a spatial entity, a boundary rotating around the
earth, and recognizing that, apart from the dateline, there is another line where the
date also changes , and that the two changes per circumvention compensate each
other. Five of the successful learners tried further to elaborate their models by
350
Wolfgang Schnotz
means of mental simulations (cf. Brown, Burton, & de Kleer, 1982). The
following extracts from subjects' protocols indicate the performance of such
simulations, which finally resulted in a discovery of the midnightline:
If I start in the late evening going from Frankfurt to New

York , and if the plane is not as fast as a rocket but is
slower than the sun. . . . In that case I would arrive in
New York the following day . . . even according to the
local time there, although I did not cross the dateline.
The question is whether there are two datelines .
The dateline has two different sides . On the left side,
there is a different day than on the right side . I imagine
now that the dateline generates these different days on
each of its sides . Now, the two days extend over the earth .
. . . The earth i s round. . . . Yes, those two days have to
meet somewhere . There should be another dateline . . . .
In the protocols of the unsuccessful learners, only two indications of such mental
simulations were found. Instead, those subjects tended to answer the question
(why could one not reach the past or the future by frequently crossing the dateline)
by stating that, in certain cases, the correction of the dateline had to be
suppressed. Related remarks made included the following:
If one returns to the same time zone from which one
started, one should not correct the date at the dateline .
The correction of the date at the dateline should probably
be performed only once in 24 hours .
Instead of elaborating the structure of the respective mental model, these subjects
performed, as it were, ad hoc repairs of their procedures by reducing the
applicability of those procedures . Indications of such procedure repairs were found
with three of the unsuccessful learners but with none of the successful learners.
Learning Time and Information Access

With an average of 31 minutes (SD =12.7), the successful learners did not invest
much more time in the learning process than the unsuccessful learners with 26
minutes (SD = 7.6) . This difference was not significant. Figure 4 shows the
average learning time used by both groups across questions Q1 through Q7 .
351
Whereas the groups hardly differed from each other between Q2 and Q7, the
successful learners used nearly twice as much time before answering the rather
easy question Ql (475 sees (SD=207 as did the unsuccessful learners (245 sees
(SD = 109 (p<.05, Mann-Whitney ).
500
400
300
200
100
0+----f---1---f-----f---+---+----i
01
02
03
04
05
o6
07
Fig. 14.4. Pattern of the average learning time invested by successful and
unsuccessful learners across questions QI through Q7.
Concerning information retrieval, there were also significant differences between

successful and unsuccessful learners: The average number of times that
Information was accessed was 41 for the successful learners (SD = 16.6), whereas
it was only 25 (SD = 6.5) for the unsuccessful learners (tOO) = 2.13, p< .05) .
Figure 5 shows the pattern of the average number of times that information was
accessed within both groups across Ql through Q7. The unsuccessful learners
showed a rather flat pattern of information access: They retrieved about the same
number of text paragraphs across the different questions. In the final interview,
four of the unsuccessful learners said that they could not imagine the facts in the
last two questions very well and that they had no clear idea for what information
they should search. In contrast, the successful learners showed a rather marked
profile: The frequency of information access clearly increased with Ql, Q3, Q5,
and Q7. The test for significance of the respective differences between the groups
resulted in p<.05, p<.10, p<.10, and p<.05, respectively (Mann-Whitney). No
one among the successful learners commented on difficulties in knowing what
information should be sought.
352
Wolfgang Schnotz
10
8
6
4
/
<,
''0--- --d'
O+---+---t------1f----+--+---+---i
02
-0 3
01
04
05
o 6 07
Fig. 14.5. Pattern of the average number of times that information was
accessed by successful and unsuccessful learners across questions Ql through Q7
Interpretation
The results just presented suggest that the two groups used different strategies to
answer the learning questions and, accordingly, interacted in different ways with
the hypertext system. The successful learners obviously used a model-based
strategy from the beginning. Because they first had to perform the initial
construction of a mental model, they retrieved much information and spent much
time at this point. Later, they concentrated their information access at those
points where an elaboration of the model structure was required to answer the
questions, that is, at Q3 (where the date was questioned the first time), at Q5
(where crossing the dateline was to be considered the first time), and at Q7 (where
both the dateline and the midnightline were to be accounted for). The successful
learners combined model construction, model analysis , and symbol processing in
a flexible way. Inconsistencies within the subjects ' knowledge were usually
solved by elaborating the respective model structure.
In contrast, at the beginning, the unsuccessful learners tried to get along with a
pure symbol-processing strategy and tackled the model construction only in the
last questions . The respective model structures were obviously not well
elaborated. Inconsistencies within the subjects' knowledge were solved by means
of ad hoc repairs of procedures rather than by a systematic elaboration of the
model structure . Because answering of questions with mere symbol processing
requires relatively little information, especially if learners mainly follow the
available examples of how to calculate time differences, those subjects showed a
lower and rather flat pattern of information access. Yet, due to the cognitively
demanding initial construction of a mental model, one would expect increased
353
information access as soon as the learner begins to construct a mental model. This
was not the case, however, as can be seen from Figure 5: The frequency of
information access was not remarkably higher in the last two questions than it had
been before. This pattern was possibly the result of difficulties in specifying the
aim of information search, because, as mentioned earlier, the unsuccessful learners
had no clear idea what information they should search for with these questions.
Discussion
Learners with more prior knowledge are usually better able to find and process
information that is relevant for further learning than are learners with less prior
knowledge (O'Donnell, 1993). In the present study, however, no such effect could
be observed; there was hardly any difference in prior knowledge between successful
and unsuccessful learners. Most subjects remarked that they had not yet thought in
a systematic way about this learning topic, so they could be considered novices .
Clear differences were found, however, concerning the strategies used in answering
the learning questions: The successful learners employed a model-based strategy
from the outset, whereas the unsuccessful learners tried first to get by with a
symbol-processing strategy and tackled the construction of a mental model only
during the last question . These different strategies were also reflected by different
patterns of information access among successful and unsuccessful learners.
These strategy differences can be related to a distinction between a local and a
global orientation of knowledge acquisition (cf. Hacker, 1986). With a local
orientation, the individual reacts ad hoc on the topical demands, whereas with a
global orientation , the attempt is made to systematically construct and elaborate
the respective knowledge structure in order to be useful also for other known or
anticipated tasks. Accordingly, the unsuccessful learners in the preceding study had
a local orientation in that they restricted their information access to the topically
relevant numerical-algorithmic information. The tendency to solve inconsistencies
by means of ad hoc repairs of procedures, instead of by elaborating the model
structure, can also be interpreted as an indication of a local orientation in solving
learning tasks. In contrast, from the outset, the successful learners in this study
were rather globally oriented in the strategies used to solve the learning tasks and
in terms of their information access strategies. At the beginning, they retrieved
much information for the initial construction of a mental model, which was then
elaborated step by step according to the respective task demands.
As mentioned earlier, the unsuccessful learners frequently reported that they did
not know exactly for which information they should search, especially in the last
questions. This corresponds to the finding that search for information is usually
rather difficult with complex search goals (Armbruster & Armstrong, 1993).
Because those subjects also reported having difficulties in imagining the facts, and
354
Wolfgang Schnotz
because their learning behavior indicated that they delayed tackling the
construction of a mental model, one can speculate that mental models constructed
by learners also function to specify the information required for further learning
(cf. Mayer, 1989). Accordingly, mental models would be not only the result of a
retrospective integration of previous information with topical new information
but also the basis for a prospective search for further information to continue the
knowledge acquisition process . Further research will be necessary to clarify this
point.
Instructional Implications
What practical consequences can be drawn from these considerations? Because the
acquisition of knowledge is usually embedded in and triggered by the use of
knowledge, the respective learning processes can be fostered not only by an
adequate design of hypertext systems, but also, in order to stimulate learners to be
active builders of their own knowledge, by selecting adequate learning tasks and
by bringing these tasks into an appropriate sequence.
Selection of learning tasks. If one follows the considerations mentioned earlier,
learning tasks are to be selected in such a way that learners are prompted to
construct mental models on the basis of propositional representations and to
systematically relate both forms of representation to each other. For this purpose,
questions can be asked that serve to localize entities correctly within the model
and that afford a determination of relations between these entities in the model
structure. With regard to the learning content of this study, such questions could
concern, for example, where the dateline is, whether it is earlier or later on its
western side than on its eastern side, where on earth (measured according to the
sun 's position) the time is 6:00 am and where it is 6:00 pm. Learners should be
forced to operate with their models in such a way that cognitive conflicts would
emerge if the achieved model structure is not yet adequate. Furthermore, they
should be prevented from utilizing ad hoc repairs in case of cognitive conflicts
without considering the consistency of their knowledge structure. Learners should,
for example, be encouraged to compare the topical solution with the solutions of
previous tasks and to check them for compatibility.
Sequencing of learning tasks. The selected learning tasks have to be brought
into a certain sequence, which prompts learners to successively elaborate their
knowledge and to use this knowledge in different ways. For this purpose, one can
follow the concept of a successive model progression (cf. White & Frederiksen,
1990). The first step is to determine for each task the model structure that is
required to solve the task with a model-based strategy. The second step is to place
those task-specific model structures in a sequence of increasing complexity. The
result is not only a sequence of model structures but also a sequence of learning
355
tasks, which should motivate the learner to move from one model structure to the
next, more complex structure.
Display of information. The concept of a successive model progression also
provides some indication about when learners need which kind of information to
proceed from one model structure to the next. In view of a possible cognitive
overload of the learner, in the case of unrestricted access to a very broad set of
information, it would be sensible to restrict the accessible information for
learning with hypertext in a task-specific way. According to the preceding
considerations, however, the information offer also has to be large enough that it
is possible for learners not only to gather information from the local view of the
momentary learning task, but also to have the opportunity to perform a more
globally oriented search for information, having all the information available that
has been necessary so far to construct a model on the task-specific elaboration
level. In order both to restrict the available information to a reasonable extent and
simultaneously to make this information offer broad enough, it would be possible
to use specific display techniques as, for example, so-calledfisheye views (Furnas,
1986). A fisheye view is a kind of presentation that displays (analogous to a lens
with a very wide angle or fisheye) nearby objects in more detail and the periphery
in less detail. In this way, the topical part of the information base could be
foregrounded, whereas other parts remain in the background although they are still
accessible to the learner. To date, very little is known about the cognitive effects
of such hypertext -specific types of staging (cf. Meyer, 1975).
Knowledge acquisition from hypertext is a complex research area that has
enlarged the traditional scope of text processing research by a number of additional
problems. It would be too simplistic to look only for positive or negative effects
of this new medium on learning outcomes. Instead, process-oriented studies are
required that analyze how different learners interact with hypertexts, which
strategies they use, and which mental representations they construct. Besides
cognitive investigations, affective and motivational aspects should also be
analyzed in order to achieve a comprehensive view of knowledge acquisition from
hypertext. A better understanding of these processes will gradually enable us to
design the respective learning environments in a way that takes into account the
psychological principles of knowledge acquisition as far as possible.
Acknowledgments
The author gratefully acknowledges the support of Emmanuel Picard in
developing the hypertext sytem used in this study. He also wants to thank Roger
Saljo and Barbara Burge for their comments on a former version of this article.
356
Wolfgang Schnotz
References
Armbruster, B. B., & Armstrong, J. O. (1993) . Locating information in text: A focus
on children in the elementary grades. Contemporary Educational Psychology, 18(2),
139-161.
Baddeley, A (1992). Working memory : An introduction. International Journal of
Psychology, 27(3/4), 96 .
Baker, L. (1985). How do we know when we don't understand? Standards for ev aluating
text comprehension. In D. L. Forrest -Pressley , G. E. MacKinnon, & T. G. Waller
(Eds.), Metacogn ition, cognition, and human performance. Vol. I: Theoreti cal
perspectives (pp. 155-205). Orlando, FL: Academic Press .
Brown , J. S., Burton, R. R., & de Kleer, J. (1982). Knowledge engineering and pedagogical techniques in SOPHIE I, II, and ill . In D. Sleeman & J. S. Brown (Eds.) ,
Intelligent tutoring systems (pp. 227-282) . London : Academic Press .
Brown, J. S., Collins, A, & Duguid, P. (1989) . Situated cognition and the culture of
learning. Educational Researcher, 18(1), 32-42 .
Brown, P. J. (1990). Assessing the quality of hypertext documents . In A Rizk , N.
Streitz, & J. Andre (Eds.), Hypertext: Concepts, systems , and applications (pp . 112). Cambridge, England : Cambridge University Press.
Carey , S. (1986) . Cognitive science and science education . American Psychologist,
41, 1123-1130.
Carmel, E., Crawford, S. F., & Chen, H. (1992) . Browsing in hypertext: A cognitive
study . IEEE-Transactions on Systems, Man, and Cybernetics, 22(5), 865-884 .
Chafe, W. L. (1979). The flow of thought and the flow of language . In T. Givan (Ed.),
Syntax and semantics, Vol. XII: Discourse and syntax (pp. 159-181). New York:
Academic Press .
Chi, M. T. H., Glaser, R., & Rees, E. (1982) . Expertise in problem solving . In R. J .
Sternberg (Ed.), Advances in the psychology of human intelligence (Vol. I, pp. 775). Hillsdale, NJ: Erlbaum .
Conklin , J. (1987) . Hypertext: An introduction and survey. IEEE Computer, 20(9), 1741.
De Young, L. (1990) . Linking considered harmful. In A Rizk, N. Streitz, & J. Andre
(Eds.), Hypertext: Concepts, systems, and appli cations (pp. 238-249). Cambridge,
diSessa, A (1983) . Phenomenology and the evolution of intuition . In D. Gentner & A.
L. Stevens (Eds .), Mental models (pp. 15-33). Hillsdale, NJ: Erlbaum.
Furnas, G. W. (1986, April). Generalized fisheye views . CHI Proceedings, 16-23.
Garner , R., & Reis, R. (1981) . Monitoring and resolving comprehension obstacles :
An investigation of spontaneous text lookbacks among upper-grade good and poor
comprehenders. Reading Research Quarterly, 26, 569-582 .
Garrod, S. C. (1985) . Incremental pragm atic interpretation versus occasional
inferencing during fluent reading. In G. Rickheit & H. Strohner (Eds.), Inferences in
text processing (pp. 161-181) . Amsterdam: North-Holland.
Glenberg, A M., Wilkinson, A. C., & Epstein, W. (1982) . The illusion of knowing :
Failure in the self-assessment of comprehension. Memory & Cognition, 10, 597602 .
357
Gloor , P. A. , & Streitz, N. A. (Eds.). (1990) . Hypertext und hypermedia. Berlin:

Springer.
Gray, S. H. (1990) . Using protocol analyses and drawings to study mental model
construction during hypertext navigation. International Journal of Human Computer
Interaction, 2(4) , 359-378 .
Guthrie, J. T., & Mosenthal, P. (1987). Literacy as multidimensional: Locating
information and reading comprehension. Educational Psycholog ist, 22(3&4), 279297 .
Guthrie, J. T ., Britten, T., & Barker, K. G. (1991 , April). Effects of document and
strategy type on cognitive components on search. Paper presented at the American
Educational Resear ch Association Convention , Boston , MA.
Hacker, W. (1986) . Arbeitspsychologie. Psychische Regulation von Arbe itstatigkeiten. Berlin : VEB Deutscher Verlag der Wissenschaften.
Johnson-Laird, P. N. (1983) . Mental models. Towards a cognitive science of language ,
inference, and consciousness. Cambridge, England: Cambridge University Press .
Johnson-Laird, P. N. (1987) . The comprehension of discourse and mental models . In
M. Nagao (Ed.), Language and artificial intelligence (pp. 253-261). Amsterdam :
North-Holland .
Jonassen, D. H., & Grabinger, R. S. (1990) . Problems and issues in designing
hypertext/hypermedia for learning . In D. H. Jonassen & H. Mandl (Eds.), Designing
hypermedia for learning (pp. 3-25) . Berlin: Springer.
Jonassen, D. H., & Mandl, H. (Eds.). (1990). Designing hypermedia f or learning.
NATO A5I Series F, Vol. 67. Berlin: Springer.
Jonassen, D. H., & Wang, S. (1993) . Acquiring structural knowledge from
semantically structured hypertext. Journal of Computer-Based Instruction, 20(1), 1-
8.
Kobasigawa, A. (1993) . Children 's retrieval skills for school learning. The Alberta
Journal of Educational Research, 29, 259-271.
Kosslyn, S. M. (1981) . The medium and the message in mental imagery : A theory .
Psychologi cal Review , 88, 46-66.
Lucarella, D. (1990). A model for hypertext-based information retrieval. In A. Rizk, N.
Streitz, & J. Andre (Eds.), Hypertext: Concepts, systems , and appli cations (pp. 8194). Cambridge, England : Cambridge University Press.
Marton , F., & Saljo, R. (1976). On qualitative differences in learning : II. Outcome as a
function of the learner's conception of the task. British Journal of Educational
Psychology, 46, 115-127.
Mayer, R. E. (1989) . Models for understanding. Review of Educational Research,
59(1), 43-64 .
McAleese, R. (Ed.). (1989) . Hypertext: Theory into practice. Worcester, England:
Billing & Sons .
McAleese, R., & Green, C. (Eds.). (1990). Hypertext: State of the art. Oxford, England :
Intellect.
McGrath, D. (1992) . Hypertext, CAl, paper or program control: Do learners benefit
from choices? Journal of Research on Computing in Education, 24(4), 513-532 .
Melton, R. F. (1978) . Resolution of conflicting claims concerning the effect of behavioral objectives on student learning . Review of Educational Research, 48, 291302 .
358
Wolfgang Schnotz
Meyer, B . J. F. (1975) . The organization of prose and its effec ts on memory.

Amsterdam: North-Holland.
Mohageg , M. F. (1992) . The influence of hypertext linking structures on the efficacy
of information retrieval. Human-factors, 34(3), 351-367.
Morrow, D. G., Green span, S. L., & Bower , G. H. (1987). Accessibility and situation
models in narr ative comprehension. Journal of Memory and Language, 26, 165-187.
O 'D onnell, A (1993 ). Searching for information in knowledg e maps and texts.
Contemp orary Educational Psychology, 18(2), 222 -239 .
Palmer , S. E. (1978). Fundamental aspects of cognitive repr esent ation . In E. Rosch &
B. B. Lloyd (Eds.) , Cognition and catego rization (pp. 259-303) . Hillsdale, NJ :
Erlb aum.
Pichert, J. W ., & Anderson, R. C. (1977) . Taking different perspectives on a stor y .
Journal of Educational Psychology, 69,309-315.
Pinker, S. (1990) . A theory of graph comprehension. In R. Freedle (Ed.), Artificial in telligen ce and the future of testing (pp. 73-126) . Hillsdale, NJ: Erlb aum .
Rizk , A , Streitz, N., & Andre, J. (Eds.) . (1990) . Hypert ext: Concepts, systems. and
appli cations . Cambridge, England: Cambridge Univer sity Press.
Sam arapungavan, A, & Beishuizen , J. (1992). Hyperm edia and knowled ge acquisitio n
from non-linear expository text. In B. van Hout-Wolters & W. Schnotz (Eds.), Text
comp rehension and learning from text (pp. 53-69). Amsterdam : Swets & Zeitlinger.
Saxer, K. H., & Gloor, P. A (1990). Navigati on im Hyperraum: Fisheye views in
HyperCard. In P. A Gloor & N. A. Streitz (Eds.), Hypertext und Hypermedia (p p.
190 -204 ). Berlin: Sprin ger.
Smith, J. B., & Weiss, S. (1988) . Hypertext. Communications of the ACM, 3 1(7 ),
8 16- 8 19 .
Spiro, R. J., Feltovich , R. L., Jacobson, M. J., & Coulson, R. L. (1991). Cognitive
flexibility, constructivism, and hypertext: Random access instru ction for advanced
knowledge acquisition in ill-structured domains . Educational Technology, 31(5), 2433 .
van Dijk, T. A , & Kintsch , W. (1983). Strategies of discourse comp rehension. New
York: Academic Press.
Vo sniadou , S. , & Brewer, W. F. (1990). A cross-cultural investigation of children's
conceptions about the earth, the sun, and the moon : Greek and American data. In H .
Mandl, E. de Corte, S. N. Bennett, & H. F. Friedrich (Eds.), Learning & instruction:
European research in an international context, Vol II :2 Analysis of complex skills
and complex knowledge domains (pp. 605-629). Oxford, England: Pergamon.
White, B . Y ., & Frederiksen, J. R. (1990). Causal model progressions as a foundation
for intelligent learning environments. Artificial Intelligence, 42 , 99-157.
Yankelovich, N., Landow , G., & Cody, D. (1987). Creating hypermedia materials for
English literature students . ACM SIGCVE Outlook, 19, 12-15 .
Part Four
Accountable Talk: Learning to Reason
Chapter 15
Talking About Reasoning: How Important

Is the Peer in Peer Collaboration?
Colaboratory for Research on Electronic Work (CREW), University of Michigan,
Ann Arbor, MI, USA
Abstract
Using data from two studies of scientific reasoning , this chapter explores whether
transactive discussion is the basis of productive peer collaborations and questions
what role the partner plays in the apparent effectiveness of this type of discussion. In
the first study, dyads who engaged in transactive discussion showed more
improvement than dyads who did not have transactive discussions . In the second
study, both dyads and children working alone showed improvement related to talk in
general. However, dyads produced more transactive types of talk and showed a more
complex understanding of the problem that they generated more quickly . Having a
partner was not a necessary or sufficient condition for producing transactive talk but
increased likelihood that it would occur. The data from these studies suggest that the
value of peer collaborations may be that the presence of a partner provides a natural
context for elaborating one's own reasoning.
The past ten years of research on children's collaborations have demonstrated that
children who work with peers can show improved reasoning and problem solving .
Although not all collaborations are successful, a growing number of studies have
pointed to the conditions under which children who work with peers learn more
than children who work alone. The earliest studies demonstrated that the degree
of mutuality is an important factor in determining the success of the
362
collaboration. Specifically, children learn more when they actively participate in

the collaboration (e.g., Gauvain & Rogoff, 1989). Using increasingly
sophisticated analyses of children's interactions, most recent studies stress the
importance of the communicative exchanges that constitute mutual participation.
Most of the current explanations of successful collaboration emphasize that the
communication between partners is responsible for learning (Ellis & Gauvain,
1992). In this chapter, I present data from two studies to explore the role that a
particular type of communication, called transactive discussion, plays in peer
collaboration and to question what role the partner plays in the apparent
effectiveness of this type of discussion . In particular, I will explore the
effectiveness of transactive discussion as it applies to children who are
collaborating on computer-presented scientific reasoning tasks.
Transactive Discussion
With the growing emphasis on the communication between collaborating

partners, the question has now turned to asking which kinds of talk lead to
learning. Despite differences in the coding systems used by various researchers,
discussions that can be generally characterized as transactive seem to lead to the
most productive collaborative outcomes. Transactive discussion, a term first
adopted by Berkowitz and Gibbs (1983), refers to a type of interaction in which
each child uses his or her own conversational turn to operate on the reasoning of
the partner or to clarify his or her own ideas. The display of reasoning
distinguishes transactive statements in conversational turns. Specifically, a turn is
considered transactive when it extends, paraphrases, refines, completes, or
critiques the partner's reasoning or the speaker's own reasoning. For example, in
a discussion about the time of day, a turn would not be considered transactive if
only an assertion was made (e.g., It's after 6 p.m .), unless there was also some
display of reasoning (e.g., I know it's after 6 p.m . because the sun is low on the
horizon). Table 1 lists various types of transactive statements and a definition of
each type of transact. Note that transacts can occur in the form of spontaneously
produced statements or as questions and resulting answers. Each type of transact
can also be coded as self-oriented (operates on the speaker's own reasoning) or
other-oriented (operates on the reasoning of a partner).
The theoretical link between transactive reasoning and cognitive change is
based on Piaget's (1932/1965) proposal that, when children operate on each
other's reasoning, they become aware of contradictions between their own
reasoning and that of their partner. Resolution of the cognitive conflict
(disequilibrium to equilibrium) provides the basis of a newer, higher level of
understanding that signifies cognitive growth. Piaget believed that the relative
equality of peers makes it likely that children will reason with each other and
integrate conflicting viewpoints. A study by Kruger and Tomasello (1986)
Talking About Reasoning
363
confirmed that children produced more transactive reasoning when working with
peers than they did with adults.
Table 15.1. Types of Transactive Statements
TYPE
Feedback Request
Paraphrase
Justifi cation Request
Juxtaposition
Completion
Clarification
Refinement
Extension
Critique
Integration
IRANSACTIYE MOVE
Do you understand or agree with my position?
I can understand and paraphrase your position or reasoning.
Why do you say that?
Your position is X, and my position is Y.
I can complete or continue your unfinished reasoning.
No, what I am trying to say is the following .
I can elaborate or qualify my position to defend against your
critique .
Here is a further thought or elaboration.
Your reasoning misses an important distinction, or involves a
questionable assumption .
We can combine our positions into a common view.
Note. From "Measuring the development of features of moral discussion" by M. W .

Berkowitz and J. C. Gibbs, 1983, Merrill -Palmer Quarterly, 29, pp . 399-410. Copyright
1983 by Wayne State Univer sity Press. Adapted by permission.
Although initial proposals about transactive reasoning were founded on Piagetian

theory, in practice today, this perspective is not incommensurate with other
theoretical positions. For example, although transacts can be based on conflict
between partners in such forms as Justification Request and Clarification, they
can also be based on agreements such as in Paraphrase, Completion, and
Integration. Transacts that involve agreement more closely resemble the process
of coconstruction described by Damon (Damon & Killen, 1982; Damon &
Phelps, 1989) than the interpersonal conflict typically associated with Piaget's
theory (see Berkowitz, 1985). In addition, transactive statements can operate on
the speaker's own reasoning (Self-Oriented Transact) as well as on the reasoning
of a partner (Other-Oriented Transact), although other-oriented transacts are
proposed to be more effective than self-oriented transacts (e.g., Kruger, 1992).
Given the increasing interest in Vygotskian theory, especially the concept of
intersubjectivity, coding transactive reasoning can also be used as a tool to
capture how and when interacting partners enter into the same here and now
(Giles & St. Clair, 1985) when they work together on problem-solving tasks.
In support of the hypothesis that transactive reasoning is the basis of productive
peer collaborations, several recent studies have shown that children who engage
in interactions characterized as transactive learn more from the collaboration than
364
those who do not. This evidence comes from studies of children collaborating on
tasks in various domains . In the domain of moral reasoning , several studies have
found that children show higher levels of moral judgment when they engage in
discussions that act on their own and their partner's reasoning (Kruger &
Tomasello, 1986; Kruger, 1992). Damon and Phelps (1989) have provided
evidence that dyads learn more from discovery and reflection tasks (e.g.,
proportional reasoning, spatial perspectives) when they engage in mutually
responsive dialogue in which ideas about the logic of the task are shared . Tudge
(1992) has shown that children benefit from collaboration when they shared their
reasoning about solutions for a mathematical balance beam task. Thus , these
studies on moral, conceptual, and mathematical reasoning suggest that peer
interactions can support learning, especially when children use trans active forms
of discussion to communicate ideas.
Transactive Discussion and Scientific Reasoning

If trans active types of communication are the key to successful collaborations,
scientific reasoning provides an excellent domain in which to investigate this
claim and to extend our understanding of the value of the collaborative process.
Various factors related to improved reasoning are decidedly dialogic: interpreting,
explaining, and justifying experimental outcomes appropriately. In theory,
transactive discussions should be particularly effective in scientific reasoning
tasks, because this type of discussion forms the basis for the epistemic actions
(Pontecorvo & Girardet, 1993) relevant to this domain . For example, work by
Kuhn and her colleagues (e.g., Kuhn , Amsel, & O'Laughlin, 1988) has
demonstrated that careful interpretation of evidence is critical for successful
reasoning. Similarly, Klahr and his colleagues (e.g., Dunbar & Klahr, 1989;
Klahr, Fay, & Dunbar, 1993) have shown that constructing and revising coherent
explanations of the data are an essential part of the reasoning process . Because
transactive discussions specifically act on reasoning, this type of discussion
should have a significant effect on children's success with scientific reasoning
tasks by supporting the types of activity that Kuhn, Klahr, and others have shown
to be crucial to successful experimentation.
The only evidence to date that supports the relationship between trans active
discussion and scientific reasoning comes from a recent study by Azmitia and
Montgomery (1993). In this study, fifth graders were asked to work with partners
to figure out which variables were poisonous in two problem types: variables that
made plants sick or pizza ingredients that caused death (a task adapted from Kuhn
& Brannock, 1977). In both types of problems, the dyads were shown several
outcomes (i.e., healthy vs. sick plants, living vs. deceased people) and a set of
variables that were related to each outcome (i.e., leaf lotion, soil type, or
mushrooms, sausage). Dyads were videotaped as they worked together for one
18-35 minute session to isolate which variables led to negative outcomes. The
365
results showed that evaluating problem outcomes and participating in transactive

conflicts were significantly associated with better problem solving . That is, dyads
who monitored their partner's reasoning and engaged in conflicts critiquing their
partner's reasoning were more likely to identify correctly the causal status of
variables in a subsequent posttest. Transactive conflict was typified in Example 1.
(The specific codes of turns considered transactive are listed in parentheses as T =
TYPE OF TRANSACT.)
Example I
Cl : So it has to be onions and green olives .

C2: Yeah, onions but it can't be green olives. See, this dead
one doesn't have it. (points to dead one). (T = CRITIQUE)
Cl : That doesn't matter, it (green olives) just has to be in
the dead one but not in the alive . See, no green olives,
no green olives, no green olives (points to alive). (T =
CRITIQUE)
C2: But it has to be in all of them (the dead) . Otherwise,
they wouldn't all die . What killed this one then? (2Ts,
CRITIQUE and TRANSACTIVE QUESTION)
Cl The onions, see, they (dead) all have onions . (T =
TRANSACTIVE RESPONSE)
C2: Well, it's just onions. I'm writing just onions.
Cl: And green olives. Onions AND (enphasizes) green olives .
C2: Let's just go on to the next one .
Cl : OK. (Whispers) But it was onions and green olives.
(Azmita & Montgomery, 1993, p. 221)
Note that, although the dispute ended, there was apparently no agreement
between the dyad partners on the causal status of the green olives. In this case,
even though the partners were acting on each other's reasoning, consensus was
not the result, and Child 2 maintained her own position. Although these children
experienced a conflict between positions, at least one member of the dyad did not
end up with a higher level of understanding. However, it is clear that both
partners understood each other's position and ended the discussion, even though
they did not agree.
Although the Azmitia and Montgomery (1993) study generally supports the
claim that trans active discussion is related to successful collaborative outcomes,
there were two differences in the results of their study compared with previous
studies of transactive reasoning. First, not all types of transactive discussion (e.g.,
transactive questions and responses) were correlated with better task
performance. The authors note that the overall trans active properties of the
discussions generated by their scientific reasoning task were best captured in the
366
coding system as trans active conflicts; other individual categories occurred

infrequently. This difference suggests that we should not overlook the obvious:
The occurrence of particular types of transactive statements may be more closely
tied to the nature of the task than to the effectiveness of the collaborators. Second,
the association between task accuracy and other-oriented transacts was not
significantly higher than between task accuracy and self-oriented transacts. These
data suggest that operating on a partner's reasoning may not be more
advantageous than operating on one's own reasoning. This result calls into
question whether the value of transactive discussion is as an interpersonal process
or an intrapersonal process.
Transactive Reasoning in Two Microworlds

In a series of studies, my colleagues and I (Teasley, 1995; Teasley & Resnick,
1993; Teasley & Roschelle, 1993) have been studying children's peer
collaboration using computer-presented scientific reasoning tasks. The data from
two studies are reexamined here to illustrate the importance of transactive
reasoning in collaborations and to suggest that the benefits of this type of
discussion may not necessarily be dependent on the reasoning of a partner.
In both studies described next, children were working with computer
microworlds where the task was to determine how some aspect of the microworld
operated. Although the microworlds were different, one involving racecars and
the other spaceships, the structure of the two tasks was the same: Children
designed a series of experiments to generate information about the microworld,
and then they evaluated the outcome of these experiments to formulate a
hypothesis.
Reasoning About Racecars

Using a computer microworld about racecars, called Daytona (see Schauble,
1990), 48 fourth- and fifth-grade children were asked to work with same-sex
partners to determine the causal status of five features of racecars (Teasley &
Resnick, 1993). Specifically, these children were asked to work together to figure
out which features (Le., engine size, wheel size, muffler, tailfin, color) affected
the cars' speed and which did not. Children explored the casual structure of the
microworld by engaging in a cycle of experimentation wherein they designed cars
with varied features, made predictions about the cars' speed, and ran roadtests of
the cars. Each road test could include one, two, or three cars. Speed was
367
represented in the microworld by how far the cars could travel in a standard
period of time . Children's predictions about speed were indicated by their
selection of one of several distance markers located along the racetrack.
In a study of children working alone, Schauble (1990 ) found that the students
who came to a more complete understanding of the micro world were those who
evaluated both the patterns of evidence they constructed and their own changing
theories. By using Schauble's racecar task in a collaborative context, we were
able to specifically examine whether children engaged each other in trans active
discussions about theories and evidence and the extent to which this type of talk
was related to theory change.
Children's understanding of the causal status of car features was measured by
their ability to predict accurately the speed of the cars . Because the children were
asked to say how fast each car would go before the cars ran along the racetrack,
we could determine how well each dyad unders tood the combined effects of any
particular set of car features. We looked for improvement in predictions over the
course of the 20 minute session to see which dyads were using their accumu lating
evidence to revise their theories about the feat ures. Of 24 dyads, 15 dyads
improved in their ability to predict the speed of any individual car . Correlational
analyses showed that improved prediction accuracy was not associated with the
number of cars constructed or the number of roadtests run. In addition , the dyads
who showed improvement in prediction accuracy did not construct more cars or
run more roadtests than did the dyads who did not improve.
Since the activity with the microworld itself did not determine which children
would better understand the racecar features , what did? Analyses of the children's
protocols showed that the verbal activity of the improvers and nonimprovers was
characteristically different. As will be seen, the dyads who engaged in transactive
discussion during more of their experiments were those who improved in
prediction accuracy.
Following the method of Berkowitz and Gibbs (1983) , I coded the presence or
absence of transactive discussion (not the number of turns codes as transactive).
Unlike Berkowitz and Gibbs, however, I considered whether or not transactive
discussion occurred in two phases of experimentation: evidence generation and
evidence evaluation . Specifically, discussions were coded as transactive if they
indicated reasoning about how to generate evidence (e.g., indicating why a
particular feature was chosen, how to test the effect of any feature) or evaluate
evidence (e.g., explaining the racecar's speed, stating a hypothesis about causal
effects of racecar features) . Discussions that were simply descriptive of the
child's activity during evidence generation (e.g ., how to operate the computer,
what commands were being entered) or of the racecar's behavior during evidence
evaluation (e.g ., merely describing the roadtest outcome without assessment or
speculation about why the car was fast or slow) were not coded as transactive.
Then, for each phase of experimentation considered separately, I calculated the
frequency of transactive reasoning on a roadtest-by-roadtest basis (as proportion
of roadtests with discussion coded as transactive). In doing so, I assessed whether
368
the frequency and timing - not total volume - of transactive reasoning are
important.
Transactive Discussion and Evidence Generation
Dyads who improved in prediction accuracy engaged in trans active discussion

during more of the roadtests that they were constructing than did dyads who did
not improve. To illustrate transactive discussion, consider the exchange found in
Example 2:
Example 2
Cl : Put on fins.
C2: It seems like when we put no fins on it, it
Cl:
Yeah
goes farther .
C2:
Cl:
than
when it does have a fin .
C2: Okay .
Cl: So let's try it with those same features that we had on
that one [previous car], except let 's put a fin on it
next time .
C2: Okay.
Cl : See how it, if it goes the same thing.
This discussion was coded as trans active because the partners were discussing
their reasoning about how to construct the roadtest. In fact, these partners were
talking about why they were selecting particular car features (the tailfin), as well
as their overall plan for the roadtest (to see if adding the tailfin affects the
outcome compared with a previous car configuration). Contrast the discussion
just presented with that in Example 3:
Example 3
Cl : I'm going to have a blue car with a big engine and no

muffler . It'll go to flag three.
C2: And the next car is going to be blue, big engine, tail fin
on , muffler off, flag f our .
Cl : Press go .
369
In this discussion, neither child explained why the part icular features were
selected for each car; there were no explanations about why the car is expected to
travel a certain speed (the prediction flag chosen); and there was no rationale for
choosing to run these two cars in the same roadtest. Because the partners did not
display any reasoning, the evidence generation phase of this experiment was not
coded as transactive.
Analyses of the conversation of the dyads who improved in their ability to
predict speed showed that they engaged in trans active discussion during the
evidence generation phase of 36% of all roadtests. In contrast , the dyads who
showed no improvement in prediction accuracy engaged in transactive discussion
during the evidence generation phase of only 19% of their roadtests. This
difference was significant, t (22) =1.72, P < .05.
Engaging in transactive discussion during evidence generation was also related
to prediction accuracy, because these discussions usually involved a plan or
strategy for how to construct the roadtest in a meaningful way. These discussions
could lead to more systematic experimentation and therefore more intelligible
outcomes. Consider the discussion in Example 4:
Example 4
Cl:
Let's try to think of, like, for slow ones that color
didn ' t matter and both [cars] go to that [prediction
marker for slowest cars].
C2: Yeah .
Cl : If we leave this [the second car constructed] the same,
they will probably go to there.
C2: Yeah.
Cl: It will prove that color doesn't matter .
C2: Now we can prove that col or doesn' t matter ?
Cl : Yeah , because they'll both go the same .
C2: Oh. ...
In this discussion, Child 1 suggests how they might construct the roadtest with
two objectives: (a) making the cars go slowly and (b) testing to see if color
matters. Although Child 2 seems to agree on the proposed plan for the roadtest,
she does not understand the concept of a controlled test of variables (where all
features are held constant except one). Without Child l' s statement that leaving
the cars the same except for color will prove that color does not matter, Child 2
would not have been likely to understand the significance of this test or possibly
any subsequent controlled test of features. Thus transactive discussion served to
externalize each partner's logic and to make it available to the other partner for
agreement , dispute, or, in this example, clarification . That this dyad went on to
produce further controlled tests, sometimes suggested by Child 2, supplies further
370
evidence that both partners understood the nature of controlled tests and could
reason based on evidence from this type of roadtest.
Transactive Discussion and Evidence Evaluation
Children who improved in prediction accuracy engaged in transactive discussion

about roadtest outcomes significantly more often than did the children whose
predictions did not improve. During evidence evaluation, discussions were
considered to be transactive when explicit interpretations about the outcome of
the experiment were discussed between the dyad partners . These discussions
showed explicit reasoning about why the cars went as fast as they did, compared
with discussions that just noted how far each car went. Compare the two
discussions in Examples 5 and 6.
Example 5
Cl : I think the tailfins do matter because it puts on weight.

'Cause that one was exactly the same but it had
C2:
It had
Cl :
I t didn 't
have a muffler like down from the tailfins.
C2: Yeah. I think that the tailfins [matter) . ..
Cl : And the muffler doesn't [emphasis) matter .
C2: The tailfins do , because that one [the racecar] was
exactly the same except the color and the muffler , and
they had tailfins .
So I think the tail fins shouldn't be on the car because
it puts on more weight .
Example 6
Cl : ah, it's fast [first car) .

C2: Yeah .
Cl : This one's slow [second car) .
C2: Real slow.
The first discussion (Example 5) was considered transactive; the second

discussion (Example 6) was not because no reasoning about causal mechanisms
took place. Transactive discussions of roadtest outcomes characterized 60 % of
the roadtests of the dyads who improved in prediction accuracy. In contrast, the
dyads who did not improve in prediction accuracy engaged in trans active
discussion of roadtest outcomes in only 25% of all their roadtests. The frequency
of outcome discussions was significantly different between the two types of
371
dyads, t (22) = 3.27, p < .001. The outcome discussions of the non-improvers
were more likely to be simple descriptive statements about what the cars were
doing rather than reasoning about why the car had a particular speed.
Summary
The results of this study show that short-term learning and theory change in a
causal reasoning task vary as a function of the difference in the kind of verbal
activity children engaged in while they were collaborating. Dyads who were able
to integrate their evolving theories about specific racecar features into global
assessments of car speed showed improvement in prediction accuracy over the
course of the session . These dyads engaged in transactive discussion as they
generated and evaluated experimental evidence . These transactive discussions
helped children to follow what Schauble (1990) calls normative scientific
reasoning. However, simply having a partner was not enough. Not all of the
children in this study were able to coordinate individual theories into correct
predictions about the combined effects of those features. In addition, the dyads in
this study engaged in many of the problematic behaviors that Schauble (1990)
and others (e.g., Klahr et al., 1993; Kuhn et al., 1988) have found to be typical for
children working individually on reasoning tasks, such as generating
uninformative evidence and misinterpreting experimental outcomes.
In considering the possible ways that transactive discourse might support
learning in peer collaborations, it is important to notice that we are analyzing
discourse in dyads without acknowledging that it may be the talk itself (of any
kind) and not the partner that facilitates learning . In any study of collaboration,
the collaborative condition necessarily confounds two factors: having a partner
and talking aloud. We cannot conclude that a particular type of talk, such as
transactive discussion, is the important factor in collaboration until we investigate
whether producing talk while working alone also facilitates reasoning.
Reasoning About Spaceships

Using a computer microworld about a spaceship designed by Klahr and his
colleagues (see Klahr, Dunbar, & Fay, 1990; Klahr, Fay, & Dunbar, 1993), 70
fourth-grade children were asked to work alone or with a partner to determine
how a mystery key affected the behavior of the spaceship (Teasley, 1995). Half of
the children who worked alone were asked to talk aloud, and half of the children
in dyads were asked not to talk to each other while they worked. Using these four
experimental groups (Talk Alone, No-Talk Alone, Talk Dyad, No-Talk Dyad), I
assessed the effects of talking and having a partner as independent variables . I
372
was most interested in the performance of the two talk groups. Given the
literature on children's private speech (see Diaz & Berk, 1992), it seemed likely
that asking these children to produce verbal protocols as they worked alone could
improve performance. However, would the children who talked as they worked
alone engage in transactive reasoning (e.g., produce self-oriented transacts)?
Further, what are the comparative effects of transactive utterances with and
without a partner (a test of self-oriented versus other-oriented transacts)?
The spaceship micro world was very similar to the racecar microworld as
children reasoned about the causal status of unknown variables and ran
experiments to test their theories about these variables. In the spaceship
micro world , children were asked to determine the causal effect of only one
variable: the mystery key. Children were shown how five LOGO-like commands
affected the spaceship (e.g., forward, backward, right turn, left turn, fire) and then
asked to determine how the new mystery key worked. Unknown to the children,
the mystery key acted as a repeat key, making all commands entered before it
repeat. For example, FORWARD 2 FIRE 3 MYSTERY KEY 2 made the
spaceship repeat the FORWARD 2 and FIRE 3 commands in sequence two
additional times, for a total of three Forward-and-Fire sequences. Commands
entered in a program after the mystery key were executed after the repeated
sequence was complete. For example, FORWARD 2 MYSTERY KEY 2 FIRE 3
would make the spaceship go forward six times and then fire three times. Success
on the spaceship task involved identifying the correct function of the mystery
key. Children's understanding of the mystery key was assessed after they worked
with the microworld for one 20-minute session. Their understanding could be
very general (e.g., knowing that it made the program repeat) or more complex
(e.g., knowing how many times the program would repeat and that commands
after the mystery key would not repeat).
The Effects of Talk
Analyses of children 's hypotheses about the mystery key showed that talking
affected performance but having a partner did not. Children who talked aloud as
they worked, either alone or with a partner, came to a better understanding of the
mystery key than did children in either silent group. Furthermore, the amount of
speech was related to performance because the total number of utterances! was
significantly correlated with children's understanding of the mystery key, r (23) =
.49, P < .05. These data suggest that the production of talk of any kind was
associated with better hypotheses. Previous collaboration studies finding superior
IBecau se I was coding the talk produced by children working alone as well as by dyads , I
coded at the level of utterance rather than the conversational turn. The unit of utterance,
although roughly corresponding to a sentence, was not considered here in the strict
grammatical sense but as it commonly occurs in spoken English (see Schegloff, 1991).
373
performance in dyads (who talk) as compared with children working alone (who
do not talk) may have been measuring the benefits of talking and not the benefits
of having a partner. Given that talk improved performance, were children in both
the alone and dyad conditions producing transactive reasoning ? Was there an
added advantage for having a partner?
As in the racecar study, the dyads' discussions were coded separately for the
evidenc e generation phase and the evidence evaluation phase of experimentation
to capture the timing of transactive reasoning . Because the previous analysis
suggested that the volume of talk was important, transacts were coded on an
utterance-by-utterance basis. Utterances were coded as transactive if they
indicated reasoning about how to generate evidence (e.g., indicating why a
program was designed, how to test the effects of the mystery key) and evaluate
evidence (e.g., assessing why the spaceship moved in a certain way, stating a
hypothesis about how the mystery key works) . Utterances that were simply
descriptive of the child's activity or of the spaceship's behavior were not coded as
transactive .
I asked whether children who worked alone produced any transactive reasoning
and, if so, how it compared with the transactive reasoning of the children in
dyads . It was not surprising that most of the children who talked aloud as they
worked alone did not produce as much talk overall (transactive or otherwise) as
did the children with partners . Children working alone showed longer pauses and
greater variability in the amount they talked (see Diaz, 1992, for discussion of the
difficulty in eliciting verbalization from children working alone) . Although the
children working alone talked less than the dyads did, they produced utterances
coded as transactive.
Because alones talked less than dyads, I followed the standard convention of
analyzing the transactive utterances as a proportion of total talk (e.g., Kruger &
Tomasello, 1986). Given the total amount of talk produced by any individual or
dyad, this analysis calculated how many of those utterances could be considered
transactive. The analysis showed a general pattern in which dyads produced
significantly more transactive utterances than alones, t (21) = 3.22, P < .01,
whereas alones produced more descriptive utterances than dyads, t (21) = 4.29, P
< .001 (see Figure 1). Transactive utterances were positively correlated with
performance, r (23) = .57, P < .01, and descriptive utterances were negatively
correlated, r (23) = -.51, P < .05. These results suggest that having a partner may
make trans active statements more likely and that performance is likely to be
enhanced, regardless of whether the transacts are self-oriented or other-oriented.
374
100
Alones
. Oyads
90
80
8
~
70
60
::> SO
"50
....
40
30
I0
20
,~
10
,
,
,
,
'4l;
1..0'"
,,",
"
Interpretive
Descriptive
Fig. 15.1. Mean Proportion of Transactive and Descriptive Talk for Children Working
Alone and Children Working in Dyads,
Transactive Discussion and Evidence Generation

Consider Examples 7 and 8 of transcripts illustrating transactive talk during
evidence generation. The first transcript is from a dyad, and the second is from a
child working alone.
Example 7
C2:
Cl:
C2:
The question mark must. affect everything.

You think?
Yeah , maybe .
[pause]
Cl :
I know what we'll do, what we'll do is we'll go- fire

something, and we'll do question mark something and do
fire something and then we'll be able to see if they
both are
C2:
No ! you can't do both fires because it will
just go fire, you know.
Let's do
Cl : Can we do up something?
You know, fire, question, fire ?
Because if we do both fires and then we'll see if both of
them are doubling.
375
C2: Well, which one you want to do is- we want to put one ,
like turn or something .
Example 8
Cl :
Now, the repeat, the mystery key repeats the programs

you punched in, punched in .
Neat , this is cool .
Okay, now let 's see, I'm trying t o think of a new one .
You could do like seven ups and then one fire which is
eight , and if it does it sixteen times then you 'll know
it repeats because eight plus eight is sixteen .
And it matters how long you want it to repeat .
Like if you want to repeat it once, then it would be
sixteen , so.- no, repeat it once, it would be sixteen.
And then like if you wanted to push in, right, right
turn three.
And then you wanted to push, and you pushed in down
five , then it would be down five, that would be eight.
And if you repeated it twice, two times is sixteen by
[inaudible] .
Eight, eight, eight, two times, eight plus eight is
sixteen, s o it will do it twice, equals eight is,
sixteen .
So, cool , okay, let's see, hmm , now let's see.
In both transcripts, children show much talk about their reasoning. For this dyad
(Example 7), the discussion concerned the intelligibility of a program with fire
commands entered before and after the mystery key. Both children wanted to see
if all commands are affected by the Mystery Key, but they held different ideas
about how to test that hypothes is. In the child working alone (Example 8), the
transcript shows that she was reasoning about the effects of the Mystery Key by
calculating in advance how many commands will be executed.
In contrast to the two transcripts just presented, consider the next two excerpts
from children who showed no transactive utterances while they generated
evidence. The first example is from a dyad; the second is from a child working
alone.
Example 9
C2:
Cl :
C2:
Cl :
Can we do- okay. [takes mouse]

Eight .
Backwards two, left
Did you push mystery yet?
376
C2:
[keeps entering without answering]
Example 10
Cl :
Okay, now I 'll fire .

Oh, got to clear horne first.
Fire two, right two , down two , left two, up two .
Right there, and question one.
Okay, let's go .
In both transcripts, there was no talk about why commands were being entered or
how the spaceship might move. In the dyad (Example 9), Child 2 did not describe
all the commands he entered, so his partner was not sure when or if the mystery
key had even been selected. In the alone condition (Example 10), the child was
only narrating the sequence of commands being entered . Both examples contrast
sharply with the previous transcripts (Examples 7 and 8) in the amount of
reasoning displayed in the talk.
Transactive Discussion and Evidence Evaluation
In the evidence evaluation phase of experimentation, children who produced

transactive utterances also displayed significantly more reasoning than those
children who produced mostly descriptive talk. The transcripts shown in
Examples 11 and 12 illustrate these differences:
Example 11
[Program runs]
Cl : But it went a different way .

C2: It's weird .
Cl : Of course it did! Because , because
C2:
It's so weird!
Cl:
see, see, see like you know how when 1- here, if see, if
we were right here going up, and they were going up it'd
be that way .
But if we were here going up ,
C2:
Yeah .
Cl : you wouldn't have that. It ended right here, so like if
we ended
C2:
Yeah , but let's
Cl :
here , going up with, now going that .
C2:
Cl :
C2:
Cl :
C2:
Cl:
C2:
377
If it were right there, going up would be going up that

way .
Must have repeated itself.
See now it's going down .
' Ca u s e see look at it it is going down, itself is going
down
but
No, no wait , it repeats itself .
Maybe it repeats itself as many times as you pushed it
in .
Maybe!
Yes ! !
'Cause we only pushed it in once before.
And it did it.
Example 12
Cl :
It's just supposed to do a left turn, and it hasn't done

that.
Ummm . Well, let's see, there's a left.
Maybe I commanded it wrong .
[pause , she tries to stop it with mouse]
I must have commanded it wrong . [checking the written
program]
Did the fire, turn, the back eleven, and the question
mark eleven, fire fifteen .
It hasn't done left turn twelve, so it still didn't
finish its routine .
TWo, three, four , five, six, seven , eight, nine, ten .
Going over the pattern, still going going over the
pattern .
Must not have been left turn twelve , it's still going
the same pattern .
In both preceding examples, the children produced descriptive utterances about

the behavior of the spaceship, but they also came to some conclusions about why
that action occurred. The children in the dyad and the child working alone tracked
whether the spaceship movement was as expected (e.g., in Example 11, "If we
were here going up ... you wouldn't have that"; in Example 12, "It's just
supposed to do a left turn, and it hasn't done that.") and what conclusions could
be made from the surprising results (e.g., in Example 11, "Must have repeated
itself'; in Example 12, ".. . still going over the pattern"). What is also interesting
in the dyad's transcript is how the hypothesis is jointly constructed by the
partners. Child 1 described what was surprising and what they expected to
happen , whereas Child 2 verbalized the actual hypothesis ("Must have repeated
378
itself'; ".. . maybe it repeats itself as many times as you pushed it in"). Child 1
then produces evidence that he not only agrees but also applies that hypothesis to
explain the behavior of a previous experiment. The child working alone produces
fewer transactive statements but, nonetheless , makes a conclusion based on the
observed behavior of the spaceship . Compare the preceding transcripts with the
following two protocols in Examples 13 and 14:
Example 13
[Program runs]
Cl :
C2 :
Cl:
What do you think mystery does?

[shzugs]
What?
[pause]
C2:
[both write down hypothesis, no discussion]
Huh??
Example 14
[Program runs]
Cl : [hmming]
Okay, one, two, one, two .
Okay, it's done .
These two examples typify the talk when children were producing more
descriptive utterances than transacts. In the dyad (Example 13), although Child 1
asks a transactive question, Child 2 does not respond with a transact, and no
further transactive discussion ensues. Although the absence of trans active
utterances was more typical of the children working alone, as shown above in
Example 14, there were children in both conditions who produced few transactive
utterances .
Summary
Dyads were producing more transactive utterances in both the program generation
and evidence evaluation phases of their experiments than were the children who
worked alone. Having a partner was not a necessary or sufficient condition for
trans active reasoning, however, because some of the children in the alone
condition produced transactive talk, and some of the children in dyads produced
very little . The increased likelihood of transactive reasoning by dyads shown in
these data suggests that the value of the peer in collaborations may be that the
379
presence of a partner provides a natural context for elaborating one's own

reasoning.
Is there any reason to believe that hearing and/or acting on a partner's
reasoning will provide an extra advantage to verbalizing one 's own reasoning?
Analysis of children's hypotheses about the mystery key suggests that engaging
in transactive discussion with a partner led to more complex hypotheses. Recall
that children's understanding of the mystery key could be general (knowing that it
made the program repeat) or complex (knowing how many times the program
would repeat and that commands after the mystery key would not repeat) . More
of the dyads gave answers about the mystery key that included specific
information about how it worked . In contrast, very few of the children who talked
to themselves gave answers that indicated more than a generally correct idea
about the mystery key (see Table 2).
Table 15.2. Children's Understanding of the Mystery Key
Hypothesis Type
Incorrect
Condition
Talk Alones (n = 12) 3
Talk Dyads (n = 22) 2
Correct
Correct
General
Complex
11
These data suggest that having a partner to talk with helps children to refine and
elaborate their ideas. In addition to a deeper understanding of the mystery key,
children who talked with partners also showed a second advantage: They arrived
at a correct general hypothesis much more quickly than did children who talked
alone. After only four experiments, 75% of the children who talked with partners
gave a correct general hypothesis about the mystery key . In contrast, it took 11
experiments before 75% of the children who talked alone gave a correct general
hypothesis. It seems, therefore, that talking with a partner leads to more rapid
generation of correct general hypotheses that then get further developed as
children continue to work together.
Several factors may account for slower and less complex hypotheses generated
by children in the talk alone condition . First, various cognitive studies in which
individuals have been asked to produce think-aloud protocols have found that the
production of concurrent verbalization can slow problem-solving activity (e.g.,
Ericsson & Simon, 1980) . In addition, studies of children 's private speech have
shown that the beneficial effects of verbalization may not appear immediately
(e.g., Bivens & Berk, 1990). Because all children in this study were limited to a
380
twenty-minute session, it is possible that the talk alones may have needed more
time for the full benefits of their verbalization to be realized. If the session had
lasted longer or delayed posttest performance measures had been administered ,
the talk alones may have shown more positive effects from verbalization.
Conclusion
The data from both the racecar study and the spaceship study provide evidence to
support the hypothesis that transactive talk facilitates learning in scientific
reasoning tasks . Children in both studies who came to understand the causal
relations of variables in the two microworlds were those who talked about their
reasoning as they worked. Data from the spaceship study suggest that
collaborations lead to learning not because of the partner but rather because
having a partner increased the likelihood that children would produce transactive
statements. Transacts are particularly crucial for success on the reasoning tasks
used here, because previous research has documented how easily children are
distracted by surface features of experimental outcomes and how often they fail to
coordinate theories with evidence (Klahr et al., 1993; Schauble, 1990).
The answer to the question, "How important is the peer in peer collaboration? ",
may fundamentally rest on the interdependency of cognition and social relations
in communication. Built into communication is an implicit obligation to make
sense to one's conversational partner (Grice, 1989). When children are talking to
partners as they work on scientific reasoning tasks, the situation requires
coordinating conversational coherence to maintain and make sense of the ongoing
task activity (e.g., Teasley & Roschelle, 1993) . To simultaneously manage
coherence and activity, children working with partners produce transactive
utterances that make explicit their intentions, current interpretations, and
expectations. The production of these types of utterances engages children in the
epistemic operations that are crucial to scientific reasoning (see also Pontecorvo ,
1987) . However, because some of the children working alone in the spaceship
study also produced transactive statements, the necessity of having a partner is
questionable. As children talk aloud to themselve s, the social implications of
communication are maintained (Wertsch, 1985), although the conditions are not
as stringent, because the failure to be coherent has no immediate ramifications .
Several studies provide evidence that the mere anticipation of having a partner
can lead to cognitive gains. In two studies of cognitive tuning (Cohen, 1961;
Zajonc, 1960), subjects who expected to transmit information to others were
found to have more organized cognitive structures than did subjects who expected
to receive information. Similarly, studies investigating accountability have shown
that subjects who expected to defend and justify their opinions engaged in more
careful deliberation and showed more cognitive complexity than did subjects who
381
did not anticipate justifying their opinions (e.g., Davis, Stasser, Spitzer, & Holt,
1976; Tetlock, 1983). Finally, a developmental study of children's problem
solving also found that the anticipation of having a partner produced conservation
answers that were as highly rated as in a condition in which children actually had
to work out a solution (talk) with a partner (Doise, Rijsman, Van Meel, Bressers,
& Pinxten, 1981, cited in Light & Perret-Clermont, 1989).
These studies point to the conclusion that the anticipation of having a partner
can lead to the same benefits as actually having a partner. This finding is
consistent with data presented here and provides a possible explanation of why
some of the talk alone children in the spaceship study produced proportionally the
same amount of transactive talk as some of the talk dyads. These alones may have
been more aware that their talk would ultimately be heard by the experimenter
and, therefore, maintained the social obligation to be coherent and explicit about
their reasoning. That is, the alones with a high proportion of interpretive talk may
have perceived the situation as being more social than did the alones who
produced less interpretive talk, and adjusted their talk accordingly . In contrast,
talk alones who verbalized few transactive utterances may have perceived the
situation as being less social and, therefore, produced speech that resembles inner
dialog more than external social speech (Wertsch & Stone, 1985). If this
assumption is correct, any procedure that reproduces the social obligation for
coherence that is implicit in conversation should engage interpretive thinking
with or without actual vocalization.
An interesting question for future work is whether language is ever divorced
from the implicit social contract regarding coherence. To test such a hypothesis, it
would be necessary to observe a child engaging in conversation with a partner for
whom coherence is not necessary . Interacting with a computer partner (e.g.,
computer tutoring) is one of the few situations in which natural language is
produced without the presence of a human partner. Some empirical evidence
suggests that conversational norms are changed when language is mediated by
computers (e.g., Sproull & Keisler, 1991). If the computer is the conversational
partner rather than just a medium for communication, would these norms
disappear completely? For example, one can imagine that students may not feel
obligated to explain their thinking or to be explicit about their ideas in response to
computer-based prompts . In this case, the computer prompts would not provoke
the same amount of transactive reasoning as if the prompts were delivered by
human partners. Such evidence would more directly address the issues of whether
the demand for coherence is fundamental to producing transactive reasoning and
if this coherence is an integral part of communication regardless of the actual
presence of a collaborator.
382
Author Note
The studies reported in this chapter were conducted while I was at the Learning
Research and Development Center at the University of Pittsburgh. The work was
supported by funding from the Office of Educational Research and Improvement
(OERI) of the United States Department of Education and the Andrew W. Mellon
Foundation. I am grateful for the cooperation of the schools involved: St.
Bernadette in Monroeville, Pennsylvania, and Lake Hills in Ferrysburg,
Michigan . I am also grateful for the guidance and support of Lauren Resnick over
the course of these studies. Margarita Azmitia and Clotilde Pontecorvo
generously provided helpful comments on earlier drafts of this paper.
Correspondence concerning this article should be addressed to Stephanie D.
Teasley, Colaboratory for Research on Electronic Work (CREW), , University of
Michigan, Ann Arbor , MI, USA. Electronic mail may be sent via Internet to
steasley@umich.edu
383
References
Azmitia, M., & Montgomery, R. (1993) . Friendship, tran sactive dialogues, and the
development of scientific reasoning . Social Development, 2, 202-221.
Berkowitz, M. W. (Ed.). (1985). New directions for child development: Peer conflict and
psychological growth San Francisco: Jossey-Bass ,
Berkowitz, M. W., & Gibbs, J. C. (1983). Measuring the development of features of moral
discussion . Merrill -Palmer Quarterly, 29, 399-410.
Bivens, J. A, & Berk, L. E. (1990). A longitudinal study of the development of elementary
school children's private speech. Merrill-Palmer Quarterly, 36, 443-463.
Cohen, A R. (1961). Cognitive tuning a factor affecting impression formation.Journal of
Personality, 29,235-245.
Damon, W., & Killen, M. (1982). Peer interaction and the process of change in children's
moral reasoning . Merrill-Palmer Quarterly, 28,347-367.
Damon, W., & Phelps, E. (1989) . Critical distinctions among three approaches to peer
education. International Journal ofEducation Research, 13,9-19.
Davis, J. H., Stasser, G., Spitzer, C. E., & Holt, R. W. (1976). Changes in group members'
preferences during discussion : An illustration with mock juries . Journal of Personality
and Social Psychology, 34, 1177-1187.
Diaz, R. M., (1992). Methodological concerns in the study of private speech. In R. M. Diaz
& L. E. Berk (Eds.), Private speech: From social interaction to self-regulation (pp. 5581). Hillsdale, NJ: Erlbaum.
Diaz, R. M., & Berk, L. E. (Eds.). (1992). Private speech: From social interaction to selfregulation . Hillsdale, NJ: Erlbaum.
Dunbar, K., & Klahr, D. (1989). Developmental differences in scientific discovery
strategies. In D. Klahr & K. Kotovsky (Eds.), Complex information processing: The
impact ofHerbert A. Simon (pp. 109-144). Hillsdale, NJ: Erlbaum.
Ellis, S., & Gauvain, M. (1992). Social and cultural influences on children 's collaborative
interactions . In L. T. Winegar & J. Valsiner (Eds.), Children's development within social
contexts (pp. 155-180). Hillsdale, NJ: Erlbaum.
Ericsson, K. A., & Simon, H. A (1980). Verbal reports as data. Psychological Review, 87,
215-251.
Gauvain, M., & Rogoff, B. (1989). Collaborative problem solving and children's problem
solving skills. Developmental Psychology, 25, 139-151.
Giles, H., & St. Clair, R. N. (Eds.). (1985). Recent advances in language , communication,
and social psychology. New York: Erlbaum.
Grice, H. P. (1989) . Studies in the way of words. Cambridge, MA: Harvard University
Press.
Klahr, D., Dunbar , K., & Fay, A L. (1990). Designing good experiments to test "bad"
hypotheses . In J. Schrager & P. Langley (Eds.), Computational models of discovery and
theory information. San Mateo, CA: Morgan-Kaufman.
Klahr , D., Fay , A. , & Dunbar , K. (1993) . Heuristics for scientific experimentation: A
developmental study. Cognitive Psychology, 25,111-146.
Kruger, A C. (1992). The effect of peer and adult-child transactive discussions on moral
reasoning . Merrill-Palmer Quarterly, 38,191-211.
Kruger, A C., & Tomasello, M. (1986). Transactive discussions with peers and adults.
Developmental Psychology, 22, 681-685.
384
StephanieD. Teasley
Kuhn, D., & Brannock, J. (1977) . Development of the isolation of variables scheme in
experimental and "natural experiment" contexts. Developmental Psychology, 13, 9-13.
Kuhn, D., Amsel, E., & O'Laughlin, M. (1988). The development of scientific thinking
skills. New York: Academic Press.
Light, P., & Perret-Clermont, A-N. (1989). Social context effects in learning and testing.
In A Gellantly, D. Rogers, & J. A Sloboda (Eds.), Cognition and social worlds (pp. 99112). Oxford, England: Claredon Press.
Piaget, J. (1965). The moral judgment of the child. Glencoe, IL: Free Press. (Original work
published, 1932).
Pontecorvo, C. (1987). Discussing for reasoning: The role of argument in knowledge
construction. In E. DeCorte, H. Lodewijks , R. Parmentier, & P. Span (Eds.), Learning
and instruction: A publication of the European Association for Research on Learning
and Instruction . (Vol. I, pp. 71-82). Oxford, England: Leuven University Press.
Pontecorvo, C., & Girardet, H. (1993). Arguing and reasoning in understanding historical
topics. Cognition and Instruction, 11,365-395.
Schauble , L. (1990). Belief revision in children: The role of prior knowledge and strategies
for generating evidence. Journal ofExperimental Child Psychology, 49, 31-57.
Schegloff, E. A (1991) . Conversation analysis and socially shared cognition. In L. B.
Resnick, 1. M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition
(pp. 150-171). Washington, DC: APA Press.
Sproull, L., & Keisler , S. (1991) . Connections: New ways of working in networked
organizations . Cambridge, MA: MIT Press.
Teasley, S. D. (1995) . The role of talk in children's peer collaborations. Developmental
Psychology, 3, 207-220.
Teasley, S. D., & Resnick, L. B. (1993) . Social influences and cognitive change: Peer
collaboration in a causal reasoning task. Unpublished manuscript, University of
Pittsburgh, Learning Research and Development Center, Pittsburgh, PA
Teasley, S. D., & Roschelle, J. (1993) The construction of shared knowledge in
collaborative problem solving . In S. Lajoie & S. Derry (Eds.), Computers as cognitive
tools (pp. 229-258). Hillsdale, NJ : Erlbaum.
Tetlock, P . E. (1983). Accountability and complexity of thought. Journal of Personality
and Social Psychology, 45,74-83.
Tudge, J. (1992) . Processes and consequences of peer collaboration: A Vygotskian
analysis. Child Development, 63, 1364-1379.
Wertsch, J. V. (1985) . Culture, communication, and cognition: Vygotskian perspectives.
Wertsch, J. V., & Stone, C. A. (1985) . The concept of internalization in Vygotsky 's
account of the genesis of higher mental functions . In J. V. Wertsch (Ed.), Culture,
communication, and cognition: Vygotskian perspectives (pp. 162-179) . Cambridge,
Zajonc, R. B. (1960). The process of cognitive tuning in communication. Journal of
Abnormal and Social Psychology, 61, 159-167.
Chapter 16
Seeing the Light:

Discourse and Practice in the Optics Lab
Roger Saljo and Kerstin Bergqvist
Department of Communication Studies, Linkoping University, Sweden
Abstract
The general issue addres sed in this chapter concerns the relation ship between
perception and discourse. The empirical analyses reported focus on how students and
teachers, working in the context of a physics laboratory in school, communicate about
the properties of light. It is shown that, in order to understand the behavior of light in
the experiments that are set up, one has to have access to elements of a theory of light
that make the phenomena produced appear as significant according to a particular
perspe ctive. Thus, seeing in the sense of identifying something that is culturally and
contextually significant is a sociocultural process that relies on discursive resources.
Introduction
The theoretical background of this chapter reflects an interest in the mediated
nature of human mental and practical action. The study is based on an empirical
field study of student/teacher and student/student cooperation in the context of
the physics laboratory. The specific problem analysed concerns work done in the
area of optics on such topics as light sources, reflection and refraction of light,
and similar phenomena. The teaching and learning activities take place in a toolsaturated environment in which equipment such as the optical bench, prisms of
various kinds, mirrors, lenses, diaphragms , lamps, and a range of other devices
386
Roger SlHjo and Kerstin Bergqvist
are used . Our interest here concerns the extent to which the participants share
perspectives on what is being talked about, and the sense in which they have
access to discursive tools that allow for the instruments to serve as a means of
providing shared - and new - definitions of the nature of light.
In the rhetoric surrounding teaching and learn ing within the natural sciences,
experimentation and hands-on manipulation of physical objects are often claimed
to serve as important vehicles in attempts to ground physical concepts and
explanations in the life-world of learners. Experimentation is seen as a significant
means for the creation of practical and concrete experiences in the teaching of the
abstractions that characterize theoretical concepts and mathematical models of
the physical world in science . The laboratory-like setting in the school offers an
arena in which systematic observations of phenomena can be made in a direct and
supposedly unmediated fashion, and where what is to be learned can be supported
by almost bodily sensations. The students when "using their hands, putting
together and taking apart things ... get much more out of it [the teaching]," as
one of the teachers participating in our study put it (cf. Bergqvist & Saljo, 1994,
p.153).
However, rather than discussing this general issue of the role of child activity
and experimentation in education, we focus on the more limited question of what
is presupposed by individuals in terms of their ability to perceive events in order
to make sense of certain kinds of physical phenomena illustrated in the context of
optical experiments . When "looking" at light, what do we "see"? And, how is this
"seeing" related to sociocultural experience?
The assumption that we perceive directly and by means of a simple registration
of visual stimuli is a prominent component of Western lay and scientific
conceptions of the process of seeing. The physical world is construed as objective
and out there, and, in most accounts, seeing is described as a democratic
phenomenon, as it were; when we look at the world from a similar perspective,
we have the same sensations and retinal images and, thus, "see" the same things .
In order to understand the difficulties that the young people whom we have
observed have in grasping the basic principles of optics when working with the
equipment designed to illustrate some of these principles , it becomes evident that,
ina sociocultural perspective (Wertsch, 1991), perception has to be accounted for
in a radically different way. Retinal images and sensations are part of the picture ,
but they offer very few clues to the situated meaning of what was illustrated in
the particular experiments that we observed. Yet, it was obvious that the teachers
were convinced that phenomena such as refraction of light, reflection angles, and
other similar basics of optics were there to be unproblematically seen by
everyone when looking at the optical bench and the other devices used in the
laboratory. However, from a sociocultural perspective, seeing must be construed
as intimately related to collective and individual experiences that provide a
person with perspectives on what to see in what is out there (Saljo, 1992). In the
optics lab, there is, we argue , a subtle interplay between technological and
discursive tools in the constitution of phenomena to be perceived, and the
Seeing the Light
387
artifacts very clearly embody distinctions that are discursive in nature. Categories
form part of situated practices as Goodwin (this volume) puts it.
The Scientific Construction of Visual Perception

When teachers in our laboratory ask students what they see or what happens in
the experiments, they rely on models of perception that are prominent in the
traditions of the study of perception. In their reductionist ambitions, these models
deemphasize, sometimes even neglect, sociocultural experience. All introductory
textbooks in the field of psychology point to the centrality of visual perception
for the understanding of human psychological functioning . Classically, and in
line with an empiricist stance, perception has been described in terms of what
Rock (1985) referred to as a stimulus theory in which the basic idea "is that for
every distinct kind of perceptual property - of color, size, depth, movement, and
the like - there is a unique stimulus (or type of stimulus information) reaching the
sense organ." (p. 28). This idea of a perceptual automata, restricted to processing
the incoming information and adding nothing to that, soon fell into disrepute . It
can easily be shown that our perceptions of the world cannot be explained in this
simplistic fashion. The regularities we perceive and our ability to compensate for
differences in factors, such as the distance to the objects we look at and the
angles from which we see them, make it obvious that this camera metaphor is
unsatisfactory. There is no one-to-one correspondence between a stimulus and
our interpretations of phenomena and events in the world.
The other group of theories, which we can provisionally label constructivist, is
founded on assumptions that imply that the perceiver and/or the biological
substrate of the perceptual apparatus (including the brain) playa more active role
in the construction of percepts. Thus, the "study of perception is the study of how
we integrate sensations into percepts of objects in our world (a percept is an
outcome of a perceptual process) ." (Atkinson, Atkinson, Smith, & Bern, 1990, p.
157; italics original). The distinction between sensation and perception is a
conceptual construction that has a long history in the study of perception and that
comes back in the diverse range of positions that we can describe as
constructivist. Sensations are what is produced by sense data out there in the
physical world. Percepts are what we make from these once they are processed as the metaphor used in the human-information-processing tradition and more
recent versions of cognitivism suggests - by the sensory organs and in the brain.
The transition from sensation to percept is the process through which we can
escape a constant bombardment of sensations that at any given moment exceeds,
by far, what we can successfully deal with . The efficient selectivity of
perceptions that allow for the world to appear as a meaningful and integrated
whole must be achieved , so constructivist positions tell us, by some "higher
388
Roger SaIjo and Kerstin Bergqvist
cognitive agency outside the stimulus domain" (Rock, 1985, p. 40). At this level,
a broad set of factors that have to do with our experiences, memories , prejudices,
taken-for-granteds, and fantasies can begin to playa role for perception . Thus, "in
moving from biological and sensory processes (. . .) to perception, we have
reached a point in our discussion of psychology where the mind now steps in and
makes contact with the information" (Seamon & Kenrick, 1992, p. 120). The
statement that the "mind ... steps in" is a significant and revealing expression in
this context. The interaction between the individual and the outside world is
generally conceived in terms of the mind as an inner eye or homunculus
analyzing and observing the sense impressions entering the perceptual system . In
most accounts of perception, there is very little room for sociocultural experience
to play any decisive role for perception. The higher cognitive agency alluded to
by Rock is mostly given a very narrow interpretation and is more of a gatekeeping function; "Our senses are our only means of knowing about the world,
but they are not windows that let the world in. They are extensions of the brain
that respond to different forms of physical energy" (Seamon & Kenrick, 1992, p.
119). In the transition from external stimuli, via retinal projections, to neural
processing of the signals, there is generally very little room for shared discursive
tools, what Winograd and Flores (1986) referred to as resources emerging from
"the consensual domain," to enter the picture as mediating devices in perceptual
orientation , which we return to shortly .
Our intention is not to give a full account of theories of perception, but it
should be pointed out that there are radically different interpretations of the
nature of perception that imply different positions with respect to the relationship
between exper ience and perception . One notable example is a phenomenological
perspective of the kind offered by Merleau-Ponty (1962/1974) , which obviously
does not always qualify as a theory of human perception, because it is generally
not given serious attention in main-stream textbooks . In this case, the distinction
between sensation and percept is not accepted as a point of departure. In fact,
Merleau-Ponty argued that "once introduced, the notion of sensation distorts any
analysis of perception" (p. 13). This "empiricist construction . .. hides from us
'the cultural world' or 'human world' in which . . . almost our whole life is led."
(p. 23). In phenomenological interpretations, the topic of inquiry is instead
human consciousness and the intentionality of experience. The problem of what
is objectively out there and what is subjectively perceived, referred to in the
distinction between sensation and percept, disappears as the basic point of
departure for analysis. Instead, the assumption of the primacy of the life-world in
experiential terms forms the background for the understanding of perception (and
other cognitive phenomena).
Seeing the Light
389
Perception and Sociocultural Experience

The general issue of how to construe the relationship between human perception
and sociocultural experience is, of course, a complicated matter that invokes
classical philosophical debates regarding the nature of knowledge. The strategy in
which the study of perception has been conducted implies, as we have tried to
indicate, that a sharp line of division has been maintained between what is
construed as pure perception , on the one hand, and our knowledge of the world
and our participation in cultural practices, on the other. Indeed, this is the very
definition of the study of perception . The issue of how people perceive complex
and dynamic phenomena as part of their involvement in practical action has been
transformed into the study of a static and largely passive phenomenon,
perception, located in the biological substrate of the human brain and sensory
organs. This form of reductionism is analogous to how, for instance, the human
activity of remembering has been transformed into the study of a postulated
entity referred to as memory, which is assumed to exist as an autonomous process
to be accounted for per se (Middleton & Edwards, 1990). Memory and
perception in context can, following this model, be conceived in line with the
cognition plus model of cognition, criticized by Lave (1988) , in which the
assumption is that one can simply add context on to models of cognition when
understanding situated practices (cf. Jonsson, Linell, & Saljo, 1991). However, to
understand human practices and the role of cognition in these, it is necessary to
adopt a conceptual perspective in which thinking, action , and context are
construed in commensurable terms. Thus, there is nothing more fundamental to
cognition than its manifestation in action, and our unit of analysis has to be one in
which people are acting in situated practices.
In what was referred to as a stimulus theory earlier , a mechanical account of
perception is, of course, to be expected, because percepts are essentially copies
reflecting the external world . However, also within the broad group of
constructivist theories that thus are premised on the assumption that perceiving
involves some sort of cognitive effort on the part of the individual, the role and
nature of sociocultural experience are often recognized in a limited way only, as
we have already pointed out. There is a strong tendency to account for the
constructive nature of human perception at the neural level where there is little, if
any, room for sociocultural experience to playa role . This seems to have been
true for the interpretation of perc eption within Gestalt psychology as well
(Koffka, 1935; Wertheimer, 1912). Thus, even though perception was placed at
the heart of Gestalt theorizing and even though the studies often concerned
complex perceptual phenomena, the explanations of the observations were in
most cases sought in biological processes.
However, also in modern constructivist positions that attempt to account for
perceptual phenomena at a psychological level, the role of sociocultural
experience for perception is often not considered crucial. Even in the widely
390
acclaimed theory of perception (and cognition) outlined by Neisser (1967, 1976),

in which the constructive nature of such events is held to be the cornerstone of
the functioning of the mind, the contribution of sociocultural experience is not
taken into account. As Harre and Gillett (1994) pointed out in their discussion of
Neisser's position on the nature of the constructive processes in perception , "[He]
appreciates the essential embeddedness of perception in the activity that it
informs but he does not really advance beyond tentative suggestions about the
roles of meaning and social context in this process" (p. 168). Knowledge and
experience are thus generally not construed as constitutive of cognitive processes .
Rather , such processes are seen as abstract elements that operate on knowledge
and that underlie what we do in worldy projects, but they are not themselves
products of sociocultural experience. Rock (1985), in his argumentation for a
cognitivist (and constructivist interpretation) of perception, claimed that " If
knowledge is defined as conceptually based, consciously apprehended
information one has about an object or event, by and large it does not affect the
perception of that object or event" (p. 300).
When attempting to understand how people perceive events in contexts of
human practices, it becomes evident that sociocultural experience is critical for
what we are able to see. For instance, differences in expertise and familiarity with
a particular practice will result in very different ways of seeing. Thus, individuals
sitting next to each other at a soccer or ice hockey game are at some level
exposed to identical stimuli when watching, yet the nature of their seeing and
understanding might be very different. To the expert, the strategies and tactics are
readily visible in the way in which the team is currently playing, whereas the
novice might see only players moving around the field or rink trying to get hold
of the ball or puck. There is a differentiated language with a vocabulary that
codifies significant events in a game and that is highly meaningful to the expert.
Distinctions such as offside, 4-2-4, and push-up index aspects of the soccer game
that constitutes it in a particular way for those who have the necessary expertise .
At the same time, such distinctions can be used for efficient communication
among in-groups. In a similar vein, for the judge at a dog exhibition, the
movements of a dog walking around in a circle convey information that is not
available to a person with less expertise exposed to the same situation. Variations
in the length of the steps, the coordination of the legs and the body, and the head
posture are highly "meaning -full" signals that convey essential information to the
judge about the physical status of the dog and maybe even about its temperament
and psychological stability.
The fact that there are cultural determinants of perception has been illustrated
in research of different kinds. In crosscultural work, Cole, Gay, Glick, and Sharp
(1971), for instance, have shown how the ability to make estimations of amounts
of rice differs between people for whom this task is grounded in a familiar
practice and those for whom this is not the case . As early as 1901 , Rivers,
studying visual illusions among the Toda people of India, found that they did not
perceive illusions to the same extent as did Westerners. Along similar lines,
Seeing the Light
391
Segall, Campbell, and Herskovits (1966) and Luria (1976), for instance, showed
how susceptibility to visual illusion cannot be explained by reference only to
physiological phenomena but has to be related to culture. In his studies of the
introduction of literacy among peoples in the southern republics of what was then
the Soviet Union , Luria (1976) concluded that "illusions are not universal" (p.
43) and that the higher the level of formal education, the more likely the
participants in his studies were to perceive illusions.
A different type of work, but documenting what can be interpreted as similar
results, was carried out by de Groot (1965). In his well-known studies of chess
players, de Groot found that there were very marked differences in which expert
and less expert players were able to reproduce the positions of the chess pieces
when being allowed to see the board for brief intervals only. In particular, if the
positions of the pieces could be seen to represent situations that are likely to arise
in a game played by people with considerable skills, players with a high level of
expertise - grand masters and masters - were very accurate in their ability to
describe the positions of the pieces. de Groot argued, "It is evident that
experience is the foundation of the superior achievements of the masters" (p. 329)
in such situations (although this is not necessarily the case with unusual or atypical games [cf. Djakow, Petrowski, & Rudik, 1927]), because highly skilled
chess players are able to pick up "the essential relations between thepieces, their
mobility and capturing possibilities, their co-operation or opposition" (p. 331).
Thus, the "entire perception occurs in fundamental relations and possibilities and
is in a sense dynamic" (p. 331), and the "M[aster] sees more than [the] E[xpert]"
(p. 334, all italics in original). The examples could be multiplied, but the
argument could very well be made that, when considering perception and its role
in social practices, what Rock referred to as "conceptually based, consciously
apprehended information" must playa critical role. We perceive reality by means
of the sociocultural resources that we find applicable to concrete practices
(Wertsch, 1991). The origin of these resources must be sought not in the
biological substrate of human thinking and perception, but rather in discursive
practices that have developed in specific contexts and that contain distinctions
that mediate these practices in rich detail and according to certain patterns.
Developing Perceptive Sensibility

When viewing sociocultural resources as constitutive of perception in social
practices, it becomes necessary to apply a developmental and sociogenetic (de
Graaf & Maier, 1995) perspective on how people perceive events and phenomena
in the world around them. Thus, the growth of our ability to perceive meaningful
entities is intimately related to our involvement in a broad range of
communicative and social practices . Again, we can draw on the work by Luria
392
(1976) on the psychological consequences of introducing literacy and schooling

in a community for illustrating this point. When presented with different kinds of
geometrical figures (Figure l ), there was a clear difference in how participants
varying in their exposure to formal education and literac y practices designated
these figures.
0 U
D u
D
1
I~
13
14
.,
, ,
I ,
C::J Q]6
15
~
9
i..
1\
10
11
c:J - -
"
IS
19
Fi g. 16.1. Geometrical figures used by Luria (1976, p. 33) when studying perception and
categorization.
People who had participated in discursive practices characteristic of formal

schooling and who were, in some sense, literate, characterized figures such as
these by means of what Luria refers to as "categorical names" (p. 32). This
implies that they used terms such as squares and circles as lingustic terms for the
figures. In the case of incomplete figures such as the second, fifth, and eleventh,
they would still fall back on categorical and abstract terms to accou nt for the
figure . Thus, Luria's second figure would be "something like a circle" and the
fifth and eleventh "something like a triangle" (p. 32). These observations led
Luria to a polemical statemen t against the Gestalt theorists when arguing that
"the laws of 'good form' (prtignanz) and of structural continuatio n (. . .) as
described by Gestalt psychologists are fully apparent only for subjects who have
mastered geometrical concepts, and do not appear in people who perceive shapes
in an object-oriented fashion." (p. 33). In other words, we cannot think of these
matters in terms of pure perception. People who had not been exposed to
schooling and literacy, on the other hand, tended to use concrete object names for
describi ng the figures. Thus, Luria' s fourth figure would be a mountain, second
figure a moon, eigh teenth a road, and nineteenth a thread. They also denied
perceiving any resemblance between figures that could be described by the same
generic geometrical term (for instance, the squares 12 and 13 in Figure 1).
When construing perception in context as intimately intertwined with
sociocultural experience, it will be critical to study institutionalized accounts of
phenomena in order to understand how our abilities to discriminate and describe
develop as a function of institutional experience (cf. chapter by Goodwin, this
volume) . In modern society, expertise and modes of construing reality are, to an
increasing extent, developed within social institutions that produce discursive
tools and what Shotter (1990) referred to as accounting practices by means of
which phenomena in the world can be perceived and explained. Science is one
Seeing the Light
393
example of such an institution specifically committed to the production of new

discursive tools that should be used as resources in perceiving and explaining the
world. And the study that we have conducted specifically concerns how light is
perceived and understood in experimental settings and what it takes to see what is
supposed to be seen in a set of experiments set up for educational purposes .
The Setting
In optics laboratory in the school where we have been studying student/teacher
and student/student interaction, the obvious purpose of the activities is to develop
new conceptions of light among the students. Students are to acquire, by means
of what is referred to as concrete experimentation, modes of understanding the
properties and behavior of light that are not necessarily part of their previous
experience. The data used in the following discussion derive from a participant
observation study of work in the physics laboratory among secondary school
students (aged 13-14). Students in two classes were followed during one year
through participant observation. One focus of this research was to study group
work among the students. The particular project we used as a vehicle for
illustrating some of the dilemmas of how light is construed in different
sociocultural practices dealt with some of the basic concepts of optics. During a
period of seven lessons spread out over three weeks, the students were to learn
about phenomena such 'as refraction and reflection, light sources and shadows,
and mirrors and lenses. Important concepts that were introduced and illustrated in
the experiments included inflection- and reflection-angle, and focus and focal
distance. A significant proportion of the time was devoted to hands-on
experimentation using the so-called optical bench, which is illustrated in Figure
2.
lamp
diaphragm
screen
Fig. 16.2. The optical bench.
394
Roger Sliljo and Kerstin Bergqvist
In the experiment illustrated here students learn about the refraction of light as it
passes through different substances (cf. Figures 4 and 5 where students'
difficulties in interpreting this experiment are discussed in some detail. For an
illustration of a different experiment using this device, see Figure 3). During the
lessons, two teachers (Anders and Randi) shared the responsibility for instruction .
(Details about the design and the data collected can be found in Bergqvist, 1990).
The Theory-Laden Nature of Seeing in the Optics Lab

As we have shown (Bergqvist & Saljo, 1994), the pedagogical philosophy
underpinning the lessons was heavily Piagetian in nature. The students were
encouraged to "test," "find out things," "be creative," and to examine the
equipment to see "what happened." The emphasis was very clearly on creating a
set of activities in which students would interact among themselves and with the
equipment. As Anders put it, "In a good science lesson, they should be doing
something; they should be using their hands, putting together and taking apart
things. That's when its good. They get much more out of it if they do that."
Another dimension of this learning philosophy is its inductive nature; learning
follows from observation and activity generated by the students' own interests
and spontaneous questions rather than by guided participation in which the role
of the competent member, the teacher, is critical for conceptual change to occur
(cf. Edwards & Mercer, 1987; Walkerdine, 1984). According to Anders, "Let
them test . . . try to get some sort of thread by means of their questions. They're
sitting and waiting for answers to their questions, not that I should say something
like, 'We're now going to do this and this, and then this will happen."
To the teachers following this philosophy of teaching and learning, students are
construed as approaching the learning situation equipped with questions that
allow them to see "what happens" with light in the different experiments. The
expression what happens is, in fact, critical as it is used by the teachers and
students throughout the experiments carried out during the lessons. But precisely
what is meant when saying that something happens in the context of light is
dependent on the sociocultural resources one utilizes in the situation and may
remain unclear to students whose conception of light does not always allow them
to discern in what sense and according to what patterns "something happens"
with light. This is what we explore next.
The following excerpt is taken from a section of one of the lessons in which
two points are illustrated simultaneously: (l) the fundamentals of refraction of
light, and (2) that white light is a combination oflight with different wavelengths.
The students are experimenting with various prisms , a mirror, and a light source
on the optical bench, which is arranged as shown in Figure 2, except that this
395
Seeing the Light
time there is a prism in the middle, and on the screen the lights of different
wavelengths appear.
Anita
Inga
This will be real fun

Check this, real cool!
Anita
Eva, Anita
Anita
Inga
Anita
Yees!
Wow ... Wow
Discolights
Yes
The blue one is the
nicest
But what are we supposed
to use this one for?
You probably give an
angle to the light in
some way
Check this, what a nice
light that this!
Nothing is happening it
it seems to me. What is
this then?
Yes, that's right,
that's what we're
supposed to do. You
should point the light
onto here . And then you
would see how the light
is reflected.
This was really boring
No, I don't get it
Me neither . But that's
nothing to worry about
What are you supposed
to do with that thing
then?
What a cool light! He
Eva
Anita
Anita
Eva
Anita
Inga
Anita
Eva
Inga
Inga
Anita
Eva
ANDERS
Eva
... )
It's no fun Anders .

Nothing's happening!
Nothing 's happening here .
Either we're stupid or
it's .. ..
What are you doing, then?
Nothing.
(Looking at the dispersion

of the the light source as
it passes through the
glass prism)
(Points to a mirror)
(Reports that "nothing "

is visibly happening, although
the light is refracted)
(One of the prisms)
396
rnga
ANDERS
Eva
rnga
Eva

Nothing .
r see. You're doing nothing.
Well, then nothing will happen .
Oh yes! We're doing lots of
things . Yes , indeed, we're doing
lots of things but still
nothing's happening .
We 're finding masses of these
things to do and .. .
We don 't know what it's for!
This excerpt illustrates issues of perspectivity of "seeing" in this particular

context. To understand the problems that the girls are experiencing and the
demoralization with respect to their own work that is apparent, one must
scrutinize the role of discursive resources in perception . Although the girls at one
level are actively involved in what they are doing and they are using the light
source (a lamp), the prism, and the mirror in the intended way, they are not doing
this under a conceptual interpretation that leads them to define what they see as
something very distinctively "happening." Thus, when observing that light is
reflected when it hits a mirror , the girls fail to construe this as significant
evidence. The reason for this, as we argue shortly, is that the students find
nothing in this that has to be accounted for in any particular way. In this instance,
one needs the conceptual resources of a particular theory of physics in order to be
able to see that the reflection against the mirror has to be explained rather than
taken for granted. Thus, and somewhat paradoxically, one needs to have access to
a sophisticated and institutionalized theoretical construction of the nature of light
to be able to realize that something needs to be accounted for in a perfectly
normal event. The rationale underlying the task, which the students are calling
out for when claiming that they "don't know what it's for!", becomes visible only
when attention is guided by distinctions that originate from discursive practices
produced within a certain scientific community.
A similar type of coordination problem between teachers and students is visible
in the continued shared activity in which the optical bench is used. This device is
interesting from a sociocultural perspective, because it has been designed to
allow for optical phenomena to be clearly illustrated. A number of critical
experiments on the nature of light can be carried out by means of this bench, and
the students were given ample opportunity to experiment on their own. In one of
the lessons, in which the teacher seemed to experience that the work degenerated
into play , she (Randi) decided to explain one of the fundamental principles of
light that could be discovered from a proper use of the optical bench. Again the
problem is talked about in terms of attempting to explain "what will happen"
when light behaves in specific ways, as can be seen from the following excerpts .
Seeing the Light

Armi
RANDI
Armi
RANDI
Armi
RANDI
397
Randi, what are we going to do?

You're to try to figure out what
is relevant . . ..
But we've already done that.
I'll give you a hint.
Yes, you can give us a hint about
what we should do because I don't
get it at all.
Look here . I 'll help you a little
bit because this is difficult t o
figure out by oneself.
Randi continued by making a drawing on the blackboard (Figure 3) in which the

lamp on the left served as the light source generating light that was to hit a piece
of paper. This is the base-line of many experiments with the optical bench. In her
drawing and to illustrate her point, Randi put a pen in an upright position between
the light source and the paper as shown in the figure. On the drawing, she
reproduced the shadow of the pen in black on the paper that she had drawn.
RANDI
Armi
RANDI
[while drawing) Here is the lamp, sort of.

Here is the pen . Here is the diaphragm.
The lamp here. Well, now it sends out
rays in all directions, like this really.
What will happen then? What will happen
to the rays?
They will disappear further up.
Yes, on the side of the pen it is light .
Fig. 16.3. Drawing of optical bench illustrating that light cannot go through solid objects.
398
In commenting on the illustration, she pointed out that the pen, being a solid
object and placed in front of the light beam, would produce a shadow on the
paper. As an explanation of this phenomenon, she added that "light can only go
straight forward and it cannot go through a body" (Randi), and, thus, there will be
light on the side of the pen. This is what the students were to discover themselves
by experimenting with the equipment. When students were told this, they seemed
anything but enlightened. Again we would argue that this is an interesting point
at which the discoordination in perspectives between teachers and students
looking at the same physical event must be understood by reference to
sociocultural resources. For the students, the observation that light cannot go
through solid objects seemed normal and not in need of any explanation. They
were probably also aware that light can only go straight forward. Therefore, well
before they had ever been exposed to this problem in the physics class, they
would be able to predict what would happen in a situation of this kind . The
shadow on the paper would require no special explanation.
For Randi, however, who has been socialized into the discursive communities
of physics and physics teaching , this is a fruitful way of recontextualizing a
fundamental principle in a concrete fashion . She construes this situation as a
context in which basic principles of the properties of light are displayed and there
is a problem to be accounted for. The interesting observation from our point of
view is, however, that this socialization into specialized communities of practice
in a particular field to some extent seems to make it difficult for her to see what
nonmembers can possibly see when facing the experiment. What needs to be
explained in one discursive practice can remain tacit in the other one and vice
versa.
The technology in the laboratory thus already embodies a theory of optics. The
optical bench has been designed on the basis of insights into optics that have been
made during literally thousands of years. In this sense, the equipment is heavily
theory-laden, and it contains forms of knowledge built into a physical device .
Learning About the Refraction of Light

What constitutes a situationally appropriate way of perceiving light depends on
the context in which one encounters these phenomena. The students are in a sense
at the crossroads of two, possibly more, discursive practices when participating in
these experiments: their own taken-for-granted , common-sensical understanding
and the one more or less explicitly offered through schooling. Many of the
phenomena addressed have been encountered in everyday experiences outside
formal schooling, as we have already argued, and there is some sort of
explanation that is seen as satisfactory for practical purposes in such settings . A
significant part of the lectures was devoted to the notion of refraction of light, a
Seeing the Light
399
phenomenon that students were familiar with from other contexts ; and the very
purpose of this inductive teaching strategy is to use the students' personal
experiences as a source for learning (cf. Wistedt, 1994). During the work on this
topic, the students were given the following tasks:
(I) Study and make a set of drawings showing the beams of light going through
different plastic objects, a rectangle , a triangle, a half circle , and check if the beam
rebounds. Make at least two positions for each object. Do the same thing with a lens,
which will make five beams. Make drawings on sheets of paper that you attach to the
screen.
(2) Discuss what happened with the light and why, on the basis of one drawing with a
fairly strong refraction .
Prior to working with these tasks , the teacher had introduced the topic of
refraction of light, among other things, by means of illustrations showing how
light "slows down and turns off' when it passes from a thin substance to a more
dense substance, for example, from air to plastic prism. The illustration made on
the board is reproduced in Figure 4.
Fig. 16.4. Drawing illustrating principles of refraction of light through different substances
with differing optical densities.
As can be seen in the instructions, the favorite metaphor of inductive teaching

appears again. The students were supposed to show what happened with the light
and why, and this implies that the students were indeed looking for phenomena
that would qualify as "something happening." Again, there are similar problems
in finding events within the experiment where something can be argued to
"happen." In order to assist the students in their work, one of the teachers made a
drawing on the blackboard (see Figure 5) that was present throughout the work
with the principles of refraction and that thus could be used as a resource by the
400
students. The drawing was intended as an attempt to explain refraction by means

of an analogy to the movements of a sledge passing over different materials.
sand
sno w
Fig. 16.5. Drawing illustrating the principle of refraction by means of analogy with a
sledge moving across different surfaces.
The idea communicated is that the sledge slows down when friction increases as
it passes over the sand and that it gains in speed when back on the snow again.
This concrete reference to a bodily experience would then serve as an analogue to
how light behaves when moving through objects with different densities .
Even though the students had an initial understanding of the fact that light
refracts, their work on the task given is not very successful. Although the students
might in some instances perceive that the light refracts, they have constant
difficulties in constituting this as something that calls for any explanation . In
addition, they find it difficult to move between the two presentations: what they,
at best , would see when they work with the prisms, and the example with the
sledge .
ANDERS
Edvard
ANDERS
So let's think about what happens

(Points along the line
showing refracted
when the sledge arrives here . Is
beam)
it correct that
he would have gone in that
direction?
Couldn't he possibly have gone
in the other direction?
We don't know
Yes, well think in the same way as (Again referring to the
sledge example on blackthere
board and then leaving)
Edvard
Ola
It depends on who's on the sledge

It depends on who's on the sledge,
Anders
(Calling out in a teasing

tone of voice)
Seeing the Light

Edvard
Ola
Edvard
Ola
Edvard
Ola
401
If it is you [Anders] you 'd never

know how . . . Then it's probably
quite true that it would move like
that ...
Want a smack?
What?
Do you want me to give you a smack?
I mean Anders, if it was him going,
then it would have gone there bang
then there bang, he would have come
there bang , bang, bang . . . .
Are you going to play table tennis
tomorrow?
A major reason for the difficulties students have is, thus, the explanation offered
in Figure 5 and present on the board during the lesson. The analogy between the
sledge and refraction of light is illuminating only if one already has an
understanding of the concept of optical density. In that case, one can subsume the
slowing down of the light as it passes through a solid object (with higher density
than air) and the movement of the sledge on different surfaces under the same
general principle. However, this connection has to be made on an explicit
conceptual level that codifies the slowing down as a shared feature of the two
examples. What the students at best have at their disposal is an intuitive, yet not
verbalized, understanding tha t light will turn off when it hits objects or
substances (e.g ., water), but they do not understand how this turning off is
represented in the picture through the angle on the line symbolizing the light
beam, and they are not able to grasp what the turning off (which is clearly visible
on the drawing) has to do with the light beam slowing down. Indeed, it is not
clear whether the idea that light can slow down is at all connected to the idea of
refraction for the students. Light simply seems to be there in a straightforward
manner, and there are no signs that the students, while working, think of the light
beam as moving from one point to the other. What is necessary, and what is
available within the community of practice within which the teachers operate, is a
distinct codification of the pri nciples of reflection tha t allows for the two
instances to count as similar . It is the familiarity with what Lemke (1990) refers
to as the thematic patterns of scientific reasoning that connects the two
illustrations as illustrations of something similar. In the students' work, the
movements of the sledge become the major topic of discussion, as is shown in the
preceding example when the discussion centers around how the sledge would
move if the teacher would be the person sitting on it. The connection to reflection
of light is clearly beyond reach at a thematic level.
402
Concluding Comment
These observations illustrate the very intricate way in which sociocultural
experience enters into our perceptions of events . The technology present in the
optics lab embodies perspectives on the nature of light; it is by no means theory
neutral . At the same time, it is reconstructed by teachers as instruments through
which students can experiment and make experiences more or less without
interpretation. We have shown that it is doubtful whether this is the case.
However, the main interest in this chapter has been to illustrate how
communities of practice offer a fixation of perspectives (Rommetveit, 1992) by
means of discursive distinctions . We thus learn to see phenomen a by reference to
conceptual resources that originate in specialized communities in complex
societies. Learning in such contexts implies appropriation of accounts of the
world that are neither out there in the objects themselves nor in our brains.
Rather, they are cultivated in institutional settings for particular and sometimes
highly specialized purposes. This implies that learning should not be reduced to a
matter of acquisition of information that is available. Rather, we have to consider
the decisive role of the social distribut ion of different forms of knowledge and
how knowledge is made available to groups and individuals in modern society.
The difficulties that our students have in understanding the principles of light,
and precisely how these are illustrated by means of the tools utilized in the
experiments, have less to do with problems of acquiring knowledge in a narrow
sense or with deficits in conceptual understanding and more to do with their lack
of access to communities of practice in which light is perceived and construed in
the way characteristic of physics. Thus, we might guess that most of the students
we have followed are not participants, not even peripherally (Lave & Wenger ,
1991), in cultural practices in which it makes sense to talk about light using
concepts from optics. On the contrary, in most contexts they would probably run
into problems of making themselves understood, should they try to do so.
A corollary of these observations is the profound extent to which institutions
socialize members into living in only partially shared social realities
(Rommetveit, 1988, 1992). Psychological and cognitive development in complex
societies thus is fundamentally dependent on the opportunities to participate in
institutionalized communicative practices in which specialized accounts and
ways of construing phenomena are cultivated. Thus, to the teachers in our studies,
the basic phenomena of optics illustrated in the laboratory settings were seen in a
very real sense; they were assumed to be there to be perceived by any observer
looking at the equipment. The decontextualization practices (Gustavsson, Linell,
& Saljo, 1993) offered by school subjects and scientific accounts provide filters
that become part of sociocultural resources when perceiving. Socialization of
members into communities of practice is often so effective that it becomes very
difficult to realize in what sense one's perception of an event is dependent on
conceptu al resources and distinctions that are not available to outsiders. We learn
Seeing the Light
403
to believe in the facticity of our observations, and, through the communicative

traditions built into the social practices of teaching and learning, we are led to
assume that display of phenomena will produce the interpretation among students
that we have learned to consider natural rather than sociocultural. In this sense,
we are able to maintain the assumption that our perceptions are realistic and that
the conceptual interpretation is less of a human construction of a phenomenon
and more of something that is objectively out there in the physical world.
The research reported here has been financed by the Swedish Council for
Research in the Humanities and Social Sciences. This chapter was written while
the first author was a Fellow at the Swedish Collegium for Advanced Study in the
Social Sciences (SCASSS) in Uppsala.
404
References
Atkin son , R. L. , Atkinson, R. C ., Smith, E. E., & Bern , D. J. (1990). Introduction to
psychology (10th ed.). New York: Harcourt Brace Jovanovich.
Bergqvist, K. (1990) . Doing schoolwork. Task premises and j oint activity in the
comp rehensive classroom. Link oping : Linkoping Studies in Arts and Science , 55.
Bergq vist, K., & Salj o, R. (1994). Conceptually blindfolded in the optics laboratory.
European Journal of Psychology ofEducation, 9(2), 149-158 .
Cole , M., Gay , J., Glick, A., & Sharp, D. W. (197 1). The cultural context of learning and
thinking. New York: Basic Books.
Djakow, Petrowski, & Rudik (1927). Psychologie des Schachspiels [Psych ology of ches s
playing] . Berlin: De Gruyter.
Edwards, D ., & Mercer, N. (1987). Common knowledge. The development of
understanding in the classroom. London: Methuen.
Gr aaf, W . de , & Maier, R. (Eds .). (1995 ). Sociogenesis reexamined. New York: Springer.
Groot , A. de . (1965 ). Thought and choice in chess. The Hague : Mouton.
Gu stavsson , L. , Lineil, P., & Salj o, R. (1993 ). Discourse on language and discourse in
language . International Journal of Educational Research, 19(3), 265-276.
Harr e, R., & Gillett, G . (1994). The discursive mind. London: Sage .
JOnsson , L., Linell, P., & Saljo, R. (1991 ). Formulating the past. Rem emb ering in the
police interrogation. Activity Theory, 9/10, 5-11
Koffka, K. (1935 ). P rincipl es of Gestalt psychology. New York: Harc ourt Br ace
Jovan ovich.
Lave , J. (198 8). Cognition in practice. Cambridge, England: Cambridge Univer sity Press.
Lave, J. , & Wenger, E. (199 1). Situated learning: Legitimate peripheral participation.
Cambridge, England: Cambridge Universi ty Press .
Lemke, J. (1990). Talking science: Language, learning, and values. Norwood, NJ : Ablex.
Luria, A . R . (1976) . Cognitive development . Its cultural and social fo undations. [M .
Lopez-M orillas and L. Solotar off, Trans.] Cambridge, MA : Harvard Uni versity Press.
Me rleau-Ponty, M. (1974 ). Phenomenology of perception. Lond on : Routledge & Kegan
Paul. (Original work published 1962)
Middleton, D., & Edwards , D. (Eds.). (1990 ). Collective remembering. Lond on: Sage .
Nei sser, U. (1967) . Cognitive psychology. New York : Appleton-Century-Crofts .
Neisser, U. (1976 ). Cognition and reality. San Francisco: Freeman.
Rivers, W . H. (1901). Primitive color vision . Popular Science Monthly, 59, 44-58 .
Rock, I. (1985). The logic of perception. Cambridge, MA : MIT Press.
Rommetveit, R . (1988) . On literacy and the myth of literal meaning . In R. Sii1jo (Ed .), The
written world (pp. 13-40). New YorklBerlin: Springer.
Rommetveit, R. (1992). Outlines of a dial ogi cally bas ed social-cog nitive approach to
human cog nition and com municatio n . In A. Heen-Wold (Ed.), The dialogi cal
alternative - Towards a theory of language and mind (pp. 19-44 ). Oslo : Scandinavian
Uni versity Press.
Siilj o, R. (1992) . Hum an growth and the complex society. Cultural Dynamics, 5( 1), 43-56.
Seamon , J. G., & Kenrick , D. T. (1992). Psychology. Englew ood Cliffs, NJ: Prenti ce Hall.
Segall, M . H., Campbell, D. T ., & Her skovits, M. 1. (1966 ). The influence of culture on
visual perception. Indi anapolis, IN: Bobb s-Merrill.
Shotter, J. (1990). Knowledge ofthe third kind. Utre ch t: ISOR.
Seeing the Light
405
Walkerdine, V. (1984). Developmental psychology and the child-centered pedagogy : The

insertion of Piaget into early education . In J. Henriques , W. Hollway, C. Urwin, C.
Venn, & V. Walkerdine (Eds.), Changing the subject: Psychology, social regulation
and subjectiv ity (pp. 153-202). London: Methuen.
Wertheimer, M. (1912) . Experimentelle Studien tiber das Sehen von Bewegung
[Experimental studies of the perception of movement] . Zeitschrift fUr Psychologie, 61,
161-265.
Wertsch, J. (1991). Voices ofthe mind. Cambridge, MA: Harvard University Press.
Winograd, T., & Flores, F. (1986). Understanding computers and cognition. Norwood,
NJ: Ablex.
Wistedt, I. (1994). Everyday common sense and school mathematics. European Journal of
Psychology ofEducation, 9(2), 139-148.
Chapter 17
Learning to Argue in Family Shared Discourse:

The Reconstruction of Past Events
ClotildePontecorvo and Alessandra Fasulo
Dipartimento di Psicologia dei Processi di Sviluppo e Socializzazione
Universita degli Studi di Roma "La Sapienza," Italy
Abstract
The aim of the study is to identify both modes of children's participation in family
disputes and types of argumentative moves adopted, particularly in the act of
opposing (problematizing) others or defending oneself. The corpus consists of
twenty-seven dinner conversations of ten middle-class families living in Rome
and Naples, each with one child between three and six years and at least one older
sibling .
Data are shown concerning relative distribution of six different discourse genres
(according to temporal focus and presence of problematic events : Ochs & Taylor,
1993) and family members ' role in problematization. A qualitative analysis
illustrates ways in which children are involved and act in family disputes. The
quantitative results indicate that the problematizing activity occupies about 1/3 of
family talk, allowing children peripheral participation in conflict talk; 1/2 of the
problematizations are directed to the children . When discourse concerns past
events, children show a lower rate of problematizing activity (31.3% vs. 40.9% of
the whole of conflict talk) but - when challenged on their past behavior - they
appear to have already learned at 4 or 5 years how to justify themselves and to
provide rhetorically designed answers, using appropriate temporal markers,
authority references, and visual recall devices. Children 's orientation to social
and/or family norms and values is also discussed.
Learning to Argue in Family Shared Discourse
407
Theoretical Background
Exploring learning processes through family talk requires laying out some
theoretical premises, although increasing numbers of scholars have recently
focused on natural interactive contexts for understanding the acquisition of
competences . In this view, cognition and knowledge acquisition are considered the
products of collaborative, repeated, and regulated activities within social domains.
Cognition is seen as a property both of persons and of the relationships between
them and also of the local rules and affordances of the setting: in a word, as a
property of situations in a Goffmanian sense (Goffrnan, 1964; 1974).
School, family, and academic institutions are places where particular
sociocognitive practices are active, part of which is represented by the specialized
types of discourse that circulate within them. By looking at the discursive activity
taking place in a specific setting, we have a key to understanding both the kind of
social situation that we are observing and the nature of the competence that it
demands of its members.
Goodwin's (1994) studies of apprenticeship in the field of chemical and
archeological research have shown how the role of the expert is that of monitoring
and correcting the procedures used to deal with the specific reality with which the
apprentices are presented. The accomplishment of activities within these reality
domains is the joint product of information and actions brought forth by different
persons and applied to the concrete objects they are manipulating.
From a developmental point of view, the acquisition of children's cognitive
capabilities is supported by the scaffolding of adults (Bruner, 1986), so that adult
and child constitute a cognitive system in which one's developed skills are at the
service of the other's purposes (Kaye, 1982). This view can be located in a
Vygotskian framework that considers higher psychic functions as internalizations
of social actions through the mediation of language (Vygotsky , 1934/1990) . In a
sense, the very definition of what psychic functions are (e.g., problem solving,
memory, reasoning) emerges from the discourse that sustains activity (Resnick,
1991). Discourse analyses in educational settings such as school and home have
shown that what gets transmitted is not abstract information but a set of rules on
the right methods for achieving proper knowledge (Edwards & Mercer, 1987;
Orsolini & Pontecorvo, 1992); for displaying memory functioning (Edwards &
Middleton, 1988); and for building up sound narratives (Ochs, Taylor, Rudolph,
& Smith, 1992; Pontecorvo, Amendola, & Fasulo, 1994).
Children's mastery of conversational practices is a basic requirement for their
participation in social life, and it implies far more than learning word meanings
(Ochs, 1988). Talk-in-interaction is a finely organized practice with properties of
its own that risk being neglected by conceptualizing it as "the serial
externalization of some joint arena of batches of talk, hatched in private (or even
socialized) intentions, and filled out with the docile artefacts of 'language'"
(Schegloff, 1989, p. 140). The conversational machinery provides opportunities
408
Clotilde Pontecorvo and Alessandra Fasulo
for delivery of contributions that the others can acknowledge, ignore, or treat in
different ways, whereas the structure of participation is regulated by the turntaking organization so that every turn at talk must be seen as the joint product of
speaker and listener(s). The management of turn-taking, as well as the
organization of sequences and preferences (Sacks, 1987; Pomerantz, 1978; 1984;
Schegloff, 1990), is an ability that children acquire by various degrees of
participating in interactive talk.
Basically, talk is a means of doing things. Learning it, therefore, makes it
possible to carry out negotiation and to pursue others' approval by building
arguments requiring conversational , rhetorical, and cognitive skills (Duranti &
Goodwin, 1992). Arguing , then, is clearly a demanding task in terms of
perspective taking and reasoning; nevertheless, it occurs very frequently among
children (Corsaro & Rizzo, 1990; Pontecorvo, 1993), with peers (Garvey, 1993),
parents (Caron, 1990), and teachers (Pontecorvo, 1987). Arguing requires walking
on the edge between cooperation and competition (Schiffrin, 1990), defending
one's position by proposing evidence that can be deemed acceptable by the other
conversants. In her analysis of everyday arguments, Schiffrin (1985) remarks on
how the defense of one's stance, even when relying on personal experiences, needs
to replicate to some extent the culturally salient meaning of the debated issue.
That is to say, there is a background of conventional knowledge about how beliefs
can be related to facts in such a way that the connection can be acknowledged and
accepted.
A good argumentative strategy is offered by categorization, that is, the
definition of an event/object as a special case of a given class or as a new/different
case, according to the developing rationale of the speaker. Billig (1987) points out
that this process lies at the core of the argumentative activity. In such a view,
categorization is not a once-for-all, internal assignment of objects to classes, as
conceptualized in the cognitive theories; it is a social endeavor, based on the
speaker 's stance and adapted to his or her discursive aims. It is precisely when the
assignment to a category is ambiguous and problematic that arguing and thinking
start to operate, and divergence in assignment can be the issue at stake between
two speakers.
Categorization and definition are frequently used in children's disputes about
scientific and historical topics (Pontecorvo & Girardet, 1993), as an instrument
for stating a position in an ostensibly neutral and descriptive way. It has been
observed that defining and categorizing are ways of persuading listeners by the
artifactual construction of evidence. Naming is the basic form of categorization,
and there is a sense in which the use of a word or an expression for reference
always singles out a specific dimension from the area of possible meaning covered
by that word or expressi on. Learning to argue is, in an important sense , learning
to categorize according to the context-specific presuppositions and the
interactional goals one is pursuing (Edwards, 1991).
Interactive mechanisms of conversation, such as sequential functions, are likely
to offer some of the central learning mechanisms operating in discursive settings
409
in which adults and children are involved. The sequential organization of talk
(Levinson, 1983), with the set of expectations that it projects, provides a
powerful scaffolding for understanding social meanings and for the transmission of
knowledge.
In a context of disagreement and challenge, accounts or justifications are
expected, and these can be accomplished by different types of discourse products,
such as narratives, descriptions, definitions, and causal explanations, according to
the specific topic at hand or to considerations concerning the audience.
A crucial factor is the disputability of the topic or of the social context that
permits disagreement, opposition, and conflict talk (Grimshaw, 1990) . The
pragmatic function of justifying opposition forces the speaker to reconsider his or
her own previous claim and to turn it into an articulate discursive position. In
classroom discussions (Orsolini & Pontecorvo, 1992), children can be asked to
make their positions explicit, and then they offer adequate explanations .
But how do young children learn to argue? How are they socialized to the
relevant tools for answering and initiating arguments? How do they develop the
reasoning skills? How do they practice arguing in interactional settings in which
adults aim to transmit rules, knowledge , and ways of thinking and behaving?
Family discourse provides interactional settings that precede and accompany
school experiences and where sociocognitive exchanges are carried out by
participants who share a high degree of reciprocal knowledge and of customary
practices. In their heuristic research work, Elinor Ochs and her colleagues (Ochs &
Taylor, 1992a) have analyzed family dinner conversations, showing how versions
of past events are subject to continuous revision and new display of affect by the
participants. In a study on Italian family dinners (Pontecorvo, Amendola, &
Fasulo, 1994), we also found that narrative activities were accomplished with the
contribution of more than one member , with members other than the main
narrator being entitled to do the telling even when they had not been present on
the scene of the narrated facts. Such observations have drawn us to engage in a
deeper analysis of family arguments revolving around the past.
Methods
Aims of the Study and Research Questions
The objective of the present study is to identify the modes of children's

participation in family disputes and the types of argumentative moves adopted in
dealing with this kind of discourse situation.
The following research questions were addressed:
(l) Previous research showed that Italian children like to be engaged in unguided
conflict talk, irrespective of the adult's role (Orsolini & Pontecorvo, 1992), and
410
have an early mastery of argumentative procedures (Orsolini , 1993). Can features

of family-shared discourse contribute to an explanation of such evidence?
(2) Given that criticism and opposition elicit and support collective
argumentative activity, how , in families, are the roles of opposer and opposee
(Orsolini & Pontecorvo, 1989; Pontecorvo & Orsolini, 1993) or of problematizer
and problematizee (Ochs & Taylor, 1992a) organized? Because challenging
concerns not only present behaviors but also the reconstruction of events, what
are the distinctive features ofproblematization in the past?
(3) Because children are more likely to be problematizees than problematizers,
do they learn to argue by being asked to defend themselves or in challenging
others? How do they practice arguing during dinner talk in which the family
reconstructs past problematic events?
We have been particularly interested in identifying the strategies children (and
parents) use to oppose (or problematize) others, to counteroppose an opposer, or
to defend themselves or another opposee. Justification is expected to be the overall
argumentative move through which this aim is reached. And justifications are
accomplished through a range of different epistemic operations (Pontecorvo &
Girardet, 1993), such as reconstructing events, describing, categorizing or
recategorizing, evaluating, and offering details. We discuss examples of these
operations later.
Data Collection and Corpus
Data were collected in the natural context of family dinners. Ten two-parent
families participated. Our selection criteria were the presence of one child between
three and six years of age and at least one older sibling. The families were middle
class and were living in Rome and Naples.
Each family was videotaped from three to four times during family dinner
within a 20-day period. The first videotaped dinner was not used for research
purposes. In most cases, the researcher left the camera with the family and was
not present during recording.
The corpus consisted of 27 dinner conversations of the ten families. The
conversations were fully transcribed using the CHAT transcription method of the
CHILDES system.1 Participants' activities were observed and fully transcribed,
considering (a) the verbal utterances (including repairs, interruptions, overlaps);
(b) the paralinguistic aspects (such as intonation, stresses, pauses, laughs); (c) the
nonverbal behaviors (such as movements, gazes, mimicry).
Transcriptions and coding have been subject to cross-control by at least three
researchers.
IThe CffiLDES system was developed by Brian McWhinney of Carnegie Mellon

University, Pittsburgh, PA, USA.
411
A System for Analyzing Family Conversations

Each transcription was divided into topics according to the discourse object shared
by different speakers in at least three turns. Topics were then grouped in ten
exclusive categories: Free Time, School, Children's Friends, Toys, Food, Video
Recording, Television , Family Events, Parents' Friends, and Persons'
Characteristics. This coding allowed us to regroup talk concerning the same topic
in order to allow specific studies. All the Food topics have been the object of a
study carried out with Elinor Ochs that compares American and Italian dinner
conversations (Ochs, Pontecorvo, & Fasulo, 1996).
Each topic is the basic unit of analysis for which we have analyzed the
discursive roles assumed by the participants. The topic ends when it is interrupted
by at least two turns in which a different topic is discussed. A topic can be taken
up again in a subsequent round during the same dinner. There can be a change of
genre when a topic is taken up a second time. We use the definition of genre
proposed by Bakhtin (1986), and we assume that speech genres are recognizable
and relatively stable cultural forms of discourse.
Genres Within the Dinner Talk

At a more general level, we classified each topic along two main dimensions : (1)
the temporal focus - past, present, and future; (2) the presence or absence of a
problematic event.
The distinction among the three main temporal dimensions was based on the
principal time location of the action/event/feeling referred to, within the shared
topic of conversation. For instance, a topic was categorized as concerning the past
when there were two clauses in the past tense linked by a cohesion marker
(Labov, 1972). These categories, imposed on the text by the analysts, correspond
to "distinct realms of common-sense time for our Western minds" that "fit nicely
into standard psychology, such as the distinction between memories and plans"
(Edwards, personal communication , March 1994). Obviously, our categorization
does not exclude that the different dimensions are intertwined and that there can be
overlapping or slipping from one dimension to another (Ochs, 1994). However,
this categorization mirrors specific discourse genres that have a different inner
organization guiding sequential and rhetorical links detectable in talk.
In identifying a problematic situation, our focus is on problem definition as
rhetorical business, inasmuch as problems are the outcomes of an interactive
activity . In fact, problems that appear in discourse are constructed by specific
discourse moves that need to be taken up by at least one other participant for the
problem to come into existence. In other words, it is not enough that a problem
is raised to consider the topic as a problematical one. Both a problem and its
uptake by someone else - in most cases the one who is problematized - are
412
needed. This is the operational criterion according to which both the problem and
the relative roles are identified.
The crossing of the temporal dimension with the presence or absence of a
problem gives rise to six types of discursive genres. Table 1 shows the
proportional distribution of the topics within each genre.
Table 17.1. Distribution of the different genres according to temporal focus and
problematization (%)
TEMPORAL
FOCUS
PROBLEMCENTERED
NON-PROBLEMCENTERED
PAST
SIDRY
REPORT
7 .2
11 .8
TROUBLESHOOT
BROAOCAST
21.4
47.3
PLAN
AGENDA
PRESENT
FUTURE
TOTAL
TOTAL
19 .0
68 .7
5 .5
6.8
12 .3
34 .1
65 .9
100 .0
The predominant temporal focus in the dinner conversations we studied was the
present (68.7%), followed by the past (19%), and then thejUture (12.3%) . This
overall result shows that the bulk of family conversation is devoted to the ongoing activity of the dinner (although the topic is not always food). Most
conversational topics are non-problem-centered (65.9%). However, the 34.1 % of
problem-centered narratives are particularly relevant because of the conflict talk
that the problematic rendering of actions, conditions, thoughts, and feelings can
induce.
Because the aim of this study is to describe how Italian children practice arguing
and reasoning competencies within the family learning environment, we gave
special attention to the ways in which children challenge other members or build
their defense when questioned.
413
Quantitative Results
Discourse Roles
Following the analysis proposed by Ochs, Taylor, Rudolph, and Smith (1992)
and by Ochs and Taylor (1992b), we identified, within each topic, the discursive
roles of Protagonist (the leading or principal character whose actions, conditions,
thoughts, and feelings are put "on the table"); Introducer (the participant who
makes the first move to open a topic); and Addressee (the participant to whom the
Introducer predominantly addresses discourse) (Pontecorvo, Amendola, & Fasulo,
1994).
In this study, we focused mainly on the ways in which the two discursive roles
of the Problematizer and Problematizee are implemented both by adults and by
children. The Problematizer is the participant who renders problematic an action,
condition, thought, or feeling of a protagonist or narrator. The Problematizee is
the participant whose action, condition, thought, or feeling is rendered
problematic. Actions, thoughts, or feelings may be problematized on several
grounds. They may be treated as untrue or doubtful, or they can be interpreted as
signs of incompetence or as being inappropriate on many different grounds.
We report one excerpt from a family discussion concerning a problematic topic
raised in the present to show to what these roles correspond. Topics that are
problematic in the present concern things that children or other members are
required to do hie et nunc: As in this case, food and nutrition are often in the
foreground.
Excerpt 1
Gravina family : String beans2
Dad: Paolo; Mum: Antonella; Son: Riccardo (Ricky), 7 years old; Daughters: Tiziana,
21 years old ; Silvana, 19 years old
[Arrows after the name indicate problematizing turns; arrows under the name indicate
consequent moves of Problematizee]
nun che buoni ' s t i fagiolini, vero
these string beans are good aren't they?
Ricky-> solo parcha me l i vuoi far mangiare
only because you want me to eat them
Tiziana : [aho! I
heyl
Mum:
2In all the excerpts, the words in italics, which are reported after the invented name of
the family, refer to the name of the specific topic and indicate the shared object of the
piece of talk .
414
Mum:
[no , perche sono buoni]
->
n o , ' cause
they are
really good
Silvana: [infatti. ( .) sono buoni] davvero pero.

indeed they are
really good though
Dad:
mica ha detto a te, a papa

s he mo t h e r wasn ' t talking to you but to dad
questi me l i mangio tutti io
Ricky
che me ne iIrporta a me.
I'm going to eat all
of them mysel f
I don't care takes his

to drinJr from the s traw
toy-glass and starts
eh!i
Silvana : sai che t u sei il campione, ti devi nutri!re
Mum :
y' know you are the

n o u r i s h yourself
champ ion,
you have
to
The problematization here comes from the younger child, who attributes to
mother 's positive evaluation of the vegetables the hidden intention of inviting
him to eat them. The problematic character of Riccardo's claim is evident from
the reactions elicited: Mother and the two sisters simultaneously start talking, the
younger sister with just a reproaching interjection presumably addressed to the
lack of respect shown by the child. The older sister admits the rightness of her
brother's attribution ("indeed") but repeats the mother's evaluation, whereas
mother rejects the child's accusation by stating that her comment was only due to
the goodness of the food.
This density of speaking is typical after problematizing turns, usually
displaying different alignments of the family members with regard to the "hot"
topic. The initial alignment takes place at the turns immediately after the child's
problematization (Tiziana: "hey!"; Mum: "no"; Silvana: "indeed"). Silvana later
adds an element to show her brother that his attack was unjustified (i.e., that
mother was talking to father and not to him) . And father readily agrees,
announcing his intentions of eating all the vegetables. In response, the child
reveals that he understands that he is the intended recipient of father's little show,
by remarking that he does not care. Again the child is warned, this time by
mother, to be more respectful. Silvana eventually cheers him up with reference to
his swimming activity, at the same time confirming the family intention of
making Riccardo eat and providing a reason for that.
We can observe here the complexity of the cognitive task posed to the child: He
receives both denial and confirmation of the content of his claim, plus repeated
reproaches to the way he addresses his parents. In these few lines, the social
organization of challenging authority is visible, as is the peculiar fact that the
offended authority does not defend himself or herself, but this is done by another
member of the family, who is lower in the hierarchy (the sister for the mother,
the mother for the father).
415
We now focus on the distribution of the family members in the discursive roles
of Problematizer and Problematizee (see Table 2).
Table 17.2. Problematizer Versus Problematizce (%)

PROBLEMATIZER
PROBLEMATIZEE
FATHER
26 .1
18.2
MOTHER
29 .8
24.2
CHILDREN
40.9
48.6
As expected, children tend to be more the Problematizee (48.6%) than are parents
(42.4%), but the difference is not very large. Furthermore, children (at least two in
each family) are the Problematizer significantly less often (X2= 4.93, p<.03) than
are parents (40.9% vs. 55.9 % of the parents); however, they assume this role
more often than we had expected on the basis of similar American data (Ochs &
Taylor, 1992b) . In other words, Italian children problematize parents more than
American children do.
Problematizing Roles in Talking About the Past

We now look at the proportions of the Problematizer-Problematizee roles in
stories (see Table 3).
Table 17.3. Problematizer Versus Problematizee in the Past (%)
PROBLEMATIZER
PROBLEMATIZEE
FATHER
25
16.1
MOTHER
37 .5
14.3
CHILDREN
31.3
48 .2
In this discourse context, children problematize much less (31.3%) compared with
the general distribution presented in Table 23 and with a ratio of 1 to 2 compared
with parents (62.5%) (X2 = 39.82, p< .OOOl). The increase in the parents' role of
3The proportion of children 's problematization slows down from present to past, with
future in the middle ; that is, Incidents = 44.5; Plans = 39.5; Stories = 31.3 .
416
problematizers is totally due to the increase in the mothers' share. Mothers , who
already were higher than fathers in problematizing (29.8% vs. 26.1%), increase
about 8% here. Overall, more than one third of the problematization in the story
genre is done by the mothers.
These results differ from those obtained in similar American research (Ochs &
Taylor, 1992b) where problematizing seemed to be done mainly by the father. In
the Italian families, we observed that there is more problematizing activity carried
out by children and by mothers, frequently in oppositional talk reciprocally
addressed. If we consider "who problematizes whom" (see Table 4), there emerges
a possible correspondence and reciprocity between the problematizing talk of
mothers and children.
TABLE 8.4. WHO PROBLEMATIZES WHOM IN THE PAST? (%)
PROBLEMATIZER
PROBLEMATIZEE
FAlHER
CIffi.DREN
OTlUERS
MOTlUER
SElJP
13 .4
6 .2
4 .5
CIffi.DREN
FAlHER
OlHERS
SELF
17 .9
10 .7
8.9
SmUNGS
12.5
9.8
4.5
3.6
(25)
M01HER
(37 .5)
CIffi.DREN
(31.3)
MOlHER
FATlUER
OlHERS
SELVES
0.9
0.0
0.9
[OlHERS
(6.2)]
Within problem-centered past narratives, in fact, mothers are problematized by

children twice as often as are fathers; and the reverse is also true: Most
problematization by the mother is addressed to children. Parents play the
problematizer role more frequently; but in our sample, mothers do it much more
than fathers. This happens mainly in the activity of joint past reconstruction in
which mothers appear to be involved very frequently. In the context of Italian
families, this cannot be interpreted only as a sign of power, as in Ochs and
Taylor's (l992a, 1992b) interpretation of the U.S . data. The "political"
explanation defining the subordinate position of the problematizee versus the
problematizer does not exhaust the observations drawn in our corpus, where
mothers' (but also fathers') questioning can be read as the signal of an intense and
emotionally laden exchange that, in the Italian culture, conveys a function of
involvement and encouragement to participation.
417
We hypothesize that there is a relevant learning mechanism embedded in the

ways children are questioned and consequently requested to justify themselves. In
general terms, they are offered the possibility of revisiting the past, and they
receive an oppositional attention that is not valued negatively in the Italian
culture (Pontecorvo, 1993). In particular, when children are directly problematized,
they are often offered the cognitive frame and the linguistic form in which to
formulate their accounts, justifications, or explanations. The problematizing acts
can also give cues for counter-arguing, and children learn to account for their
actions and thoughts by following recipient-design patterns of talk (Schegloff,
1989), that is, by presenting justifications in ways that can be satisfactory to the
parents .
Arguing About the Past

Challenge and Defense
Recalled past events possess only loose connections with the immediate action.
For this reason, the available deictic referents are scarce, and the discussion
requires more elaborate processes of topic introduction and sequence tying.
Moreover , when the past is rendered disputable, the reconstruction of facts and
mental states becomes the object and tool of challenge and defense. Challenge and
defense are complementary conversational moves that can be accomplished
through different argumentative choices: Deploying reconstruction as a rhetorical
device involves presenting elements of the past, such as setting , characters,
mental states , and affective reactions. What is accomplished is the recreation of
colorful and often emotionally connoted scenarios in a discursive display of
truthfulness (see Edwards & Potter, 1992). Reconstructing always implies
definitions or categorizations of the events talked about.
Challenge and defense are characterized as follows.
Challenge : It is enacted by the Problematizer in the first place and can concern
the level of facts and the level of discourse (Labov & Fanshel, 1977). It is
typically introduced (in Italian) by the marked use of the personal pronoun or by
discourse markers such as rna [but] (Schiffrin, 1987). It can be pursued through
transversal tactics when the addressee is different from the problematizee. Other
modes include prescriptions (transmitted by the use of modal verbs of obligation) ,
proscriptions (referring to things that should not have been done), and use of a
counterfactual strategy when alternative ways of behaving are proposed.
Defense: It can be enacted by the Problematizee, by his or her allies, or by the
Problematizer when counter-challenged. It is achieved by offering reasons,
motives, and explanations; referring to alternative versions of events; offering
detailed accounts of them; appealing to authority or eye-witnesses; and changing
418
the script andlor the categorization of the recounted facts. The discourse context
laid down by the challenge, although constraining the possibilities of
argumentative response, also offers elements than can be used for building the
defense.
The following is a qualitative analysis of a set of stories extracted from a corpus
of 75. The selection is meant to reproduce instances of different patterns of
participation and argumentation.
(a)
Peripheral participation
(b)
Challenge
1) being drawn in by others
2) building on another's introduction
3) introducing and challenging
(c)
Defense
1) redefining and recategorizing events
2) counter-opposing the Problematizer
Learning to Challenge
Peripheral Participation
Children are frequently the silent protagonists of parents' conflict talk. They are
referred to in the third person, although present, and their role is that of peripheral
participants (Lave & Wenger, 1991). We also find this kind of talk in the
aftermath of parent-child discussions, when parents continue the discussion
between themselves, often debating educational-organizational decisions (see
following excerpts : Violin and Coke).
In Excerpt 2, the child is used as prosecution evidence of guilt for which both
parents attribute responsibility and carelessness to one another.
Excerpt 2
Soldano family: Hands
Dad: Vittorio , 35 years old; Mum: Raffaella, 35 years old; Grandmather; Grandfather ;
Sons: Gianluca, 4 years old; Stefano, 8 years old; Daughter: Marina, 3 months old
Mum : - >
Dad:
non hai visto che rrani che c'ha Gianluca ?

didn't you see GianI uca' shands?
te l' ho detto prima non hai sentito ?
I told you before, didn't you hear i t

Mum:
si rna (
yes
419
but ...
Dad:
quando 1 ' h a i sentito gliele potevi lava te
->
well
you
could have washed them when you heard
allora scusa eh io l'ho detto qua a tavola

then excuse me uh I
said it here at
the table
In the preceding short exchange between the two parents, about the dirty hands of
the child, they each discharge their own responsibility for washing the child's
dirty hands , and they each justify themselves by attributing the fault to the other.
The argumentative strategy used by the two parents is similar and is based on
displacing in a former past (as shown by the repetitive use of the declarative and
perceptual in past forms of verbs: see, say, hear) the negligent behavior of the
other and the origin of the present problem.
Being Drawn in by Others

The quantitative distribution discussed earlier confirms that children are
proportionally more the Problematizee than the Problematizer in the past.
However, they can be engaged indirectly in the problematizing activity by an
adult, as is the case in the next excerpt of the same family as in Excerpt 2. The
behavior of the father is indirectly questioned by the mother , who addresses the
child with a marked use of the personal pronoun and involves him in the
reconstruction and evaluation of the event. Father 's action is problematized first
by the mother and then by the young child because of its collective negative
consequences (father has eaten a piece of food that was being saved for breakfast).
Excerpt 3
Soldano family: Pizzetta [Small pizza]
Dad: Vittorio, 35 years old; Mum: Raffaella, 35 years old; Grandmother; Grandfather;
Sons : Stefano, 8 years old; Gianluca, 4 years old; Daughter: Marin a, 3 months old
[Gianluca is trying to save the last sweet on the serving plate for his father while
someone else is picking it up. The mother and the grandparents try to dissuade him and
tease him about that.]
Mum: - >
tu hai voluto conservare la pizze :tta per papa=e

you wanted to
save the
small pizza
>papa s'e mangiato 'a crosta:ta< .

and daddy has
eaten the
cake.
tu non j e conserva Ie cose

don't
save
things
for
him
for daddy
420
Dad:

(Oche cosa ha conserva- che pizzetta ma- ")
what did he
( (to Mum
Mum:
save- which
orne la so' mangiata i : (h)
small pizza did he
(h)
I ate the small pizza to Dad, whispering

as trying not to be heard by the son
Grandpa:
Dad :
[hahahahah
huhuhuhuh
[ah ve:di
s 'e rnangiata la rnia pizzetta.
Uh you see,
she ate my
small pizza
Gianluca: no: [perche tu ti sei rnangiato

no: because you ate to father
Mum :
[rnh erano sala : te,
they were too salty to Dad
Gianluca : [la rnia cros- la no :stra
->
my-
Mum:
[delle volte guarda compri la roba in pasticceria
our
sometimes you know you buy things
[e salata (
that
at
the pastry
are too
turning
salty
to
Granpa
Gianluca : [crosta :ta

cake
Dad :
[rna lei- ma lei prima si e rnangiata la pizzetta,
->
but
she had already eaten the
small
pizza.
io non ho trovato la pizzetta, c'era solo la crostata

I didn't find the
the cake
small pizza.
There was
only
e rni so'rnangiato la crostata.

and I
Mum
ate the
cake.
non t' e venuto in mente che c ' erano due f ettine

it didn't cross your mind that there were
two slices
[preparate la sopra per la colazione

prepared over
Grandma :
but
Dad
Dad :->
for
breakfast
didn't
you
leave
him anything
dinner?
what crossed my mind was
[a chi?
to whom to Grandma
Grandma points toward Dad with her head))
vedi che gentili m'hanno lasciato la crostata
look how kind they were:
Mum:
for
no [m' e venuto in mente ( .)

no
Mum:
there
[rna non je : l'avevi lasciata la cena?
they left me
the
[io je lascio la cena a Vittorio?

should
leave
the
dinner
for Vittorio?
cake
421
This layered problematization carried on by the mother is apparently addressed to

the child , who insists on saving food for his father (who, it turned out, does not
consume meals regularly, when sitting with the others at the table, but has snacks
all day). The implication is that the child's kind thought is wasted on the father,
because not only did he not eat the pizza that Gianluca had left for him the
evening before, but also he ate the cake that was saved for breakfast. Evidently,
the father is the target of the problematization . He engages in defense, while the
child joins the mother as the offended party by recycling her challenge. The
problem at the core of the mother's move is dealt with by the father, who says
that he interpreted the slice of cake found on the table as a generous and kind act
from the others, thus rejecting the main accusation of insensitivity.
Grandmother's attempt to readdress the problematization to mother, by asking
whether any dinner was left for father, is readily rejected by her as out of place.
Building on Another's Introduction

Through a challenge, children can insert themselves into a discourse on past
events opened by someone else, as in the following excerpt:
Excerpt 4
Selci family : Gigi
Dad: Francesco; Mum: Monica; Aunt: Stefania ; Daughters : Manuela, 10 years old ;
Federica, 7 years old; Francesca, 4 years old; Son: Alessio , 1 year old
Mum:
Aunt :
Gigi dov'era?
where was Gigi?
a casa .
at home.
Mum :
immancabile eh
Aunt :
as always uh?
mhm .
Mum finishes
up
the food on the plate of
Federica) )
Federica : [grazie mamma
thanks
Mum:
mum
[col cane?
with his
Aunt:
dog?
eh .
yeah
Federica : Manuela !
Mum:
e col gatto?
and with the cat?
Aunt :
Aunt's
[vanno d'accordo]
they get
along well
cat
422
Manuela:
[Stefania t'e ]simpatico Gigi?
->
Mum :
Stefania do
you
like
Gigi?
che dici?
what
are
Aunt:
( (shakes
Manuela:
a me no
Aunt:
no?
you
saying?
head
as
in
"so
and
SOH
don't
no?
Federica: a parte che Gigi e tuo padre

->
not
Aunt:
non e 000 padre
to
say that
->
he
Gigi
is your father
is not my father
Federica: strano
Aunt:
Manuela :
strange
mh
perche non si e sposato con mamma Rori

because
Aunt :
Manuela:
he
mamma
Rori se vuole
mamma Rori
Aunt :
Mum:
didn't
marry mamma Rori
no .
wants
sposa z
to
get married?
no .
[e fidanzata
she's
engaged
Manuela:
[Gigi?
Federica : invece si e sposato con
he
mamma
Rori
did marry mamma Rori
In this story, the two older sisters use a joint strategy for problematizing the
young aunt. Their topic is the live-in partner (Gigi) of the aunt's mother (mamma
Rori), and they like talking about this delicate issue.
The sisters use the apparently neutral format of asking informative questions,
offering personal evaluations, and stating matters of "fact." It becomes evident
that they know very well the answers to their questions and the falseness of their
asserted facts (the aunt's mother and her partner are not married, and the partner is
not the aunt's father) . The sisters' goal of challenging their aunt's known but
unadmitted view about the situation is highlighted by the mother's first attempt
to stop Manuela (what are you saying?) and then giving a socially acceptable
definition of the relationship (she's engaged).
A similar situation in which the two sisters cooperate in problematizing adults
occurs in the following excerpt drawn from conversations of the same family. The
youngest daughter introduces a topic about her baptism and seeks information
about her godmother; then the oldest daughter (who is ten) problematizes the
mother's choice of her godmother by questioning the choice of the person (who is
not highly valued by the mother herself) and complaining about being held by an
423
unknown person . Federica uses the opportunity to criticize the mother who,
according to the girl, was very badly dressed at her baptism.
Excerpt 5
Selci family : Baptism
Dad: Francesco ; Mum: Monica; Aunt: Stefania; Daughters: Manuela, 10 years old;
Federica, 7 years old; Francesca, 4 years old; Son: Alessio: 1 year old
Federica:
Stefania e vero che tu sei la mia madrina?

Stefania is
Aunt:
it
true you are my godmother?
la tua madrina di che?

your
godmother of what?
(4 .0)
Federica :
(mo' tra l'altro cara)

now among other things
Mum:
of
Aunt:
baptism
di battesimo?
of
baptism?
Mum:
tene sei dimenticata .
Aunt :
did you forget it?

no non rni ricordo
no I don' t rememeber
Mum:
[te e Ugo] .
you
Aunt:
[ah
and Ugo ...
e vero] si
oh yes
Manuela:
dear ...
di battesimo .
that's
right
e invece a me chi e stato?

and who
has been mine?
(2 .0)
Mum :
un'arnica di mamma Rori. meglio perderla che trovarla

a friend of mamma Rori
to find her
Manuela : - >
better to
lose than
se : : ( .) a mamma perche a me I'arnica di mamma Rori ?

yeah mum why mamma Rori' s
friend
for me
Mum :
perche l'hanno voluta fare loro .
->
because
Manuela :ah.
(5 .0) ri o stavo] in braccio a una sconosciuta?
->
Aunt :
Manuela:
ah ,
mamma!
thay wanted her

was
in the arms of an unknown person?

{(to Mum
[Marilu ?]
mummy
Mum:
->
stavi in braccio a me no a una sconosciuta

you were in my arms, not in an unknown
person's
(2.0)
424

Fr anc e s c a pul i s c i t i le mani col tovagliolo
Mum:
Prancesca
clean your
hands
with the
napkin
Federica: - > eri vestita rnal issimo al mio battesimo

you dressed awfully
ah si?
oh really?
Mum:
Federica:
for my
bapt ism
con un cappetto almeno:

wi th such a coat
Mum laughs
As in the preceding example, when children challenge their parents, they often
advance complaints about the ways in which they have been treated. It is a kind of
appeal to children' s righ ts, which have not been properly attended to by the
parents.
This is also the case in the next story, a very short one in which the seven-yearold daughter questions why her father has not taken her to a nice small woods
where she would have liked to go with him.
Excerpt 6
Selci family : Nice small woods
Dad: Francesco; Mum: Monica; Aunt: Stefania; Daughters: Manuela, 10 years old ;
Federica, 7 years old; Francesca, 4 years old; Son: Alessio, 1 year old
Dad:
so' andato a prendere la terra

I
we n t
to get
so il
Fede rica : s'e' fatto i l boschetto ora?

has
Dad:
the
small
woods
g rown?
eh qua vicino
just
in
the
surroundings
Federica : rna perche non me 10 dicevi ci venivo anch 'io

->
but why didn 't

too
you
t el l
me;
wou ld
have
come
non ci sono rnai andata [in t utta la mia vita

I
never went
Aunt:
there
in all my life
[dammi i l piatto Manuela

give me
your plate Manue la
Dad:
stasera, andiamo,
( .) un'altra volta
Aunt :
this evening we
mhm
Dad:
quando vado in queste cose mi sembra di esse un ladro
->
when I
go
some other time
(1. 5)
go
for t h e s e
things
feel
like a
t hief
425
In the last two turns, the father justifies himself both by prormsing and by
furnishing a kind of justification by evaluating his uneasy state of mind in going
to the wood and taking some soil.
The stories of this particular family (Selci) show a high proportion of child
problematization. The children, three plus a baby, are problematizers four times in
ten stories and are often successful in constraining the adults to justify
themselves.
Introducing and Challenging

In relatively few cases 00 in 75 stories), children are both introducers and
problematizers of parents or siblings. Here we provide an example of this
situation.
In Excerpt 7, the 7-year-old son (the eldest) begins by reconstructing a summer
holiday scene:
Excerpt 7
Fanaro family: Brood Hen
Dad: Silvano , 34 years old; Mum: Teresa, 34 years old; Daughter: Stefania, 5 years old;
Son: Sergio : 7 years old
Sergio :
papi ti credo, (0 .2) che Ii rnangiavamo cosi tanti .

daddy no wonder we
were eating so much
sta: andavamo=la rnatti :na andavamo serrpre al mare e

we
went
to the beach every morning and
uno al rna:re , ( .) [spende energia

at
the beach one burns up energy

[ DC'
Mum
ha f ame?
gets hungry
Sergio :
a f orza di nuota-=nuota . (.) nuo t a =nuot a .

by swimming and
swimming
->
Sergio:
swimming and
be tu t' 'a conservavi perche stavi cosi seduto

to Dad
were saving it because
well you
you were
sitting this way
come 'na: sernbravi

Dad :
Mum:
swimming.
~natra.
like a.- you looked like a duck.

mh?
mh mh,
( (they both continue looking into
plates) )
their
no v er ament e . stava seduto su- su queIIe sedie che

no really he was sitting on:
chairs
on one of those
426

s i aprono e s i chiu :dono .
Marmna. :
Sergio:
that open and close to Mum} )

( ( r ep r o d u c i n g the movement of the
bends toward Dad)}
rnhrn
mhm looking up to Sergio
sembravi n' anatra .

you
Dad:
chair he
looked
like
duck
chi io ?
who me
Sergio:
eh . stavi cosi
ah you were going like this stretching his
neck and turning his head left and right
Da d :
e certo . ( . ) dovevo fa' la chioccia=dovevo fa ' .
->
of
course
had to do the brood hen I
had
perche dovevo s ta' a ttento a yoi, la mamma non

because
there,
had to
look after you muJlllllY wasn ' t
c'e :ra la chioccia la f acevo io .

I
Mum:
had to do the hen .
ti ricordi che non io so ' venuta ?

r e me mb er I d idn ' t come? to Sergio } )
infatti mi so' abbronzato praticamente di qua
do you
Dad :
indeed I got tanned practically here

di qua sotto le ascelle f ino a giu come stavo
and here from my arms down I was
sempre in piedi: sul mare no .

always
Mum:->
stand ing ,
never
in
the water
comunque sei stato meglio te che io

howeve r
you had a
better time
than
che so ' stata a Rorna da sola .

who was
in Rome
alone
Dad:
non credo proprio guarda
Mum:
non credi proprio?
really don't
you
Da d :
think
really don't
so
thi nk
so
[no non credo

no
don't
Mum :
[s ap e s s i che - che tristezza sta' qua da soli .
Dad:->
rna io me sarei riposato a sta' da solo qua guarda .
if you knew how sad it

but
would have
is to be at
rested being here
home alone
alone
Here, the child teases his father for his ridiculous position at the beach, describing
the scene and even imitating him to the amusement of the whole family (in his
second turn, he addresses his mother ). The father reformulates the metaphor,
changing it from duck to brood hen, a smart move that has the effect of
427
transforming an image of voluntary and lazy sitting into one of forced and busy
watching. At the same time, it introduces a reason for the complaint about the
mother's absence.
In what follows, the ironical dispute is transferred smoothly to the two parents,
whose topic is who "suffered" more during summer : the mother, who was at
home alone and who worked, or the father, who looked after the children at the
beach. In the last part of the story, "other voices" (Fasulo, this volume) of the
people are reported by both, as a backing for the counter-opposed interpretation of
the situation they are proposing. In this last part, the children are exposed to
alternative and conflicting ways of reconstructing and evaluating the same
situation in the past. Through peripheral participation, they are exposed to a nice
dramatization of parents' sacrificing for children's good.
Learning to Defend
When children are the direct object of challenge, they can decide to defend
themselves. We have identified the two main ways of doing this: offering an
alternative interpretation and thus denying the content of the challenge, or counteropposing the Problematizer. Both strategies can involve reconstruction of past
events .
Redefining and Recategorizing the Situation
Excerpt 8
Fanaro family : Quarrel
Dad: Silvano, 34 years old; Mum: Teresa, 34 years old ; Daughter: Stefania, 5 years old;
Son: Sergio : 7 years old
MLun :
oggi corn'e anda- eh a proposito corn'e andata oggi da:

how was- uhm by the way,
how was today
da Daniele e Fausto ?
with
Sergio:
Daniele
and
Fausto?
bene !
very good nodding

Stefania: - > no abbiarno litigato un pochetto noi due.
Dad:
no we had a
mhrn
MLun :
mhm
t u e Sergio?
you
and
little bit of quarrel us two
Sergio?
428
Stefania :
Dad:
ell
eh.
[corne rnai?
how was
Mum:
it?
[e perche?
and why?
Dad:
alza la voce Stefa
Stefania:
perche: :: parche lui faceva gli scherzi quindi (.)
speak
->
louder
because: ::
Stefa
because
he
was
joking and
poi mettevano tutto per l'aria poi quindi

then they were making a mess then so
Antonietta s'e arrabbiata con loro, ( .) con lora due

Antonietta was
of them
angry with them with
the
two
con me no perche dopa io so' stata brava

Mum:
Stefania:
with me no because afterwards I have been good

mhm rna come mai? lora che combinavano ?
mhm but how was it? what were they doing?
eh mettevano tutto per l'aria li vedevo (
eh they were making a mess, I saw them

perche dicevano cosce: cose ell .
because they were saying things: things
eh
Dad:
a te? come i bambini di un anno allora
Stefania:
mettevano- (.) allora tiravano tutto per aria tutto:
to
you?
as
one-year-old
children
because they were making a

everything
then?
mess with
Sergio :
si ti sembra come ti racconta che sta a fa' : presa-
->
yes it looks as
she was doing
though she
is
telling what
ha presQ la chitarra- ha presQ la chitarra quella

she took the guitarsmall
she took the
guitar the
piccola e ce l'ha buttata in testa cosi boing baing

one and she threw it on our heads this way
boing boing
Mum:
insomma non siete stati proprio santi insomma oggi
->
well
you
have not
really been saints
today
eh ? io immaginavo invece che facevate i bravi che

eh? instead
good that
figured
that you were
being
non facevate arrabbiare ad Antonietta no?

you were
Dad:
->
not making Antonietta angry no?
eh vi abbiamo raccomandato di fare i bravi

eh we
have warned you to be good
[vedo che ci ascoltate tanto]

I
see how you listen to us

Mum :
->
[nonostante tutto siete] riusciti pure a farla

notwithstanding all you succeeded in making
her
arrabbiare mhm ? metti la mano qui sopra poggia
angry too mhm ? put the hands on here leave
tovagliolo [(
the
[si rna questa volta abbiamo detto la verita!

yes but this time we
->
the
have told
truth
soprattutto Fausto
in
Mum :
)]
napkin
Stefania :
Sergio:
429
particular
Fausto .
robe Sergio chiunque sia non siete stati tanto bravi

well Sergio whoever has
been so good
been you have
not
questa l 'ho capito .

I
understood
this
Here Stefania, the younger child, expresses opposition to her brother's laconic
account of the morning's events by initiating her turn with a no, followed by a
radically different account. The parents, speaking in complete overlap, ask for
clarifications, and Stefania answers by recalling some "facts" that clearly put the
responsibility on the brother's and cousin's side. In support of her thesis, she
offers the evaluation of another adult (their aunt) who got angry with them. She
builds a report clearly designed to obtain the consensus of her parents (i.e.,
recipient-designed: Schegloff, 1989). Because her mother asks a question and her
father makes a comment, both seemingly accepting her version (Mum: "why?
what were they doing?" and Dad: "just like one-year-old children"), the brother
reacts by challenging his sister's version as biased and introducing a new piece of
evidence about her questionable behavior. The parents, summing up the reports in
a reproachful tone addressed to both children, provoke a change in Stefania's
strategy: leaving aside the past facts, she casts a positive light on herself by
defining her present behavior as praiseworthy, using an appeal to a more general
set of rules ("at least we told the truth"). Although she includes her brother in this
clever escape, she was the one who decided to tell the parents about the quarreling,
despite Sergio's reassuring answer .
From the children's talk, we can infer that they are completely aware that
telling is not an innocent activity or a straightforward mirroring of reality. The
children know that what they say can meet the sanction of parental judgement, and
this knowledge is detectable as a built-in feature of their accounts. Stefania's last
move shows a deep understanding not only of the moral values but also of their
hierarchical order: Telling the truth is a morally superior value to behaving
properly, something that is also implicit in the Catholic culture, where the
practice of confession warrants forgiveness for one's sins.
430
In the next excerpt, a new and different definition of the problematic situation is
offered by a four-year-old girl who is problematized by her older sister. The
mother supports the claim of the sister, although apparently keeping a more
neutral role. The justification strategy of the younger child is to redefine the
situation by partially keeping the same objective tone used by the mother. In this
very short story, the redefinition of the situation is an epistemic operation carried
out by the four members of the family, who each look at the same event from a
different point of view.
Excerpt 9
Bianco family: Police
Dad: Giovanni, 42 years old; Mum: Sara, 33 years old; Daughters: Silvia, 4 years old;
Fabiana , 9 years old
Mum: -> Silvia raccontaci bene quello che stavi facendo

Silvia,
tell
us
precisely what
you were
doing
con Clara in stanza prima che venisse su Fabiana

with Clara
in your room before
Fabiana
came up
Dad:
()
Silvia : stavamo giocando a polizia e poi uno c'ha
->
we were playing police then someone
disturbato ch'eri tu .
came to disturb us who was you
Fabiana : ti stavi per rompere la testa ho disturba :to

you were just
disturbed
about
breaking your head.
(4.0)
Dad:
ma io 1 'ho lasciate che volevano farsi un riposino

but when I
left
them they wanted to take
nap
Fabiana: no : :
Mt.un:
a dis fare i letti
to undo
the
beds.
The problematizing nature of mother's question is revealed by her request to tell

''precisely'' what the child was doing and by defining accurately the point in time
for which the account is asked. This format suggests that a little drama must have
happened after the older sister, Fabiana, went up to the room. Apparently all the
participants already share a certain amount of information about the facts.
A resolution of the negative affective outcomes of that past event seems to be
sought in the ongoing talk. This is, in fact, a distinctive character of discussion of
past events during family reunions, although often the problematic nature of them
is defined as such only in the present, as in the Baptism story, in which the child
could not possibly have realized the questionable clothing choice of her mother at
the time it happened (see Excerpt 5).
431
The child responds to her mother's challenge with a report on the activity she
was engaged in with her friend Clara: "we were playing police ." This is an act of
categorization that contains a justification. Fiction games have features of their
own. They have to be consistent with the reality domain they are assumed to
imitate: If you play police, you cannot sit down and talk quietly. The temporal
construction is also significant: The play is rendered with an iterative form, and
the precipitating event is represented by Fabiana's coming in and disturbing ,
reported with a punctual verb tense. By this temporal construction, the
problematic focus is transferred from the activity to the interruption and from one
daughter to the other. Fabiana, called into question, substantially changes the
picture, defining her entering as the providential avoidance of an impending
dangerous event: namely, her young sister's breaking her head.
As in the preceding example, we have here a contest about different versions of
a past event in which the contending parties were both present, with the
problematized party objecting to the format of the report, as shown here by the
ironical reported speech that follows the redefinition ("I disturbed"). A parallel
exchange takes place in the following turns between father and mother: He reports
the account he had been given before leaving the children alone in the room. again
an innocent one (they wanted to take a nap), and she translates it at the action
level (to undo the beds).
Confronting this episode with the preceding one, we observe that a challenge
about a similar troublesome event (children playing in a far too energetic way)
elicits different justification strategies: Sergio, the child of Excerpt 8, when
challenged, goes down as a first step to the factual level, providing factual
information from which the desired inference can be drawn by the parents .
Conversely, in Excerpt 9, the questioned conduct is accounted for with a formula,
the name of an activity that involves a certain kind of action but does not require
describing them. Some data from social cognition studies suggest that a situation
of challenge provokes a retreat to the descriptive level (Semin & Fiedler, 1988).
This certainly happens, in that rooting a claim in a detailed narrative is an
effective device to achieve an appearance of truth. Nonetheless, there can be
discourse situations in which the opposite strategy can turn out more convenient,
as in our example. We are closer, therefore, to Billig's position (1987), inasmuch
as a differential categorization is always a choice and responds to the rhetorical
demands of the discursive occasion.
In accounting and justifying, children can also use the frame offered by the
Problematizer, as in Excerpt 10 :
Excerpt 10
Aloisio family : Coke
Dad: Marco ; Mum: Laura; Daughter: Gaia, 6 years old; Son: Adriano, 11 years old
Dad:
rna la Coca e finita ?

But is there no more Coke?
432
Gaia:

seh : : .
yeah
Adriano : - > Gaia 1 'ha finita tutta . oggi porneriggio 1 ' ha aperta
Gaia has finished
opened it
it
up.
This
afte rnoon
she
e se l 'e finita tu:tta qua :nta !
and
she
finished
it
al l.
Dad :
Gaia ? guarda che e tanta quella=quella cosa di
->
Gaia
look ,
it
is a
l ot that
that muc h of
[Coca Cola eh?

Coke
uh
Mum :
[e ti fa male
->
and
Gaia :
no guarda (
no
Adriano :
it's
bad for your heal t h
l o ok
e io non rni so' bevuto pili nient e .

and
couldn ' t
drink
anything
afterwords
Gaia :
papa r : papa : guarda che ho ini zia t o starnatti:na
->
daddy daddy
Mum:
[capi :to
you
Dad:
began
this morning
see :
[peggio rni sento .

that's
Adriano :
look
even worse
eh infatti
yeah exac t ly
( ( Ga i a laughs
Adri ano :
e ride lei
and
she
laughs
Mum :
quando ti dico di non prenderla perche (
->
when I t o l d you n o t
it ' s too :
to buy it
) e troppo :
' cause
( .. . )
Dad:
no perche nessuno co nt rol l a evidenternente
->
no
Mum:
eh : : no anche perche se anche volessero non ries cono.
' caus e
well
Dad :
nobody looks
no because
after her
even i f
c l ear l y
they wanted they can't
addirittura?
really?
Mum:
Adr iano :
eh : !
a f are che ?
to do what?
Mum:
a controllare Gaia nel senso che [Gaia e un po' :]

to
look after Gaia
, cause
( ( laughing) )
Mum:
Adriano :
Gaia
is
somewhat
fe' vero ]
Gaia :
that 's
tru e
he he
he he ci ride pure capito
she laughs too , y'see
i mitating Ga ia
( ( t h ey
both
laugh
433
The father opens the problem by addressing a question preceded by a rna [but],
which in this case (and in many others we found in our corpus) is neither a
contrastive move to previous talk nor a way of blocking a possible inference.
Rather, it is a way of initiating a problematic issue, probably based on a
previously unexpressed assessment of a state of affairs. The older child, Adriano,
picks it up by accusing Gaia of having drunk all the Coke and touches on the
temporal dimension by specifically indicating when his sister opened the bottle of
Coke. After the father and mother have expressed their disapproval, Gaia selects
the temporal dimension to make up a defense, correcting her brother and claiming
she began drinking the Coke beforehand, in the morning. Her justification
strategy involves the reconstruction of the event, aimed at getting rid of the
attributed behavior of "drinking so much Coke in so little time." Here the case is
especially interesting, because her justification bumps into another blameworthy
conduct: "drinking Coke in the morning." Her false step is commented on in a
critical chorus, and the girl ends up laughing at herself.
In the second part of the story, the focus shifts to "who buys it" (the mother
reproaches the father) and "who should control the children" (the father reproaches
the mother). The shift from the level of facts (there is no more Coke) to the level
of rules and values involves all the family roles, and the children find a free space
here between the parental authorities. They learn to use it in arguing, as can be
seen in other stories (e.g., "Violin ," Excerpt 12).
The Problematizee, particularly a child, can use the same frame of discourse
proposed by the Problematizer in order to counter argue, as is the case in the
following excerpt.
Excerp t 11
Soldano family : Ghosts
Dad: Vittorio, 35 years old; Mum: RaffaeIla , 35 years old; Grandmother; Grandfather;
Sons: Gianluca, 4 years old ; Stefano, 8 years old; Daughter: Marina, 3 months old
[Stefano refuses to go to bed without his brother because he is afraid of ghosts . He says
tha t, in his friend's "Handbook of Ghosts," it is written that Rome is the town most
populated by ghosts 1
Dad : - >
scusa Roma
excuse me
tanto popolat a dai f ant a smi vero?
is Rome
so
full
of ghosts,
is it?
quindi chissa quante volte l i hai incont r a t i senza

then pe rhaps
you met
them many times without
saperlo t'hanno mai fatto male? a te hanno fatto

knowing it ,
been
d id they never hurt you?
male i fantasmi ?
hurt
by t he
g hosts?
Stefano :
una volta sono inciampato nel niente
->
once
stumbled on t he
nothing
have you
434
Dad:

e quello e stato il niente mica i fantasmi
well
that was
nothing not
the
ghosts,
[canmina dai
come on
go,
Mum :
[dai Stefano su
Stefano:
no c'ho paura vieni?
come
no I
Dad
on
stefano
am afraid,
do you come with me?
eh Stefano su!
come
on
Stefano
Gianluca: tanto Gianluca ha finito deve venire con me

anyway Gianluca
with me
Dad:
has
finished,
he must
come
ci pensi? stai colla telecamera accesa

can you
imagine? you have a
camera on you
col microfono li e qualcuno sentira che un bambino

wi th the microphone
child
and
someone'll
hear that
di quasi otto anni ridicolo ha paura dei fantasmi e

of nearly eight
ghosts
[(
-ridiculous-
he
is
afraid of
Mum:
[e vuole aspettare Gianluca
Dad:
perche da retta alle stupidaggini che gli racconta
->
because
Marco
and wants
he
to wait
for
listens
to
Gianluca
the
foolish
things
that
Marco che Ie ha viste su un Iibro che ( .) e fatto

tells him who has
seen them in a
book made up
per dire le cretinate ai bambini cosi se 10

to
tell
stupid things
to
children
so
comprano e loro guadagnano [un sacco di soldi

they buy it and they earn a
Stefano :
->
Dad:
lot of money
[c'e- c'e pure una foto

there-
there
is
even a
photo
la foto di un fantasma? allora si vedono perche le

a photo of a
because
ghost?
then they can be
seen
foto fotografano solo le cose che si vedono

photos
Stefano:
show only things
that
can be
seen
be a llora si vedono.
well
they can be
seen then
In this story, the father, adopting a rational stance, challenges Stefano, the eightyear-old son, about his fear of ghosts. In his defense, Stefano keeps within the
same line of discourse and provides evidence ("Once I stumbled on the nothing") .
435
The father, although rejecting the explanation, actually reifies a nonentity not too
different from ghosts ("and that was the nothing") . Then the mother depicts
Stefano's fear as being a shame for a child of his age, and the father builds on it,
stressing Stefano's easiness to believe in every foolish thing his friends say or
that is written in stupid books . Stefano then argues against the core argument
(i.e., that the book lies) by presenting a new piece of evidence, the photos. The
father's rebuttal is again weak: Although meant to present the visibility of a
ghost as an absurdity, it can be taken literally to strengthen Stefano's position,
which, in fact, he does.
The level of discourse established by the father is accepted and used here by the
child to dismiss the content of the challenge and to discharge the negative
evaluation attached to his conduct.
Cou nt er-Opposing the Prob lematizer

The Problematizee can also try to reverse the situation by counter-opposing the
Problematizer: that is, by challenging him on the same ground. This is a more
complex strategy, mainly enacted by older children who profit from the space
opened by a possible conflict between the parents when they reciprocally attribute
to one another the responsibility for children's misbehavior.
Excerpt 12
Naceri family : Violin
Dad: Giancarlo , 45 years old; Mum: Gaia, 44 years old; Daughters: Ludovica, 14 years
old; Irma, 10 years old ; Antoni a, 3 years old
Dad: - > doveva andare dalle sei - aIle sei e mezza

she had to go
from six-to six thirty
dal professore di violino

to
I nna :
the
violin
teacher
macche dalle sei alLe sei e mezza

from s ix to
six thirty what?
[tu non h a i capi t o niente

you
Dad:
d idn't
understand
anything
[e si e mossa da qua aIle sei e trentacinque

and
she
started off
at
six thirty-five
poi dice perche mi sei venuto a prendere

then she
say s
why d id you come to p ick me up
aIle sette meno dieci .

at
ten t o
seven
Inna :-> mamma io ho detto aIle sette meno dieci

mum I
said ten to
seven
perche lui viene sempre una mezzoretta pili tardi

' caus e
he
always
c ome s
about
half
an hour
l ate r
436

a prendermi e venuto alle otto meno dieci .
to pick me
Dad:
up.
he came at ten to eight
non si rende neanche conto che e arrivata

she
doesn't
even
realize
she
arrived
alle sei e mezza 10 sto dicendo da un sacco di tempo

at
six thirty I
am
saying it for ages
e uscita alle sei e mezza dietro i rniei forti solleciti

she went
prompts
out at
six thirty after my repeated
mo' telefono i o al professore di violino e stabilisco 10

now I ' l l
call
the violin teacher and fix
l'orario per la prossirna volta .

the time
Mum :
for
the
but the time
Dad:
next
lesson
rna l ' orario questo e .

is this one
dalle sei alle sei e mezza puntuali se ne e uscita

from six to
six thirty on time.
she went
out
alle sei e trentacinque corne e possibile?

at
Mum :
six thirty-five.how is
it
possible
rna e assurda . comunque oggi sara andata rnalissimo

well she
today
is
absurd.
anyway she probably was
bad
perche in tutta questa settirnana ha fat to

, cause
in the whole week she has
done
una sola volta esercizio

Dad:
only once this week

s i rna l'anno prossirno non fara violino
yeah but next year she won't do violin
Mum:
NO : :
Dad:
no .
neanche lei 10 vuole fare pili
not
Mum:
even she wants to do
it anymore
no: :no rna anche se 10 volesse fare non glielo farei fare
no no but even if she wished to do it I
wouldn't allow her
perche e un irnpegno per me non per lei io ogni giorno

, cause
can't
it's a
commitment
for me,
not
for her.
non posso stare a ricordare che deve fare il violino e

remind her everyday that
the violin and
she
has
to
practice
lottare contro la sua volonta che non 10 vuole fare eh .

fight against her own will
want to do it
that
she
doesn't
This story shows how the conflict between daughter and father is mediated by the
mother, who is addressed by both. The daughter partially succeeds in counter-
437
opposing the father by using the same argument ("he is always late") he was
using against her. But when the father addresses his counter-defense to the mother
by saying that he will be the one who sets the time for the violin lesson, she
answers laconically and coldly that "the time is this one." So the mother is not
only the one who does more problematizing on the past activity than any other
single participant (as it shown in Table 3) but she is also the privileged recipient
or addressee of many disputes. And, in this case, she is recognized as the deciding
judge. She has the last word in the dispute between father and daughter because
she is committed to having the daughter practicing the violin. Her conclusion has
the form of an articulated reproach; it embeds a counterfactual argument, is
expressed in the third person, and is also addressed to the father in the sense that
he is not as involved as she is in getting their daughter to study the violin.
Conclusions
Return ing to the questions posed at the beginning, we can confirm that conflict
talk is not dispreferred (Levinson, 1983) but is rather usual in Italian families'
shared discourse. The quantitative results indicate that the problematizing activity
occupies a large part of family discourse, given that about one third of family talk
is problem centered. This allows children peripheral participation in conflict talk,
both when they are and when they are not the object of the discussion. In such
cases, children are exposed to various strategies for challenging and defense.
However, half of the problematizations are directed at the children, and when the
discours e concerns past events, it is more difficult for them to have an active role
of problematizer. We found that the mother controls and sustains this family
genre.
We focused on the different kinds of problematic situations in which children
can participate, when past events are problematized. We found that it is easier for
them to build a challenge within a topic already introduced by someone else. This
also appears to be a successful strategy of turn-taking, in that their problematizing
moves get taken up more easily. Specifically, they formulate topic-tied questions
that strongly demand a second pair part (Sacks, 1992), that is, of some kind of
relevant response.
Compared with discourse interaction in school, family discourse has an obvious
difference. In teacher-pupils interaction, an opposition is referred to another's
claim and is always a second pair part because it requires that someone has taken a
position or advanced a claim (Orsolini & Pontecorv o, 1992); in family dinner
talk, the Problematizer can raise a question and problematize someone else from
the very beginning of the sequence because the referent is probably already shared
by the participants. On a fine-grained linguistic level , this can explain the use of
rna [but] at the very beginning of the problematic topic. Sometimes at the
438
beginning, there can be a description or a short report of the facts, but the story
can also begin with a question . Although older children seem to be able to
introduce a topic on which they problematize someone else, for younger children
it is easier to link their argument to a topic previously introduced by other, more
expert participants. However, children become aware rather early that questions apparently very innocent ones, as we have shown in the previous excerpts - act as
one of the more appropriate communicative avenues by which to enter into the
family discourse arena.
When some children's past behavior is challenged, even children as young as
Stefania (Excerpt 8: Quarrel), are able to justify themselves and to articulate
answers that are rhetorically designed and finely tuned to their particular audience,
by offering a seemingly accurate reconstruction of the quarrel with her brother and
by using temporal markers, authority reference, and visual recall. Moreover, while
they are exercising their conversational skills in justifying and arguing, children
are also coping with the evaluative, normative, and encyclopedic connotations of
various family discourse topics. That is the case of the superior value of truth to
which Stefania appealed as a compelling argument.
Almost every example that we have reported here could be analyzed for
unraveling the implicit normative assumptions underlying the debate, both in the
challenging and in the defending parts. For instance, in Excerpt 7, when Sergio is
teasing about his father's laziness, the father defends himself by appealing to his
superior duty to look out for the children and conjures up, in competitive
cooperation with the mother, a little scene of parental sacrifice for the children's
benefit. A shared Italian value of the centrality of children's needs is thus
dramatized by the two parents for their children's sake.
In a general way, all the problems described here are not strictly cognitive . They
are far from classical (and mostly artificial) problem-solving situations, because
the focus is on problem definition (as we said at the beginning), but they are also
more value laden than the ones in problem-centered topics oriented to the future in
which practical solutions are sought. The reconstruction of a past that is yet
problematic appears substantially structured within social-relational problems.
Once again, cognitive activity related to communication concerns social
relationships in the first place, as shown in Bruner (1990).
We have tried to show in our analysis that an underlying orientation to social
and/or family norms and values is detectable in all our short stories and enacted in
the collective discourse. More or less active participation in these types of
discursive situations, which frequently lead to arguing activities, enables children
to be exposed to and to explore different strategies. So children need to learn the
use of complex discursive tools that are embedded with both arguing and
reasoning dimensions. These include the pragmatic ways of responding to requests
and problematizations proposed by parents and siblings through argumentative
moves accomplished by particular linguistic forms. They also include the use of a
range of different justification strategies, such as recategorizing reported events,
439
transferring the problem focus from self to others, and gaining the right of
challenging and of problematizing others.
Acknowledgment
The study presented in this chapter is part of a research project on interactive
learning in different social contexts that was supported with a 1993-1994 grant
from the University of Rome "La Sapienza" to the first author.
440
ClotildePontecorvo and AlessandraFasulo
References
Bakhtin, M. M. (1986) . Speech genres and other late essays. Austin : University of
Texas Press .
Billig , M. (1987). Arguing and thinking: A rhetorical approach to social psychology.
Bruner, J. (1986) . Actual minds, possible worlds. Cambridge, MA: Harvard University
Press.
Bruner, J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press .
Caron, J. (1990) . Le developpement de l'argumentation chez I 'enfant [The
development of argumentation in children] . Paris : Presses Universitaires de France .
Corsaro, W. A., & Rizzo, T. A (1990) . Disputes in the peer culture of American and
Italian nursery school children. In A D. Grimshaw (Ed.), Conflict Talk (pp. 21-66) .
Duranti, A, & Goodwin, C. (Eds.). (1992). Rethinking context: Language as an
interactive phenomenon. Cambridge, England : Cambridge University Press .
Edwards, D. (1991). Categories are for talking. On the cognitive and discursive bases
of categorization. Theory &: Psychology, 1(4), 515-542.
Edwards, D., & Mercer, N. M. (1987) . Common knowledge: The development of
understanding in the classroom. London : Routledge.
Edwards, D., & Middleton, D. J. (1988). Conversational remembering and family
relationships: How children learn to remember. Journal of Social and Personal
Relationships, 5, 3-25.
Edwards, D., & Potter, J. (1992). Discursive psychology. Newbury Park, CA: Sage .
Garvey, C. (1993) . Diversity in the conversational repertoire: The case of conflicts in
pretending. In C. Pontecorvo (Ed.), Forms of discourse and shared thinking .
Cognition and Instruction, 11(3 & 4), 251-264 .
Goffman, E. (1964) . The neglected situation. American Anthropologist, 66(6), 133136.
Goffman, E. (1974) . Frame analysis: An essay on the organization of experience. New
York: Harper and Row.
Goodwin, C. (1994) . Professional vision . American Anthropologist, 96(3), 606-633 .
Grimshaw, A D. (Ed.). (1990) . Conflict talk: Sociolinguistic investigations of
arguments in conversations. Cambridge, England : Cambridge University Press.
Kaye, K. (1982). The mental and social life of babies. Chicago: University of Chicago
Press.
Labov , W. (1972). Language in the inner city: Studies in the Black English vernacular.
Philadelphia: University of Pennsylvania Press .
Labov, W., & Fanshel, D. (1977). Therapeutic discourse: Psychotherapy as
conversation. New York: Academic Press.
Lave, J, & Wenger, E. (1991). Situated Learning: Legitimate peripheral participation.
Levinson, S. C. (1983) . Pragmatics. Cambridge, England: Cambridge University
Press.
Ochs, E. (1988) . Culture and language development: Language acquisition and
language socialization in a Samoan village. Cambridge, England: Cambridge
University Press .
441
Ochs, E. (1994) . Stories that step into the future. In D. Biber & E. Flanagan (Eds.)
Sociolinguistic perspective on register (pp .106-135). Oxford, England: Oxford
University Press .
Ochs, E., Pontecorvo, C., & Fasulo, A. (1996) . Socializing taste . Ethnos , 61(1-2). 746 .
Ochs , E., & Schieffelin, B. (1989) . Language has a heart. Text, 9(1), 7-25 .
Ochs, E., & Taylor, C. (1992a). Family narrative as political activity . Discourse and
Society, 3, 43-57 .
Ochs, E., & Taylor, C., (1992b) . Mothers' role in the everyday reconstruction of
"Father knows best. " In K. Hall (Ed.), Locating power: Proceedings of the 1992
Women and language conference (pp. 447 -462) . Berkeley, CA: University of
California at Berkeley.
Ochs, E., Taylor , C., Rudolph, D., & Smith, N. (1992). Storytelling as a theorybuilding activity . Discourse Processes, 15, 37-72.
Orsolini, M. (1993) . "Because" in children's discourse. Applied Psy cho linguistics,
14, 89-120.
Orsolini , M., & Pontecorvo, C. (1989) . La genesi della spiegazione nella discussione
in classe [Genesis of explanation in class discussion] . In M. S. Barbieri (Ed.), La
spiegazione nell 'interazione sociale (pp. 161-190) . Torino, Italy : Loescher.
Orsolini, M., & Pontecorvo, C. (1992) . Children's talk in classroom discussion .
Cognition and Instruction, 9, 113-136 .
Pomerantz, A. (1978) . Compliment responses: Notes on the co-operation of multiple
constraints . In J. Schenkein (Ed.), Studies in the organization of conversational
interaction (pp. 79-112) . New York : Academic Press .
Pomerantz, A. (1984). Agreeing and disagreeing with asses sments : Some feature s of
preferred-dispreferred discourse. In 1. M. Atkinson & 1. Heritage (Eds.), Structures of
social action (pp. 57-101). Cambridge, England : Cambridge University Press .
Pontecorvo, C. (1987) . Discussing for reasoning : The role of argument in knowledge
construction. In E. De Corte, J. G. Lodewijks, R. Parmentier, & P. Span (Eds .),
Learning and Instru ction. European research in international context, 1, pp. 71-82 .
OxfordlLeuven: PergamonlLeuven University Press .
Pontecorvo, C. (1993) . Forms of discourse and shared thinking . Cognition and
Instruction, 11, 293-310 .
Pontecorvo, C., Amendola, S., & Fasulo, A. (1994) . Storie in famiglia. La narrazione
come prodotto collettivo [Stories in the family . Narration as a collective product] .
Eta Evolutiva, 46, 18-34.
Pontecorvo, C., & Girardet, H. (1993) . Arguing and reasoning in understanding
historical topics. In C. Pontecorvo (Ed.), Forms of discourse and shared thinking .
Cognition and Instruction, 11, 365-395 .
Pontecorvo , C., & Orsolini, M. (1993). Discussing a story in a pre -school setting . In
M. A. Pinto & M. Danesi (Eds.), Humanism in language studies: Essays in honor of
Renzo Titone (pp . 78-94). Milan : IMFE .
Resnick, L. B. (1991) . Shared cognition: Thinking as social practice. In L. B.
Resnick, J. M. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared
cognition (pp. 1-20) . Washington, DC: American Psychological Association .
Sacks, H. (1987) . On the preferences for agreement and contiguity in sequences in
conversation. In G. Button & J. R. E. Lee (Eds.), Talk and Social Organization (pp .
54-69). Clevedon, England: Multilingual Matters.
442
Sacks , H. (1992) . Lectures on Conversation: Volume 1. (G. Jeffer son, Ed.) Oxford ,
England : Basil Blackwell.
Schegloff, E. A. (1989). Reflections on language , development, and the interaction al
character of talk-in-interaction. In M. H. Bornstein & J. S. Bruner (Eds.), Interaction
in human development (pp. 139-153). Hillsdale, NJ: Erlbaum.
Schegloff, E. A. (1990). On the organization of sequences as a source of "coherence" in
talk-in-interaction . In R. Freedle (Ed.), Advances in discourse pro cesses:
Conversational organization and its development (Vol. 38, pp . 51-77). Norwood ,
NJ: Ablex.
Schiffrin, D. (1985) . Everyday argument: The organization of diversity in talk . In T .
van Dijk (Ed.), Handbook of discourse analysis, Vol. 3. Discourse and dialogue (pp .
35-46) . London : Academic Press .
Schiffrin, D. (1987) . Discourse markers. Cambridge , England: Cambridge University
Press.
Schiffrin, D. (1990). The management of a co-operative self during argument. The role
of opinions and stories. In A. D. Grimshaw (Ed.), Conflict Talk (pp. 241-259) .
Semin, G., & Fiedler, K. (1988) . The cognitive functions of linguistic categories in
describing persons: Social cognition and language . Journal of Personality and
Social Psychology, 54(4), 558-568.
Vygotsky, L.S. (1990) . Pensiero e linguaggio [Thought and language] . Bari, Italy:
Laterza . (New Italian critical edition by L. Mecacci, based on the first Russian
edition; originally published in 1934) .
Chapter 18
Discourse in the Adult Classroom:

Rhetoric as Technology for Dialogue
Juan D. Ramirez! and JamesV. Wertsch?
1
2
Laboratory of Human Activity, University of Sevilla, Spain

Department of Education, Washington University, St. Louis, MO, USA
Abstract
The aim of this chapter is to present observations on the role of literacy in adult
education, with special emphasis on the cognitive and discursive transformations
that represent the appropriation of a new communicative technology such as
writing (Goody, 1977; Ong, 1982). Starting with the contributions of rhetorics of
this century (Burke, 1969; Billig, 1989), we have studied an activity frequently
promoted by teachers inspired in the main ideas of Paulo Freire: cl assroom
debates . Through the study of this kind of adult classroom activity, we have
analyzed the following aspects involved in them: types of semiotic means (e.g. ,
contextualized vs. decontextualized signs, speech genres, social languages;
Bakhtin, 1986a; Wertsch, 1985, 1991) used in debates and similar forms of
conversation by students (all of them women of several adult centers situated in
southern Spain) ; modes of argumentation and ways for persuading the audience
displayed by participants in debates ; motives that emerge in the course of this
form of dialogue .
444
Introduction
Research into literacy has been carried out in many of the social sciences. Many
of the findings from research in psychology, communication ethnography, and
literary theory have contributed to a better understanding of educational processes
in general and of the learning of writing in particular. However, these findings
have not yet filtered through to one of the educational areas that is presently
undergoing rapid development: literacy and adult education. This study aims to
probe into the complex world of adult education . We will do so by observing and
analyzing the transformations that literacy produces in orality in people attending
courses at different educational levels at the adult centers that we have been
studying in the Andalusian Adult Education Program, southern Spain, Cala,
1991; Laboratory of Human Activity , 1993).
We have been studying a series of group debates on subjects that, more often
than not, concerned and motivated the participating students, who were practically
all women . These discussions, on topics such as children's education,
unemployment, and drug abuse in youngsters , were often stimulated by the
teachers and provide an ideal setting for obtaining generative meaning with which
to develop texts that are understandable and stimulating for the student (Freire,
1970, 1973; Freire & Macedo, 1987). From our standpoint, it seems that they
also function as an interesting setting for testing the changes that a
communication technology such as writing can promote within the framework of
oral communication. However, we believe that orality and literacy each have an
influence on the other, and, if it is true that the former is affected by the second,
then it is also necessary to practice some types of dialogue, such as a debate, to
increase mastery of oral forms of discourse to begin with, and then later on to
develop the necessary interpretative skills for the comprehension of written texts.
In future research, we hope to clarify the mutual transformations between both
forms of communication in the framework of the research being undertaken in the
Laboratory of Human Activity . The underlying hypothesis in this chapter relates
to this two-way relationship. We believe that the technologizing of the word by
writing has changed our orality significantly, giving rise to rhetorical forms of
arguing that came up in the debates we observed. But we should also a<tl
something that is rarely taken into consideration by researchers into literacy,
while at the same time being widely acknowledged by adult educators: Promoting
this type of discourse is necessary to stimulate the understanding of our audience's
discourse in the use of words. In the case we discuss here, this is tantamount to
saying the discourse of others participating in the debate. Enlarging the capacity
for understanding oral discourse is fundamental for developing oral skills that will
facilitate the interpretation and comprehension of written discourse.
Discourse in the Adult Classroom
445
Literacy and Technology of Dialogue

Literacy represents the construction of a new system of communication
superimposed on another already existing one (orality) that is common to all
humanity. If oral language builds meaning from sound (e.g., phonemes, syllables,
articulated words), writing represents a particular form of grapholect (Ong, 1982)
that endows meaning with a new form, a new body, in its own particular space,
bound in by the edges of a piece of paper, capable of being recovered or, put
differently, of being re-read on different occasions. Such a grapholect is not linked
to any dialect, gender, or style other than that which transcends it. The written
text may serve a wide variety of genres and even serve artificial languages created
ex professo for a certain setting of cultural activities. Good examples are the
following symbols taken from the language of chemistry: C, 0, H. They all form
part of the Lati n alphabet of many written languages. At some point in the
evolution of these languages, their users decided to use this alphabet rather than
another one (e.g., Greek, Germanic, or Hebrew). At some early stage in the
history of modern chemistry, some scholars decided to use these letters as
symbols for very specific entities. The aim was to transmit a particular form of
knowledge and manipulate certain substances outside the framework of the
laboratory, using symbols (C, 0 , or H) and rules enabling the explorations of the
complex nature of these substances and their possible combinations, before
moving on to actual experimental manipulation. So, writing represents a type of
grapholect for very different forms of communication and action, serving a wide
body of activities in varying settings of cultural creativity.
Following Walter J. Ong's brilliant analysis, let us say that writing can never
be freed from orality in the sense that it is a secondary cultural system and, as
such, it relies on a primary system of communication. It hardly needs stating that
this primary system is spoken language . "Oral expression may exist and in fact
has done without writing, but one never finds writing without orality" (Ong,
1982; p. 8). However, once the speakers of a language take on this language
through a second system of communication (writing), orality is changed radically.
According to Ong, we can distinguish between a primary orality, one that is
unpolluted and that appeared before the emergence of writing, and secondary
orality, which results from the interaction between primary orality and the written
code . The reason for the changes that occur as a result of the influence of writing
has to be found in the possibility of returning to any type of knowledge, once this
has been laid down in a written text. The text may be attended to again and again
through that complex activity that we call studying, something that is impossible
with oral text. A new level of intellectual functioning is reached through
studying, and it is completely different from that occurring on the plane of
primary orality. Knowledge linked to the written text may be organized in classes
capable of being analyzed at different moments in the life of the researcher; it can
also be seen in the light of different perspectives developed by the reading and
446
understanding of other texts. Learning in an oral culture or in a nonliterate

sociocultural context varies radically from learning in a written culture (or one
dominated by writing). The term we use to refer to the person who learns differs
accordingly. The word student is reserved for written culture, whereas in
predominantly oral cultures, the more appropriate word is apprentice (Rogoff,
1990; Wertsch, Minick, & Arns, 1984). Although there is certainly a great deal of
truth in Ong's statement that cultures that are transmitted using an oral tradition
can develop a wide and complex body of understanding, they are nevertheless
excluded from something that is exclusive to cultures with a written tradition: the
activity of study. Even sophisticated forms of fundamentally oral activity, such as
rhetoric, could not exist without the systematization of the principles and
techniques that make it possible; and these have been codified, revised, and
reflected on in written texts by different academics from antiquity to the present
(see Billig, 1989; Burke, 1969).
If the development of writing introduced these great changes that Ong and many
others have observed in the socio-historic field (Goody, 1977; Ong, 1982), we
must assume that reading -writing learning, to the extent that it transforms or may
come to transform forms of oral communication, will allow the development of
new forms of describing daily life and the appropriation of new tools for
discourse . We make a special reference to these tools, because we agree with Ong
when he defends a close connection between literacy and rhetoric . The arguments
in defense of one's own point of view, the replies to the points of view defended
by our audience, and many other processes involved in persuasion, the central aim
of rhetoric, must be closely related to the semiotic changes promoted by literacy.
A related question concerns whether these changes are produced in the area of
ontogenesis and if literacy acquired in formal education can produce the necessary
semiotic changes in the individual to make possible the development of
mechanisms for dialogue in public.
The Functions of Debate

The debate of ideas has been considered traditionally as a fairly sophisticated
academic practice that puts to the test sometimes extremely complex techniques
and forms of dialogue. These may be understood as procedures for defending points
of view that the orator adopts and, complimentarily, as a means of persuading the
audience. However, these procedures, tending as they do toward persuasion, were
considered not only from a discursive angle, but they were also looked on as a
path of moral development. The relation between rhetoric and moral development
was crucial for philosophers and pedagogues such as Quintillian and Cicero , who,
in many cases, coupled the ideal of the orator to the ideal of the citizen insofar as
447
this was a person characterized by compassion for others and respect for other
people's basic rights (see Burke, 1969).
Two aspects of rhetoric should be highlighted from an ethical perspective: the
role of persuasion of the audience and the personal transformation of the speaker
in the use of words. Understanding the rhetoric act as a communicative event,
motivated exclusively by the desire to persuade the audience, represents just the
ftrst of the two aspects referred to; the second is understanding this as an act that
affects both the speaker and the other interlocutors . For this double affect to
happen , we need to accept a dynamic vision of language that will allow us to
understand the rhetoric act in terms of the real action: a particular form of action
that affects the audience and simultaneously transforms the speaker who carries it
out. Taking into account the two-way direction of rhetoric, the concept of
persuasion, normally understood as the fundamental motive, gives way to another
motive of great importance, the achievement of social cohesion, so necessary for
living in groups. As Duncan, one of the most illustrious proponents of the new
rhetoric, has stated, the attaining of such an important objective is conditioned by
the fact that men act "rhetorically on themselves and on others" (Duncan, 1989,
p. 154).
The role of rhetoric can be found in relation to the commitment that emerges
with the use of the word. Far from the neutrality that formalist approaches have
tried to implant on meaning, the use of the word always involves addressivity,
otherness, and, in short, commitment and responsibility of the speaker toward
those listening to him or her (Burke, 1969, 1984; Clark & Holquist, 1984;
Voloshinov, 1986). In Burke's opinion, speech is not a neutral event. In fact, it is
far from lacking judgments and moral assumptions; it is dominated by them. All
of this is intensely moral. Names for objects contain emotional suggestions that
give us clues concerning how we are to act in relation to them (Burke, 1984). A
word such as automobile will normally involve a concealed selection (it
designates not merely an object but also a desired object). Speech is not
exclusively guided by the mechanical act of referring; instead, it acts as a system
of attitudes. Burke continues his reasoning, saying that calling a man friend or
enemy is the same as suggesting a program of action with which to consider him .
An important ingredient in the meaning of such a word is the set of attitudes and,
consequently, the availability of the action that this involves . And the author
adds: "regardless of whether we should call the implicit program of action adeqJ.ate
.. . or inadequate .. . these emotional or moral weights inherent in spontaneous
speech tend to reinforce the act itself, hence making the communicative and active
aspects of speech identical" (Burke, 1984, p. 177).
The perspective adopted by Burke is important to us for one fundamental
reason: By eliminating any supposition of neutrality, the word is re-infused with
the morality that it had lost, along with its capacity for creating attitudes for
action. If we also combine the two ideas we have already mentioned (that the
rhetoric act influences both the speaker and his/her audience, and that the use of
the word is moral and generates attitudes for action), we come to a conclusion that
448
coincides with the standpoints defended by Bakhtin and his circle: namely, that
human experience is created through discourse, and, as a result, it lacks real
existence outside the field of expression. From the very moment of expression , in
the act of producing an utterance, we are creating experience or recreating our
internal experience. But, at the same time, internal experience is created by the
same semiotic material that gives form and expression to its external
objectivation . From this, we can conclude that no experience exists that is not
linked to signs (Bakhtin, 1986a; Mead, 1934; Voloshinov, 1986, 1987).
If the use of the word is not a neutral event but the creation of attitudes, and if,
as Bakhtin points out, it is the way through which the speaker can externally
recreate internal experience, in spite of the word having, like all signs, an external
origin, we can conclude by saying that the rhetoric act affects both the audience
and whoever carries it out. But, in addition to this, expression-experience, the
utterance once it is produced, may turn itself into a motive for reflection for its
transmitter and enable him to discover new senses that were latent in him.
The Decontextualization of the Sign and Its Function in

Arguing
The concept of agency refers to the capacity the agent has for carrying out actions
without taking into consideration the subjective intentions motivating him at the
time he is carrying them out (Giddens, 1989). The developing action will depend
both on the actual agent and on the conditions in which it develops. An agent has
the "capacity for . . . (doing something)" when he or she possesses adequate
control of the tools for carrying it out and knowledge of the situation.
For an adequate consideration of the concept of agency, it is necessary to
emphasize an aspect that is rarely considered by specialists in this area: Agency
does not rule out authorship. While he is the agent, the speaker is responsible for
what he does. Furthermore, insofar as this action is a rhetorical action, that
responsibility turns into authorship. Authorship can, in the purest sense of the
term, be described by saying that a speaker is author of the utterance he produces,
in the same way that he is writer of his text, whatever the literary genre chosen.
So, the speaker-agent is the author of the rhetoric action (insofar as he tries to
persuade his audience to get them to identify with his own motives) . He depends,
sometimes without knowing it, on his interlocutors, those whom the tries to
affect with his rhetoric action.
Given the important role of interlocutors in the course of rhetorical action, we
shall label the type of agency involved in it shared agency. This concept could be
described from a point of view of discourse in the following manner: To achieve
his or her goals, the speaker depends on what he knows about the arguments that
his or her interlocutor uses, because he or she has to answer them.
449
The analysis of this concept does not stop here. To gain a better insight into it,
we need to analyze the means/tools with which the agent acts, because in addition
to human agency being shared, it is also mediated agency, given the fundamental
role that tools play in it (Wertsch, Tulviste, & Hagstrom, 1993). Mediated agency
is, in Bateson's terms, that which extends beyond the skin (Bateson, 1972;
Wertsch, 1991; Wertsch, Tulviste, & Hagstrom, 1993). Apart from depending on
the behavior of others at the time of carrying out one's own action, one also needs
the tools for acting . So human agency is mediated in the sense that it is carried
out with the use of cultural instruments, of mediating procedures, such as the
tools of work and language. Whatever their structure and the type of activity that
they facilitate, tools enable us to extend our mind beyond the skin, beyond the
narrow limits of our bodies.
Apart from the way in which we have tackled the analysis of the agency
responsible for rhetoric action, we can see the latter as a subset of what Wertsch
and colleagues have labeled mediated action (Wertsch, Tulviste, & Hagstrom,
1993). Given the important role played by the settings and tools used for carrying
out the action, we must assume that variations in the settings will produce
important changes in the course of and in the results of the actions in which they
are involved. If we consider that rhetorical actions involve a peculiar use of signs,
of word, and, in general, of semiotic settings, we will have to accept that the level
of decontextualization of these will play an important role when they are carried
out.
Several social and interactive processes are directly responsible for the
deconiextualizaiian of signs and indirectly responsible for the development of
higher mental functions (Vygotsky, 1978). By adopting a cultural-historical
perspective, Vygotsky took into consideration all the social processes as a whole
to explain the genesis of higher mental functions. In practice, however, he
actually tackled only the study of the sign in the framework of basic social
processes, such as interaction between dyads and relationships within the
framework of small groups (Vygotsky, 193411987; Vygotsky, 1978). Subsequent
experience has shown that the development of these instruments and their
progressive decontextualization do not depend exclusively on interactions within
the framework of small groups; they also depend on large scale institutions
developed for this reason . In what is undoubtedly his most representative book,
Thinking and Speech (193411987), Vygotsky was aware of the importance of
formal education in the process of decontextualization of the word's meaning. It
was clear to him that, without the framework of formal education, the child could
not reach a level that would be sufficient for developing scientific concepts .
In the Western world, the school has evolved as a setting in which society
provides the instruments and sufficient knowledge for people to be able to live
within it. The general set of practices that are carried out there provide the student
with new meaning that would be unlikely to be transmitted without it. It is rare
for these meanings, which Vygotsky labeled scientific concepts, to have arisen as
the result of individual work . Instead, they represent the efforts of working groups
450
(e.g., scientific, professional, religious) that are also talking communities in

continuous debate about the result of their activity.
Dialogue and Agency: Social Languages and Genres in

the Art of Arguing
The collective nature that shrouds the construction of the meaning, something so
often overlooked by language experts, was not only a fundamental research
concern for Vygotsky but was also one of the most interesting subjects for
Bakhtin and his colleagues. If, through Burke, we say that the meaning of the
word is never neutral, we must add, following Bakhtin, that neither is it
impersonal, as if it were a dictionary meaning. We should not forget that, before
its appropriation by a specific speaker, the word "exists in other people's mouths,
in other people 's contexts, serving other people's intentions : it is from there that
one must take the word, and make it one's own" (Bakhtin, 1986b, p. 294).
If we were saying earlier that the meaning of the word is decontextualized in the
setting of formal education, we should now add that this is possible thanks to the
different discursive forms generated there (e. g., the discourse of the social
sciences, the natural sciences, language, and literature). Bakhtin labeled each of
these forms of talking a language depending on the activity in which the speaker
is placed as a social language (1986b). Even though the author did not leave us a
clear definition of the term, we can define it, following Holquist and Emerson
(1986), as "a peculiar form of discourse for a specific stratum (professional,
religious, age group, etc), within a given social system and at a fixed time"
(Holquist & Emerson, 1986, p. 430 ). We should not find it strange that school
encourages decontextualization of the meaning of word, because, through
confronting us with diverse social languages, it creates the necessary conditions
for successive reappropriations of words known to the student and puts him or her
in a situation of revising and redefining their meanings; the longer the educational
process, the more the revisions that will have to be undertaken and the greater the
level of decontextualization achieved.
Apart from the concept of social language, Bakhtin proposes another
complementary concept: that of the speech genre. If the social language has to do
with the socio-ideological orientation of the meaning of the different group and
strata that make up a given society, genre is related to the form that the discourse
adopts, depending on the particular communicative situation and on the social
context in which the speakers relate to each other. Bakhtin defined the genre as a
typical form of utterance in which the word adquires its expressivity. "Genres
correspond to typical situations of speech communication, typical themes, and,
consequently, also to particular contacts between the meanings of words and actual
concrete reality under certain typical circumstances" (Bakhtin, 1986a, p. 87).
451
What seems to prevail in the genre, in contrast to in social language, is its

dependence on the situation in which the communication is produced. The result
is that the genre presents surface markers that are absent from social language
(Wertsch, 1991). We could say, in short , that genre is an instrument of
communication whose routinized and predictable use enables the speaker to
achieve certain social ends in particular situations in daily life (Bauman, Irvine, &
Philips, 1987; Brigg, 1993; Philips , 1987; Wertsch, 1991).
The skill of the speaker in control of series of genres may allow flexibility in
its use and even the creative combination of some with others in order to achieve
certain rhetoric objectives. A skillful use of these enables the person
communicating to carry out complex combinations, such as combining various
primary genres in order to reach another more complex one. Bakhtin named the
results of these types of combinations in the history of literature secondary
genres. From our point of view, these generical combinations, which are
fundamental in explaining the literary product, may also be useful to us for
explaining conversational events from daily life. In the same way that the novelist
draws on different literary genres in a story, the speaker may carry out generic
combinations in the course of a conversation or during a debate.
Types of Utterances
The utterances we present next are representative examples of the discourse uttered
by the women, all of them students at different educational levels of the Adult
Education Program, taking part in debates. In all the debates observed in the adult
classes , we can confirm the existence of utterances that reflect a wide variety of
personal experiences (utterance-type), in terms of both their content and the
interdependence of the dialogue with preceding utterances. During one of these
adult classes, on the subject of household chores , the women taking part produced
utterances similar to the following:
1. My husband doesn 't help me at home

2. My husband doesn 't even help me to clear away the plates
There is one common thread running through the possible meanings and senses
reflected by these utterances. They all depend on the speaking voice concentrating
on itself and on its daily experiences, and, given the conditions created by the
dialogue, the speaker decides to exteriorize dramaturgically a part of herself
(Habermas, 1984) through the presentation of fact from her daily existence .
452
However, these are not the only utterances expressed by those taking part in the
debate. Others of a more general content may transcend the limits of that existence
and place the problem described in a more complex social context. The following
examples reflect more general facts:
3. The man should help with the housework

4. Men should share the household chores
When someone produces an utterance like this, it does not refer to just one
particular person; it refers to many. The woman who says something like this is
producing an utterance whose meaning could be assumed by many other women.
Most likely it is applicable to such a wide variety of couples that it goes beyond
the boundaries of its group of reference. It is fairly common to see expressions
like 3 and 4 in propaganda sent out by political parties at election time in an
attempt to win over the female vote. Example 3 could also come from a feminist
text, or it might appear in the more progressive press reflecting present-day issues
affecting women.
Utterance 1 (My husband doesn 't help at home) reflects the use of
contextualized signs. It takes the self of the speaker as a reference point or, to put
it differently , as a deictic center (Levinson, 1983). In addition to this, the referent
is one particular person, and, as a result , the sign that refers to him cannot be
anything other than a contextualized sign (my husband). The same thing happens
with the predicate. In the latter, a very specific type of help is mentioned: the
chores that the speaker carries out in her own home. So, all the signs are highly
contextualized, and, as a result of that, we get a particular and specific image of
the situation .
Utterances 3 and 4 transmit a different meaning (the man should help with the
housework, and menshouldhelpwith the household chores). The speaker here has
used a different type of tool. The signs used in these two cases are
decontextualized signs (the man, men, the housework, household chores), and so
the effect achieved is that of a general meaning, that has no particular individual as
a referent. Instead, it takes a general individual or individuals that do not even have
to be men from the same reference community (the men from this town, from
this neighborhood, and even from this country) as referent, and it may refer to the
male collective in its totality.
If formal education is the most privileged setting in our culture for producing
the decontextualization of meaning, the question to be asked in relation to the
utterance-types presented here has to do with their varied use by the adult students
taking part in the debates. Of all the utterances produced by the students at the
first levels of training (total or partial illiterates) , those similar to type 2 (My
husband doesn 't evenhelp me to clear away the plates) were the most frequently
found in their verbal productions. However, the utterances of types 3 and 4 tended
453
to occur in greater proportions among students of final levels . In other words,

they appeared among those with greater school experience, who were close to
ending their studies in the Adult Education Program. It should also be added that
the majority of the women attending the final courses have not normally worked
up from the lower courses (basic literacy and post-literacy) ; instead, they normally
go straight into them. In other words, the process of decontextualization started
long before they attended the center. We might also assume that the appropriation
of the different social languages that are characteristic of formal education, despite
being interrupted for whatever reason, must have started during childhood. So,
perhaps it is not surprising that the women in the final courses produce the most
general utterances, such as types 3 and 4. This indicates, above all, that the level
of semiotic decontextualization achieved is the product of their attending school
during their childhood and that there are added discursive practices carried out
beyond the limit of the school. The combination of these factors with the
experiences that the Adult Education Center provides makes itself evident in the
debates studied.
The analysis of the discourse produced in the different debates goes beyond what
we have analyzed to this point. By adopting a clearly rhetorical perspective, we
can see that the two opposite forms of producing utterances refer to different forms
of arguing. The opposition between the two reminds us of the two forms of
discussion referred to by Michael Billig (Billig, 1989). Utterances 1 and 2 point to
the form of arguing based on particularities (arguing about particulars), whereas
Utterances 3 and 4 seem to be closer to the categorial form (arguing about
categories). However, over the course of the different debates, both the literacy
students and those in the graduate course (final level) produced complex forms of
expression and/or arguing by combining two or more expressions of both types in
the same utterance. The following examples illustrate the combination of
utterances into more complex statements:
5 . Men should help with the housework. Although this is not my case. My
husband doesn 't help me at all.
6 . The man should share the household chores with the woman. Although my
husband doesn't help me with anything . He doesn 't even help to clear the
plates away.
A quick glance at Utterances 5 and 6 shows us a curious combination of both

types of utterances (general and particular) to form other, more complex types.
The speakers begin to describe general situations referring to the behavior of men
at home, and then they incorporate particular descriptions, which, in most cases,
refer to themselves. These examples represent complex forms of arguing in which
categories and particularities are present at the same time. In our view, we are
faced here with forms of arguing that demonstrate the ability of some of the
speakers in the debate to successfully combine discursive genres. The first of them
454
is a clearly colloquial genre that working-class Spanish women frequently use in

their conversations about daily life (e. g., conversations about their family life,
problems with the children, and gossip) The second is the genre that is closer to
using rhetorical devices in debate; as such, the utterances produced in it present a
more general meaning. To argue, combining particular and general ideas, would
not be possible without the control of two complementary genres that combine to
form another more complex one. But neither would it be possible to argue
rhetorically without the use of certain social languages (especially in the case of
general utterances) with the purest and most genuine voice of the speaker.
Some Final Thoughts

In light of the different forms of rhetorical action reflected in the utterance types
described and of the diversity of procedures used by the participants in the debates,
it is necessary to reflect on the motives involved in this type of dialogue . The role
of motives in discourse represents one of those typical problems that, despite its
importance, is rarely tackled by language experts. More specifically, our
reflections here have to do with the different rhetorical motives that are
externalized and/or promoted throughout any of these debates.
We should emphasize the existence of institutional aims, which are imposed by
the curriculum through the teacher, such as: stimulating dialogue and
comprehension between the students on the program as a way of developing a
democratic life-style, using the path of dialogue for the adequate development of
reading understanding in all its forms, and extracting generative words. To reach
these objectives , the teacher makes use of specific, well-established strategies in
the programs of literacy and adult education: for example, spontaneous
conversation, semistructured dialogue, and debate.
Faced with these aims that make up the set of strategies from which the
educational process is evaluated, we should also consider the tactics developed by
the students (De Certeau, 1984) to use to their own advantage the educational
experiences that the adult education center offers them, depending on their own
motivation. Let us suppose for a moment that the objective of all these tactics is
that of arguing in the terms that are laid down by the debate. This form of arguing
will be guided by a primary rhetorical motive: the persuasion of the speakers and
the identification of the audience with the speaker's point of view. However, in
our analyses of these debates, we have found a motive that is more basic than the
previous two; given that it has points in common with Goffmann and Habermas 's
concepts , we have labeled it the dramaturgical expression motive. The dialogue
may produce utterances referring to problems affecting the speaker or comments
about her daily life. From there, through the production of utterances that take as
a reference point the person who is speaking, this person uses the debate as a
455
pretext, or as a tactic for expressing herself dramaturgically . The attempt by some

of the participants in the dialogue to transform the debate into something
approaching a psychotherapy group seems to emphasize the dramaturgical nature
of the discursive production, especially in the literacy group. So, the motive of
the dialogue is expression through expression itself. As we pointed out
previously, the search to make the audience identify with what the speaker says
about her life comes about through empathy. Those listening put themselves in
the place of the speaker and try to share her experience. In these circumstances, we
can say of the speaker that she needs, above all, to be accepted and understood.
A dialogue in which each of the participants describes particular situations
revolving around her own daily experiences cannot be considered a real debate,
because the most important part is to be heard and understood by the other
interlocutors. On the other hand, when there is a dialogue in which the speakers
produce general utterances (e. g., the man should help with the household chores)
with the structure of real arguments, having as a basic aim persuading the
audience to identify with their points of view, it creates a real debate in the purest
sense of rhetorical tradition. Identification is produced when the audience considers
the utterance to be correct, assuming it totally or partially.
The motives driving dramaturgical expression and those acting in the purer form
of arguing are, however, not mutually exclusive. It would be a mistake to
suppose that when some motives are dominant they automatically block out
others. Expressiveness does not exclude arguing; it can instead be coordinated with
this, producing a complex web of motives . When a person argues in terms of
general utterances she may, for example, give them a content with particular
utterances referring to her daily life or to that of another person (a member of her
family or someone she knows), producing a form of arguing that combines
particularities with categories of the type described by Billig, in which social
languages and individual voices combine in more complex argumentative genres.
The existence of particular utterances dominated by dramaturgical expression, in
contrast to general and hybrid (particular and general) ones, and the fact that the
women in the literacy class are those who use it most, allow us to hypothesize in
two possible directions. On the one hand, we can imagine that the particular
utterances are the result of poor control of the semiotic tools that make
argumentation and any form of rhetorical action possible (e. g., decontextualized
signs, social languages, and genres). On the other hand, it could also be assumed
that the more frequent production of some utterances in comparison to others
reflects the prevalence of some motives over others for diverse reasons. We
believe that nonliterate women who attend for the first time or are in their second
year at the Adult Education Center find a place where, perhaps for the first time,
they can speak about themselves, express themselves freely, and be understood by
a receptive audience. In these circumstances, it is logical to think that
dramaturgical expression, insofar as it is a motive, prevails over the motive of
persuading and obtaining the identification of the audience through the channel of
456
arguing. The particular utterances will be under the control of the first motive
mentioned.
We should stress, however, that both hypotheses could mutually complement
each other. The recently acquired habit of expressing oneself in the educational
framework, faced with a public made up of the teacher and the rest of the women
in the classroom, makes the motive of expressing oneself dramaturgically take
precedence while at the same time highlighting the lack of tools and skills for
carrying out more complex rhetorical actions. But, at the same time, we should
also point out that only continued discursive practices (e. g., semi-structured
dialogues or debates) will allow the appearence of new motives and the
development of more sophisticated rhetorical instruments subordinated to them.
Whatever the different perspectives adopted in the study of debate, teachers,
experts, and researchers of literacy and adult education share one thing in common:
the great interest represented by the application of a rich variety of discursive
practices previously studied by rhetoric but largely forgotten by pedagogy and
other educational sciences.
It is worth considering the transforming role of literacy when this is acquired in
an appropriate educational framework. In the adult centers in which the researchers
from the Laboratory of Human Activity have carried out their observations, the
literacy practices are combined with various forms of dialogue. These dialogues
are sometimes informal conversations through which women students have an
opportunity to speak spontaneously about a wide variety of subjects. On other
occasions, the dialogues are organized in debates coordinated by the teacher. We
believe that only the appropriate combination of a new communication
technology, such as literacy, with the oral dialogue in all its variations can
promote the adequate decontextualization of the sign and the generation of new
forms of discourse (social languages and genres). Through these new semiotic
settings in the service of communication, the adult starting at the beginning of
the educational process or coming back to it after many years will be able to
widen his or her cultural horizon and find ways of adapting to the demands of a
changing world.
457
References
Bakhtin, M. M. (1986a). Speech genres & Other late essays. Austin, TX: University of
Texas Press .
Bakhtin, M. M. (1986b). The dialogic imagination. Austin, TX: University of Texas
Press.
Bateson, G. (1972). Steps to an ecology of mind. A revolutionary approach to man 's
understanding of himself. New York: Ballantine .
Bauman, R., Irvine, T., & Philips, S. (1987). Performance, speech communication and
genre: A critical review of concepts in the ethnography of speaking . Working
Papers of the Center for Psychosocial Studies (No. 11). Chicago: Center for
Psychosocial Studies.
Billig, M. (1989) . Arguing and thinking. Cambridge , England: Cambridge University
Press.
Brigg, C. L. (1993) . I'm not just talking to the victims of oppression tonight - I'm
talking to everybody: Rhetorical authority and narrative authenticity in an AfricanAmerican poetics of political engagement. Journal of Narrative and Life History, 3
(1), 33-78.
Burke, K. (1969) . A rhetoric of motive. Berkeley , CA: University of California Press .
Burke, K. (1984). Permanence and change. An anatomy of purpose . Berkeley , CA:
University of California Press .
Cala, M. J. (1991). El programa de educaci6n de adultos y el papel social de la mujer: un
estudio sobre valores , actitudes y modos de pensamiento [The Program of Adult
Education and the social role of the woman. A study about values, attitudes and
modes for thinking] . Memoria de investigaci6n . Sevilla, Spain : Instituto Andaluz de
la Mujer.
Clark, K., & Holquist, M. (1984). Mikhail Bakhtin . Cambridge, MA: Harvard
University Press.
De Certeau, M. (1984) . The practice of everyday life. Berkeley, CA: University of
California Press .
Duncan, H. D. (1989). Communication and social order. New Brunswick, NJ:
Transaction Publishers.
Freire, P. (1970) . Pedagogy of the oppressed. New York: Seabury Press.
Freire, P. (1973). Education for critical consciousness. New York: Seabury Press .
Freire, P., & Macedo, D. (1987) . Literacy: Reading the word and the world. Hadley,
MA: Bergin and Garvey.
Giddens, A. (1989) . Constitution of society. Cambridge, England : Polity Press.
Goody, J. (1977) . The domestication of the savage mind . Cambridge, England:
Habermas , J. (1984). The theory of communicative action (Vol. I). Boston, MA:
Beacon Press .
Holquist, M., & Emerson , C. (1986) . Glossary for The dialogic imagination (pp. 423 434). Austin: The University of Texas Press.
Laboratory of Human Activity. (1993). La educaci6n de adultos y el cambio de valores,
actitudes y modos de pensamiento en la mujer: Un estudio etnognifico del discurso
[Adult education and change of values, attitudes, and ways for thinking: A discourse
ethnography study]. Memoria de Investigaci6n. Madrid, Spain: CIDE.
458
Levinson, S. C. (1983) . Pragmatics. Cambridge, England: Cambridge University

Pres s.
Mead, G. H. (1934) . Mind, self, and society from the standpoint of a social
behaviorist. Chicago : University of Chicago Press .
Ong, W. J. (1982) . Orality and literacy. The technologizing of the word. London:
Methuen .
Philips, S. U. (1987) . The concept of genre and the study of language and culture.
Working Papers and Proceedings of the Center for Psychosocial Studies (No. 11) .
Chicago: Center for Psychosocial Studies.
Rogoff, B. (1990) . Apprenticeship in thinking. New York : Oxford University Press.
Voloshinov, V. N. (1986). Marxism and the philosophy of language. Cambridge, MA:
Voloshinov , V. N. (1987). Freudianism. A critical sketch. Bloomington, IN: Indiana
University Press .
Vygotsky, L. S. (1978) . Mind in society. The development of higher psy chological
processes. Cambridge, MA: Harvard University Press.
Vygotsky, L. S. (1987). Thinking and speech (N. Minick, Trans). New York: Plenum.
(Original work published 1934)
Wertsch, J. V. (1985) . Vyg0tsky. The social formation of mind. Cambridge, MA:
Wertsch, J. V. (1991) . Voices of the mind. A sociocultural approach to mediated
action. Cambridge, MA: Harvard University Press.
Wertsch , J. V., Minick, N., & Arns, F. J. (1984) . The creation of context in joint
problem solving . In B. Rogoff & J. Lave (Eds.), Everyday cognition: Its
development in social context (pp. 151-171) . Cambridge , MA: Harvard University
Press .
Wertsch, J. V ., Tulviste, P., & Hagstrom, F. (1993). A sociocultural approach to
agency. In E. Forman, N. Minick, & C. A. Stone (Eds.), Context f or learning :
Sociocultural dynamics in children development (pp. 336-356). New York : Oxford
University Press .
Author Index
Abraham, C. S., 145, 168

Allerton, M., 304, 307, 311
Alter, N ., 273, 285
Amendola, S., 407 , 409 , 413, 441
Amsel, E., 364, 371, 384
Anderson, A., 174, 199,227,237
Anderson, B., 42, 61
Anderson, R. C., 341, 358
Andre, J., 337, 358
Appel, K., 270, 285
Apter, A , 172, 199
Argles, J., 145, 168
Armbruster, B. B., 353, 356
Armstrong, J. 0 ., 353, 356
Arns , F. J. , 446, 458
Atkinson, J. M ., 145, 168
Atkinson, M ., 54, 59, 68, 85
Atkinson, R. C., 387 , 404
Atkinson , R. L., 387, 404
Auchlin, A., 95, 96, 105, 106, 109
Audetat, M., 280, 285
Azmitia, M., 364, 365 , 283
Baddeley, A, 340, 356
Baker, G. P., 133, 137
Baker, L., 338, 356
Bakhtin, M. M., 11, 19, 63, 73, 82,
85,88, 172, 173, 184, 199, 204,
205,222,411 ,440,443,448,450,
457
Balmelli, A, 271, 286
Bangert-Drowns, R. L., 275, 285
Barbieri, M . S., 229, 237
Barker, K. G., 338, 357,374
Barley, S., 42, 59
Barnes , P., 230, 238
Bateson , G., 448, 457
Bauman , R., 182, 199,451 ,457
Beach, K., 250, 264
Beishuizen, J., 338, 358
Bell, N., 269, 286
Belyaeva, A , 227, 237, 270, 285
Bern, D. J., 387, 404
Bergqvist, K., 12, 112, 386, 394, 404
Berk, L. E ., 372, 379, 383
Berkowitz, M. W ., 362, 363, 367,383
Berlin, B., 111-114 , 120,129,131,
132, 134-136, 137
Berstein , B., 124, 137
Besnier, N., 205, 222
Bessiere, c, 272, 287
Bibby, A , 232, 237
Billig , M., 8, 18,19, 145,168,408,
440,443,446,453 ,457
Bivens, J. A , 379, 383
Blanche-Benveniste, C., 205, 222
Blaye , A , 227-228, 229 , 236, 237,
238
460
Blom, J.-P., 170, 199
Blomberg, J., 42, 59
Bower, G. R., 339, 358
Bowers, 1., 42, 59
Brackenridge, A., 232, 237
Brannock, J., 364, 384
Brassae, C., 89, 94, 95, 108, 110
Brewer, W. F., 337, 358
Brigg, C. L. , 450, 457
Britten, T., 338, 357
Brown, J. S., 65, 85 , 296, 310, 339,
350,356
Brown, P. J., 338,356
Brun-Cottan, F., 42, 43, 45, 51, 57,
59, 123, 137
Bruner, J., 112,137,407,438,440
Bruner, J. S., 9, 19, 115,137,138
Buchanan, K., 158, 168
Buckingham, N ., 226, 238
Burke, K., 443, 446, 447, 457
Burton, R. R., 296, 310, 350, 356
Butor, M ., 280, 285
Butterworth, G., 232, 236, 238, 269,
286
Buttny, R., 145,168
Byrne, R. M. J., 90, 108
Cala, M. J., 457
Campbell, D. T., 391, 404
Campion, P., 64, 85
Carey, S., 337, 356
Carmel, E ., 338, 356
Caron, J., 408, 440
Carraher, D. W., 14, 19,292,295,
296,299,300,310,311
Carraher, T. N ., 291, 292, 294, 296,
299, 310
Cesareni, D., 15, 19
Chafe, W. L. 338, 356
Chaiklin, S., 115, 137
Chalfont, B., 266, 286
Chambres, P ., 236, 237
Chen, R., 338, 356
Chavez, L. R., 171, 199
Author Index
Chi, M. T. H., 337, 356
Cicourel, A. V., 112, 121, 137, 173,
199
Clark, A., 59
Clark, R. R., 314, 334
Clark, K., 447, 457
Clement, J., 316, 334
Cody, D., 337, 358
Cohen, A. R., 380, 383
Cohen , R ., 270 , 285
Colbourn, C., 228 , 238
Cole, M., 3,19,117,128,137,174,
202, 245, 246, 264, 270, 285, 304,
310, 389, 404
Collier, J., 144, 151, 168
Collins, A., 356
Collins, H. M., 117, 137,339,356
Condor, S., 145, 168
Conklin, R., 114, 132, 137
Conklin, J., 338, 356
Corsaro, W. A., 408, 440
Coulmas, F., 82, 85
Coulson, R. L., 337, 358
Crawford, S. F., 338, 356
Crook, C. K., 228, 237
Cuisset, P., 273, 285
Culley, L., 232, 237
Curnock, D., 164, 168
Damasio, A. R., 136, 137
Damasio, R., 136, 137
Damon, W. , 363, 364, 383
Davillard, C., 91, 108
Davis, J. R., 381, 383
De Certeau, M., 454, 457
de Kleer, 1., 350, 356
Demuth, C., 179, 199
De Young, L., 338, 356
Diaz, R. M., 372 , 373, 383
DiBello, L., 245, 258 , 259, 264
Dickson, L., 305, 307, 310
diSessa, A., 337, 356
Djakow, 391, 404
Doise, W., 225, 226 , 237 , 269, 285
Author Index
Donaldson, M ., 226 , 232, 237
Douady, R , 307 , 310
Drew, P., 64, 85
Droogers, A., 172, 199
Dubuisson, C., 280, 285
Duchastel, D. D ., 356
Dugdale, A., 42, 59
Duguid, P., 65, 85, 339, 356
Dunbar, K., 364, 371, 380, 383
Duncan, H. D., 447, 456
Duranti, A., 11, 16, 170, 172, 175,
181-183, 186, 199,200,408,440
Edwards, D., 10, 19, 145, 168,269,
285,389,394,404,407,408,417,
440
Eibl-Eibesfeldt, 1., 190, 199
Elbers, E., 269, 285
Ellis, S., 362 , 383
Emerson, C., 450 , 457
Engestrom, Y., 115 , 128, 137
Engle, R A., 312, 314, 331, 334
Epstein, W., 338, 357
Ericsson, K. A., 379, 383
Evans, J. S. B. T ., 90, 108
Faggin, E, 274 , 285
Falzon, P., 89, 109
Fanshel, D., 417 , 440
Fasulo, A., 11, 16, 17,206,210,212,
222,407,409,411 ,413,441
Fay, A. L. , 364 , 371, 380 , 383
Feltovich, R L., 337 , 358
Fiedler, K., 431 , 442
Filippi, G., 42, 59
Fish, R., 266, 286
Flores, E, 64, 86, 388, 405
Foot, T ., 228, 238
Foppa, K., 64, 85
Forbes, K., 59, 137
Foucault, M ., 218, 222
Forman, E. A., 236, 237
Forsythe, D., 42, 59
Frederiksen, J. R., 354, 358
461
Freire, P., 443, 444, 457
Fujimura, J., 42, 59
Furnas, G. W., 355, 356
Gaik, E, 216, 222
Gal, S., 170, 200
Galegher, J., 266 , 286
Gane, M ., 145, 168
Gardner, H., 289, 310
Garfinkel, H., 42, 49, 59 , 61, 63, 85 ,
88, 108, 112, 137
Garner, R , 341, 356
Garrod, S. C., 341, 356
Garvey, C., 408, 440
Gaudin , T., 268, 285
Gauvain, M., 362, 383
Gay, J., 245, 264, 304, 310, 389, 404
Gegeo, D., 179,202
Ghiglione, R., 90, 94, 95, 106, 108
Gibbs , J. C., 362, 363, 367, 383
Giddens, A., 448, 457
Giles, H., 363, 383
Gillet, B., 273 , 285
Gillett, G., 389, 404
Gilmour, A., 226 , 238
Girardet, H., 364, 384, 408, 410 , 441
Glaser, R., 357,356
Glenberg, A. M ., 338, 357
Glick, A., 389, 404
Glick, J., 13, 245, 250, 256, 259, 264
Gloor, P. A., 337, 338, 357, 358
Goffman, E ., 9, 19,82,85, 95, 105,
108,115,137,160,168,205,222,
407,440
Goldman, S. V., 174,201
Gonzales, P., 24, 40
Goodenough, W.H., 114, 137
Goodman, N ., 3, 19
Goodwin, C., 4, 7, 16,24, 36,40,42,
45,47,48,51,54,59,64,76,85,
112 ,118,134,135,137,170,173,
200,387,392,407,408,440
Goodwin, M . H., 24, 36, 40, 42 , 45,

47,48,50,51,54,56,59,64,85,
462
Author Index
112, 115, 131, 137, 138, 173, 200

Goody, J., 174,200,443,446,457
Gorsuch, C., 226, 238
Graaf, W. de, 391, 404
Grabinger, R. S., 338, 341, 357
Grando, N., 300, 310
Gras, A., 274, 285
Gray, S. H., 344, 357
Greatbatch, D., 64, 85
Green , C., 337, 357
Greenfield, P., 13,19
Greenfield, P. M., 112, 137, 138, 270,
285
Greenhough, P., 232, 237
Greeno, J. G., 312, 314, 315, 316,
331,334
Greenspan, S. L., 339, 358
Grice, H. P., 380, 383
Griffin , P., 227, 237
Grillo, R. D., 171, 200
Grimshaw, A. D., 409, 440
Groot, A, de, 391, 404
Grossen, M ., 7, 13, 269, 274, 286,
287
Grosslight, L., 313, 334
Grusenmeyer, C, 4, 7, 91, 108
Guggisberg, V ., 271, 286
Gumperz, J. J., 170,199,205,222
Gustavsson, L., 402, 404
Guthrie, J. T., 338, 357
Healy, L., 228, 237

Heath , c, 7,42,47,48, 52, 54, 58,
60, 61, 64, 67, 76, 85, 112, 115,
116, 138
Heath, S. B., 174,200
Heidegger, M ., 112, 115, 138
Heritage, J., 54, 59, 64, 68, 85, 90,
108,112,115,117,121,124,133,
138, 145, 168
Herskovits, M. J., 172, 200, 391, 404
Hewitt, J. P., 218, 222
Hill, J., 16, 19,170,172,173,182,
200, 204, 205, 222
Hill, K. C., 170, 172, 173, 182, 200
Hindmarsh, J., 42, 60, 64, 67, 85
Hoc, J. M ., 89, 108
Hoekstra, T., 269, 285
Holowchak, M., 16, 19
Holquist, M., 447 , 450, 457
Holt, R. W., 383
Hoogsteder, M ., 269, 285
Howe, C., 228, 237
Hoyles, C., 228, 237
Hughes, J., 42 , 43, 60, 64, 85
Hughes , M ., 232, 237, 290, 310
Hughes, S., 88, 108
Huguet, P., 236, 237
Hutchins, E ., 4, 6, 24, 40, 42, 60, 64,
85, 115, 133, 138
Hymes, D., 170, 182,200
Habermas, J., 451, 457

Hacker, P. M., S., 112, 133, 137, 138
Hacker, W., 353, 357
Hagstrom, E, 449, 458
Haken, W., 270, 285
Hampson, S. E., 145, 168
Hanks, W . F., 126, 138, 170, 172,
173,200
Harper, R., 42, 43, 60, 64, 85
Harre , R., 389, 404
Hartland, J., 60
Hastie, R., 88, 108
Hatano, G., 293, 303, 310
Inhelder, B., 304, 311

Irvine, J., 16, 19,204,205,222
Irvine, T., 451, 457
Iser, W ., 82, 85
Jacobson, M. J., 337, 358
Jacoby, S., 16, 19, 24,40
Jefferson, G., 122, 138, 139, 154,
168, 182, 202, 206 , 208, 209, 222,
223
Jirotka, M ., 42, 60, 64, 67, 85
Johnson-Laird, P. N., 90, 108, 110,
339, 340, 341, 357
Author Index
John-Steiner, V., 3, 19
Joiner, R., 228-230 , 238
Jones, A, 228, 238
Jonassen, D. H., 337, 338, 341, 357
Jonsson, L., 389, 404
Jordan, B., 42, 50, 51, 55, 56, 59, 60 ,
61, 137
Karsenty, L., 89, 109
Kay, A, 283, 286
Kay, P., 111-114, 120, 129, 131, 132,
134-136, 137
Kaye, K., 407 , 440
Keating, E ., 173, 200
Keisler, S., 381, 384
Keller, C., 4, 19
Keller, J. D., 4, 19
Kendon,A, 76, 85,173,200
Kenrick, D. T ., 388, 404
Kerr, L. K., 312 , 314,331 ,334
Kiesler, S., 273, 287
Killen, M., 363, 383
Kindred, J., 245, 258, 264
Kintsch, W., 338, 339, 358
Kirshner, D., 315, 334
Klahr, D., 364, 371, 383
Klausen, T., 24, 40
Knorr-Cetina, K., 42,61,117,138
Kobasigawa, A, 338, 357
Koffka, K., 389,404
Kohn, AS., 312, 313 , 317, 333, 334
Kosslyn, S. M., 340 , 357
Kraut, R., 266, 286
Kroskrity, P. V., 170,200
Kruger, A C., 362-364, 373, 383
Kuhn , D., 364, 371, 384
Kulick, D., 172, 200
Labov, W., 411 , 417, 440
Lacoste, M., 89, 109
Landow, G., 337, 358
Larkin, 1. L., 315, 334
Larrue, J., 90, 94, 95, 96, 97, 106,
109, 110
463
Latour, B., 42, 46, 61, 115, 117, 138
Lave, J., 4,9, 13, 19,54,61, 115,
133, 137, 138, 139 , 174, 180, 201,
246,261 ,264,289,294,310,311,
389,402,404,4]8,440
Lawler, R. W., 296, 3]0
Lee, B., 204, 222
Leech, G., 204, 222
Leichter, H. J., 174, 201
Lemke, J., 401 , 404
Lerner, G. H., 153, 168
Levine, J. M., 269, 287
Levinson, S., 88, 89, 109, 409, 437,
440,452,458
Levy, P., 273, 286
Light, P., 11, 226-230, 232, 234,
236, 237, 238, 269, 272, 286, 289,
304,306,307,310,311,381,384
Linell, P ., 389, 402, 404
Littleton, K., 11, 230, 232, 238, 272
Livingston, E., 42, 61
Lodge, D., 82, 85
Lucarella, D., 338, 357
Luff, P., 42, 47 , 48, 52, 58, 60, 61,
64, 67, 85, 112, ] 15, 116, 138
Luria, A R., 288 , 310 , 391, 392, 404
Lynch, M., 42, 61, 112, 115,117,
121, 126, 135, 138, 139
Macauley, R. K. S., 205, 222
Macedo, D ., 444 , 456
Mackenzie, M ., 227, 237
MacKinlay, D. R. E ., 145, 168
Macpherson,
170, 201
Maier, R., 269, 285, 391,404
Mandl, H., 337, 357
Mannheim, B., 171,201
Markova, I., 64, 85
Martin, L. M. W., 250, 264
Marton, F., 341, 357
Mason, J., 269 , 286, 304, 306, 307,
311
Matz, M., 315, 334
Mayer, R. E ., 354, 357
c.
464
McAleese, R., 337, 357
McCaffrey, M., 232, 239
McDermott, R. P., 174, 200, 246 ,
264
McClelland, I., 228, 238
McGrath, D., 338, 357
McNeill, D., 128, 139
Mead, G. H., 2, 19, 448, 458
Melton, R. F., 341, 358
Mercer, N., 228, 238 , 269, 285 , 394,
404, 407, 440
Merleau-Ponty, M. , 388, 404
Messer, R., 230, 238
Mevarech, Z. R., 228, 238, 272, 286
Meyer, B. J. F., 355, 358
Middleton, D., 10, 19, 115, 137, 145,
146, 158, 164, 168,389,404,407,
440
Miller, P., 179, 201
Minick, N ., 236, 237, 446, 457
Moeschler, J., 95, 96, 105, 106, 109
Mohageg, M. F ., 344, 358
Monteil, J. M ., 236, 238
Montgomery, R., 364, 365, 283
Moore, J. L., 312, 314, 331, 334
Moore, W. S., 119, 139
Morgan, M., 183, 201
Moricot, C., 274, 285
Morrow, D . G., 339, 358
Mosenthal, P., 338, 357
Mugny, G., 225, 226, 237, 269, 285
Mulkay, M., 42, 61
Nathan, M. J., 312, 313, 317, 333,
334
Neisser, D., 389, 404
Newman, J., 226 , 238
Nicholls, G., 7
Nicolet, M ., 269, 286
Nisbett, R. E. , 90, 109
Nordon, C., 270, 285
Nunes, T ., 14, 19,269, 286, 289,
292, 295 , 300, 301, 304, 306, 307,
310, 311
Author Index
Ochs, E ., 11, 16, 19,24,40, 170,
172,175-177,181-183,199,201,
202,407,409,410,411 ,413,415,
416,440,441
O'Donnell, A , 353, 358
O'Laughlin, M ., 364, 371, 384
Oliver, R. R., 112, 137
Ong, W. J., 443, 445 , 446 , 457
Orr, J., 42, 61
Orsolini, M., 407, 409 , 410, 437, 441
Palen, L. , 4, 6
Palmer, S. E., 340, 358
Papert, S., 270, 286
Parsons, T., 117, 119, 139
Pea, R. D., 270, 271, 286
Pennington, N ., 88, 108
Perakyla, A, 153, 164, 168
Perret, J.-F., 269 , 286
Perret-Clermont, A-N., 225, 226,
237-239,269,285-287,289,310,
381, 384
Perriault, 1., 13, 19, 268 , 286
Perrin-Glorian, M. J., 307, 310
Phelps, E. , 363, 364, 383
Philips, S., 451, 457, 458
Phillips, T., 228, 238
Piaget, J., 89, 90, 109, 225, 239, 291,
304 ,311,362,384
Pichert, J. W., 341, 358
Pickering, A , 42,61,115,117,139
Pinker, S., 341, 358
Plunkett, S., 296, 311
Pochon, L.-O ., 7, 13, 270, 274, 286
Podak, C., 266, 286
Poirot-Delpech, S. L., 274, 285
Polanyi, M., 118, 139
Pomerantz, A , 408, 441
Pontecorvo,
15-17, 19,364,380,
384 ,407-411,413,417,437,441
Potter, J., 145, 168,417,440
Pozzi, S., 228, 237
Radley, A , 145, 168
Ramirez, J. D., 18
c,
465
Author Index
Randall, D., 42, 43, 60
Reed , H. J., 294, 311
Rees, E ., 337, 356
Reis, R., 341, 356
Resnick, L. B., 5, 16, 19, 88, 109,
269,287,312,313,317,333,334,
366, 384, 407, 441
Retornaz, A, 90, 110
Rheingold, H., 275, 276, 287
Rivers , W. H., 389, 404
Rizk, A., 337, 358
Rizzo, T. A., 408, 440
Robbins , H., 226, 238
Rock, I., 387,388,389,404
Rogard , V., 89, 109
Rogoff, B., 115, 133, 139, 178,201,
268, 287, 362, 383, 446, 458
Romaine, S., 171, 201
Romanelli, P., 15, 19
Rommetveit, R., 269, 287, 402, 404
Rosaldo , R., 171,201
Roschelle, J., 313,314,327,334,
366, 380, 384
Roskill, M., 120, 139
Ross , L., 90, 109
Roulet, E., 95, 96, 105, 106, 109
Rubattel, C., 95, 96, 105, 106, 109
Rubtsov, V., 227, 228, 237, 239
Rudik, 391,404
Rudolf, D., 175, 201
Rudolph, D., 407, 413, 441
Ruhleder, K., 42, 61
Sabel , C., 250, 264
Sacks, H., 8, 10, 19,49,59, 66, 77,
86 , 112, 115, 121, 122, 133, 139 ,
145,158,165,168,182,202,208,
209,212,222,223,408,437,441,
442
suje, R, 17, 112,269,287,341,
357,386,389,394,402,404
Salmon , M. H., 16, 19

Samarapungavan, A, 338, 358
Saussure, F. d., 113, 114, 117, 119,
139
Saxer, K. H., 338, 358
Scardigli , V., 274, 285
Schaefer , E. E , 314,334
Schauble, L., 366, 367, 371, 380, 384
Schegloff, E. A, 96, 109, 112, 115,
122, 124, 133,139, 182,202,208,
209,212,222,223,372,384,407,
408,417,429,442
Schegloff, M., 54, 61
Schelling, M., 95, 96, 105, 106, 109
Scheper-Hughes, N., 127, 139
Schieffelin , B. B., 16,19,170,177,
179, 201, 202
Schiffrin, D., 408, 417, 442

Schliemann, A D., 14, 19,292,295,
296,299, 300, 31~ 311
Schnotz, W., 15
Schoenfeld, A H., 317, 334
Schubauer-Leoni, M. L., 269,287
Schwarz, B. B., 14,312,313,317,
333,334
Scribner, S., 13,19,135,139, 174,
202, 245, 246, 248-251, 253, 258,
264
Seamon, J. G., 388, 404
Searle, J., 94, 95, 109
Segall, M. H., 391, 404
Semin, G., 431, 442
Shapin, S., 124, 139
Sharrock, W., 42,43 ,60, 61
Shapiro , D., 42, 43, 60
Sharp, D., 245, 264
Sharp, D. W., 389, 404
Sheldon, S ., 228, 229, 237
Sherzer , J., 182, 199
Shills, E., 117, 139
Shore, B., 182, 186,202
Shotter, J., 392, 404
Silverman, D., 164, 168
Simon, H. A , 3, 19, 379,383
Smith, C., 313, 334
Smith, E. E., 387, 404
Smith, J. B., 338,358
466
Smith, N., 407, 413, 441
Smith, R, 175, 201
Snir, J., 313, 334
Soldatova, G., 228, 237
Somekh, B., 228, 238
Spiro, R. J., 337, 358
Spitzer, C. E., 381, 383
Sproull, L., 273, 287, 381, 384
Star, S. L. , 42,61,117,123,139, 140
Stasser, G., 381, 383
St. Clair, R. N., 363, 383
Stone, C. A , 236, 237, 381, 384
Streitz, N. A , 337, 357, 358
Suchman, L. , 4, 6, 42, 45, 50, 52, 54,
58,59, 61, 62, 64, 86, 88, 109,
112, 115, 116, 123,137, 140
Szeminska, A, 304, 311
Tannen, D ., 205, 223
Taylor, C., 175,201,407,409,410,
413,415,416,441
Teasley, S., 15, 16, 228, 269, 287,
366, 371, 380, 384
Tetlock, P. E., 381, 384
Theureau, J., 42,59
Thurstone, L. L., 289, 311
Todorov, T., 82, 86
Tolmie, A, 228, 237
Tomasello, M., 362, 364, 373, 383
Traweek, S., 42, 61
Trigg, R., 42, 59, 62, 137
Trognon, A, 4, 7, 89, 90, 94-97, 106,
108,109,110
Tudge, J., 364, 384
Tulviste, P., 449, 458
Turner, V., 206, 223
Underwood, G., 232, 239
Underwood, J., 232, 239
Vacherand-Revel, J., 272, 287
Vanderveken, D., 94, 95, 109, 110
Van Dijk, T. A , 338, 339, 358
Varenne, H., 174,201
Author Index
Vergnaud, G., 307, 311
Vitale , B., 270, 287
Voloshinov, V. N., 172,202,204,
223,447,458
Vosniadou, S., 337, 358
Vygotsky, L. S., 2, 3, 16, 19,86,
172,202,268,287,288,311 ,407,
442,449,458
Waern, Y., 266, 271, 287
Walkerdine, V., 394, 405
Wallace, T. E., 253, 264
Wang, S., 338, 357
Wason, P. C., 90, 110
Wathen, S. H., 16, 19
Watson, A R, 144, 145, 151, 168
Watson-Gegeo, K., 179, 202
Watt, 1., 174, 200
Weiss , J., 280, 285
Weiss, S., 338, 358
Wenger, E., 4,9, 13, 19, 54, 61, 133,
138, 180, 201, 261 , 264, 402, 404,
418 ,440
Wertheimer, M., 389, 405
Wertsch, J. V., 4, 16, 18, 19,82, 86,
268, 287, 380, 381, 384, 386, 391 ,
405 ,443,446,449,451,458
Wetherell, M., 145, 168
Whalen, J., 42, 57, 58, 61, 62, 64, 86
White, B. Y., 14, 19,313,334,354,
358
Wilkes -Gibbs, D., 314, 334
Wilkinson, A C., 338, 357
Winograd, T., 64, 86, 388, 405
Wiser, M., 313, 335
Wistedt, 1., 399, 405
Wittgenstein, L., 111, 112, 115, 129,
133, 136, 140
Wooigar, S., 42,61 , 115, 139
Yankelovich, N., 337, 358
Yates , J., 45, 62
Zajonc , R. B., 380, 384
Author Index
Zazanis, E., 245, 250, 255, 256, 258,
264
Zeitz, C. M., 16, 19
Zentella, A. C., 171,202
Zimmerman, D., 42, 58, 62
467
Subject Index
Accorrrrnodation, 177, 191, 192, 197

Accountability, 16-18,42,49,50,58,
129, 145, 205
Accountable knowledge, 129, 131, 134
Airline cockpit, 24-30, 35, 39, 54
Algebraic symbols, 315, 316,321
Anthropology, 41, 112, 120, 126,
129, 173
Apprenticeship, 54, 55, 111, 116,
122, 133, 407, 446
Area concepts, 292, 304-307, 309
Argumentation, 443, 448, 449, 452454
Argumentative reasoning, 408-410,
412,419,433,438
challenge , 406, 409, 410, 412,414,
417,421,422,424,425,427,429,
435,437-439
defense, 408, 410, 412, 414, 417,
418,421,437,438
Arithmetic
oral, 292,294-303, 305, 308, 309
written, 292-298, 300-303, 305,
308, 309
Attunement, 314, 317, 332
Authoritative knowledge, 44, 55, 56
Autobiography, 205, 206
Balance Beam, 228
Caregiving, 169,176-179,192
hierarchical organization of, 176180, 192, 194, 196
socially distributed, 176-181, 191,
192, 194
Centers of coordination, 41-43, 45,51 ,
57,116
Chemistry, 445
Classification
processes of, 112-115, 126, 129,
135, 136
Classroom interaction, 393-398
Coaction, 230, 232-234
Co-construction, 363, 379
of knowledge, 229
Code-switching, 170, 171, 105, 243,
255-258
Cognition, 2, 64, 112-116, 124, 127,
133-136,236,246,249,250, 389,
390,402
developmental, 243, 245-248, 261263
distributed, 5-8, 12, 15-17, 24, 90,
106, 115, 133
embodied, 118, 127-131, 134, 136
situated, 1,3,4,9, 12, 15-17,24,
30,31 ,39,115,117,128,246-250,
254, 261, 262, 289-292, 303, 304,
387, 389, 396, 398, 407
470
situated action in, 112, 115, 116,
126, 128, 134
Cognitive
conflict, 362, 363, 365
development, 265, 267-269, 274,
279
Collaboration (see also Learning,
Problem Solving), 16, 63, 66, 74,
82, 90, 181, 224, 225, 227
peer, 361-364, 367, 369,371-373,
377-381
Color categories, 111,112, 114, 115,
121, 124, 131, 132-135
Communities
of innovation, 254, 262, 263
of practice, 111-113, 117, 128, 132134, 250, 252-254, 398
Competent practitioner, 44, 47, 116
Computer
use of, 265-267, 270-276, 278, 281
representations, 272, 274
Conceptual
change, 394,403
integration, 255, 259
knowledge, 391, 396,402
learning, 396, 401
models, 312, 313
Context, 224, 236
Contextualization, 451, 452
decontextualization, 448-450, 452,
456
Contingency, 44, 56-58
Control,49
Control room , 42, 43, 45-47, 50-54,
64, 67
Conversation (see also Dialogue,
Discourse , Talk), 90
analysis, 68, 88, 90, 91, 94-106,
112, 115, 133, 145, 182
family, 407, 409-411, 413, 414,
418-438
structure of, 90, 106
transactive, 15
Subject Index
Conversational processes, 314
of confirmation, 312, 329-331, 333
of convergence, 312, 327, 329-331,
333
of creation, 312, 329
turn-taking, 282, 312, 314, 327,
329, 331, 333
Debates, 443, 444, 446, 450-452 , 454,
455
Decomposition , 296, 297
Design, 43, 44, 57, 58
Design-in-use, 44, 56, 57
Dialogue, 443, 445, 449, 450, 453456
Didactic contract, 269
Dilemmas of uncertainty , 143, 146,
157, 158, 163
circumstance, 144, 146, 149, 153,
155, 160, 165
ordinariness/extraordinariness of
family life, 143, 144, 146, 151,
15~ 15~ 158, 16~ 16~ 165
outcome, 144, 146, 153, 155, 157,
158, 165
reasonableness , 144, 158, 164, 165
trustworthiness, 146, 149, 159, 162,
164
using technologies of care, 144,
151, 161
Discourse (see also Conversation,
Dialogue, Talk), 2, 3, 5, 17, 29, 38,
47-49,175,204,205,207,221,
243, 246, 256, 262, 263, 385, 398,
443,444,447,453
analysis, 8, 9, 206, 407, 452
classroom, 392, 395-397,400,401
communities,S, 9, 10, 17
genres, 406, 411, 412, 443, 449,
450, 454-456
narrative, 31, 32,34, 38
oral, 444, 445, 448
patterns of, 4
471
Subject Index
roles, 409, 411-413, 415
written , 444, 445, 448
Distributed activities, 43-45, 46, 50,
51,53,56
Division of labor, 42, 43, 53
Environment
home
cultural organization of, 169, 183,
189,197
spatial organization of, 183, 184,
186, 196
work, 45, 51-53, 55, 56, 67, 83,
254
technology-intensive, 42, 44, 57,
91,92
Ethnography, 255, 258, 444
of science, 117, 119, 124, 128, 134,
182
Ethnomethodology, 49, 87-89, 94,
112, 121, 145
Ethnotechnology, 13,41-42,265,
267-269, 274
Experimental contract , 269
Expertise
transmission of, 280
Informings, 50, 75, 80, 81

Interaction, 43, 44, 47, 51, 53-55, 57,
63, 64, 66, 73-76, 80-83, 87-89, 91,
105,121 ,133,170,173,175,176,
178, 181-183,224,225,227,229,
230,232-236,393-398
human-machine, 42, 265-268, 270,
271, 274, 275, 281, 283
peer, 224, 225, 228, 236, 269
technology-mediated, 43-45, 47, 50,
56, 64, 65, 81, 83, 93
Interactional analysis (see also Talk),
145, 146, 15~ 154, 157, 16~ 162
collaborative completion of talk,
146, 152, 153
talk in parent support groups, 143150, 153-165
Interactional order, 43, 51, 54, 95
Interactive space, 265, 267, 274, 275,
282, 283
Interlocution, 87, 94, 95, 97, 105, 170
Intermediate abstraction, 14,312-314,
317,320,326,331 ,333
Intersubjectivity, 267, 269-272, 282,
363
Intertextuality, 203, 204
faaaloalo [respect], 178

faapalagi [Western way], 175
fa'aSamoa [Samoan way], 175
Knowledge acquisition, 337-339, 345,

353-355,407,408
Gender, 224, 230, 232-236

Genre, 82, 117, 172
Gesture, 29, 31-35, 37-39, 126, 128,
135
Learners
Language socialization , 170, 175
Habitual spaces, 53, 54

Heteroglossia, 11, 172, 173,204
Highlighted, 122, 123, 125, 126, 134,
135
Hybridization, 172, 173, 194,204
Hypermedia, 228, 337
Hypertext, 15, 273, 280, 336-339,
341, 342, 344-346, 352, 354, 355
successful, 336, 339, 346, 347, 349353

unsuccessful, 336, 337, 346, 347,
349-353
Learning, 5, 12, 55, 57, 267, 269,
361-364,394,399,402,403
as apprenticeship, 13, 116
to argue, 406, 408
collaborative, 312-314, 317, 320,
323, 326, 327
optics, 385, 386, 402
physics, 386, 393, 396, 398, 399-
Subject Index
472
402
to reason, 15
tasks, 339, 354, 355
Linearity/nonlinearity, 336, 338
Linguistic markers, 243
Literacy, 173, 174, 246, 444-446,
452,455
activity, 170-175
church-based, 175, 182
home-based, 175, 182
syncretic, 169, 170, 172, 173, 175,
182, 183, 191
Material practice, 44, 45
Materials Requirement Planning
(MRP), 250, 251-257, 259, 260,
262
Mathematical operators, 248, 249,
312,315,317,319,321,327,331,
333
Measures
isomorphism of, 307-309
product of, 307-309
Mediated agency, 448, 449
Mental models, 340, 341, 344, 350,
354
analysis of, 336, 341, 342, 348,
349, 352
construction of, 336, 339, 341-343,
348-350,352-354
model-based strategy, 337-339, 341343, 348, 352, 353
Mexicano speakers, 173
Microworlds (see also OPERA,
Planner system), 366, 367, 371,
372, 380
Modification of use, 273, 274, 276279
Multiculturalism, 11, 169-173, 175,
176, 183, 196, 197
Multilingualism, 170, 172, 173, 175,
196
Mutual intelligibility, 49
Mutual orientation, 53
Nahuatl , 173
News , 65-67, 70, 80
Normal order, 44, 49
OPERA, 313, 317-323, 327, 331-333
Opposition, 409, 410, 416, 417, 429
counter opposition, 410, 417, 418,
427,433,435-437
Organizational routines, 50
Organizationally relevant objects, 57,
58
Partial perspectives, 49, 52
Past events reconstruction, 406, 410,
411,416,417,419,421 ,427,433,
437,438
Perception , 17, 112, 114, 116, 119,
121, 123-126, 131, 135, 145,385393,396,402
visual, 386, 387, 393
Perceptual development, 391, 403
Peripheral monitoring, 51
Peripheral participation, 418, 427, 437
Philosophy , 117
Piagetian theory, 225, 228, 267-269,
362, 363
Planner system, 313, 317
Pragmatics, 89
Problem-centered topics, 411, 412,
437,438
Problem resolution, 87, 90, 91, 94106
Problem solving, 361, 363, 365, 379,
381
collaborative , 312-314,317,323,
326, 327
computer based, 224, 227, 228, 230,
234, 235, 269, 272
group, 24, 30, 31, 37, 39,47-49, 51
mathematical, 288-290, 292-294,
296, 300, 303, 306-309
Problematization, 406, 410, 412, 413416,419,421 ,422,425,437,438
Subject Index
Problematizer/problematizee, 410,
413,415-419,427,431,433,435,
437
Professional design, 56, 57
Prompting, 178, 179
Psychology, 132, 206, 224, 236, 266,
387, 444
Psychosocial support, 144, 145, 149,
151, 165
common knowledge, 146, 162
formulating memberships, 146, 147,
149, 150
paediatric renal care, 144-146, 149151, 153, 154, 159, 160, 164, 165
problematic of normality, 149, 151,
154, 163, 164, 221
Rationalilty, 90, 91
Reading a scene, 44,47
Reasoning, 3-4, 15
as action, 4
practical, 52
principles , 228-292, 296, 297, 300,
303, 304, 308
scientific , 364, 371, 380
Repeated groupings, 300
'Representational devices, 46
Rhetoric, 443, 446-448, 450, 455
Samoan American community, 169,
171
Samoan
culture, 182
language, 182
Scaffolrung, 228,407,409
Schooling, 245, 246, 261, 263
Semiotic tools, 267-270, 279, 282,
283,443,446,447,452,454,455,
458
Sensemaking, 49
Shared meaning, 312, 314, 327
Shift changeover, 91
Signed numbers, 301-303
Situated identity, 206
473
Situativity theory, 4, 5, 12
Social comparison, 225, 235, 236
Social language, 443, 450, 452, 454456
Social marking, 226
Sociocognitive conflict, 225, 226,
228, 254, 267, 269-272, 279, 283
Sociocultural, 4, 12, 170, 268, 270
experience, 388-394,396,398,402,
403
theory, 385, 386
Sociology, 41,89, 117, 124,266
Spanish, 173
Speech
direct, 204-207, 211, 213, 214
private, 372, 379
reported, 203-205, 209, 210, 213,
214, 220, 221
Stories, 65-71, 72, 73,77,78
Structures of participation, 44, 45, 50,
53, 58
Study, 445, 446
Symbol processing, 336, 340-342,
348, 349, 352, 353
Symbolic representation , 14, 36-37,
46,94-106,117,175,279,294,
312, 314-322, 326, 327, 332, 333
non-symbolic representation, 320,
321, 332
Syncretism, 11, 172, 173, 176, 183,
192-194,196, 197
Synesthesia, 124, 128, 129, 134, 136
Systems of signs, 288, 289, 291, 292,
297, 303, 304, 308, 309
Talk (see also Conversation, Dialogue,
Discourse), 7, 8, 63, 66, 68-72, 79,
82,369,371-373,
accountable, 5-6, 8, 15
coherence in, 380,381
conflict talk, 409, 412, 418, 431,
437
dinner, 406, 409, 410, 412, 437
effects of, 371-373, 381
474
mediated by computers, 372, 381
therapy, 203, 207, 208, 215-218 ,
219, 220
Task completion , 181
Technology, 64, 83, 249, 250
of accountability, 50, 58
communication, 45, 47,53,58,
443, 444
information , 45
innovation, 275, 276, 283
Temporal order, 42, 44-46, 50
Territories, 43
Text, 68-72, 75, 76, 81, 82
Tools, 340
in measurement, 290, 304-309
Tower of Hanoi, 228
Transactive discussion, 361, 372, 374,
376, 378-381
categories, 362, 365
types of, 362, 363, 365
Transactive reasoning, 362-368, 372,
373, 375, 376, 378-380 .
Transformations
cognitive, 443
discursive, 443
Turn-taking (see also Conversational
Processes), 408, 431, 437
Typified action sequence, 49, 50
Uncertainty
discursive analysis of, 159-164
discursive reasoning, 145, 146, 150,
153, 154, 156, 158, 160, 162, 163,
165
Utterances (see also Verbalization),
451-454
Verbalization (see also Utterances),
367, 373, 379, 380
Vision , 112, 118
Visual behavior, 69, 72, 74-79 ,83
Vygotskian theory 2, 4, 16, 124, 128,
225,228,267,268,363,407,449
Subject Index
Word processing, 265, 267, 275, 280282
history of, 265, 275-279
Work practices, 43-45,51-53,57,6567,111,1 26,133,134,248-250,
252-254, 262
Workplace
design, 43, 53, 56, 57,65, 254, 263
learning, 243, 248, 249, 253, 261263
acquisition, 243, 248, 249
coordination of knowledge
systems, 243, 248-252, 255, 256,
258, 260
organization , 258, 260, 262, 263
constructive, 258-260, 262, 263
procedural, 258-260, 263
Workplace Project, The, 43, 126, 135
Workspaces,44, 53, 54
Written algorithms, 292-297, 300
The NATO ASI Series F Special Programme on

ADVANCED EDUCATIONAL TECHNOLOGY
67:
76:
78 :
80 :
81:
84 :
85 :
86:
87 :
89:
90:
9 1:
92:
93 :
95 :
96:
97 :
104:
105:
107:
109:
111:
112:
113:
115:
116:
117:
119:
121 :
122:
125:
128:
129:
132:
133:
135 :
137:
140 :
Designing Hypermedia for Learning

Multimedia Interface Design in Education
Integrating Advanced Technology into Technology Educat ion
Intelligent Tutoring Systems for Foreign Language Learning
Cognitive Tools for Learning
Cornputer-Based Learning Environment s and Problem Solving
Adap tive Learning Environments: Foundations and Frontiers
Intelligent Learning Environrnents and Knowledge Acquisition in Physics
Cogn itive Modelling and Interactive Environments in Language Learning
Mathernat ical Problem Solving and New Inforrnation Technolog ies
Collaborat ive Learning Through Compute r Conferencing
New Direction s for Intelligent Tutor ing Systems
Hypermedia Courseware : Structures of Comrnunication and Intelligent Help
Interactive Multimedia Learning Environments
Cornprehensive System Design: A New Educational Technology
New Directions in Educat ional Technology
Advanced Models of Cogn ition for Medical Training and Practice
Instructional Models in Computer-Based Learning Environrnents
Designing Environments for Constructive Learning
Advanced Educational Technology for Mathernatics and Science
Advanced Educational Technology in Techno logy Education
Cognitive Models and Intelligent Environments for Learning Programming
Item Banking : Interactive Testing and Self-Assessment
Interactive Learning Technology for the Deaf
Learning Electricity and Electronics with Advanced Educational Technology
Control Technol ogy in Elementary Education
Intelligent Learning Environment s: The Case of Geometry
Automating Instructional Design. Developm ent, and Delivery
Learning from Computers: Mathematics Education and Technology
Simulation-B ased Experiential Learning
Stud ent Mod elling: The Key to Individualized Knowledge-Based Instruct ion
Cornputer Supported Collaborat ive Learning
Huma n-Machine Communi cation for Educational Systems Design
Design of Mathematical Modelling Courses for Engineering Education
Collaborative Dialogue Technolog ies in Distance Learning
Technology Education in School and Industry
Technology-Based Learning Environments
Automating Instructional Design:
Computer-B ased Development and Delivery Tools
141 : Organizational Learning and Technological Change
142: Dialogue and Instruction. Modeling Interaction in Intelligent Tutoring Systems
146: Comp uters and Exploratory Learning
148: Knowledge Acq uisition . Organization, and Use in Biology

160 : Discourse, Tools, and Reasoning: Essays on Situated Cogn ition
NATO ASI Series F

IncludingSpecialProgrammes on SensorySystemsfor Robotic Control(ROB) and on
Advanced Educational Technology (AET)
Vol. 1: Issues in Acoustic Signal- Image Processing and Recognition. Edited by C. H. Chen. VIII, 333
pages. 1983. (out of print)
Vol. 2: Image Sequence Processing and Dynamic Scene Analysis. Edited by T. S. Huang. IX, 749
pages. 1983.
Vol. 3: ElectronicSystems Effectiveness and LifeCycleCosting. Editedby J . K. Skwirzynski.XVII, 732
Vol. 4: Pictorial Data Analysis. Edited by R. M. Haralick. VIII, 468 pages. 1983.
Vol. 5: International Calibration Study ofTraffic ConflictTechniques. Edited by E.Asmussen. VII, 229
pages. 1984.
Vol. 6: Information Technology and the Computer Network. Edited by K. G. Beauchamp. VIII, 271
pages. 1984.
Vol. 7: High-Speed Computation. Edited by J. S. Kowalik. IX, 441 pages. 1984.
Vol. 8: ProgramTransformation and ProgrammingEnvironments.Reporton a Workshop directed by
F. L. Bauer and H. Remus. Edited by P. Pepper. XIV, 378 pages. 1984.
Vol. 9: Computer Aided Analysisand Optimizationof Mechanical System Dynamics. Edited by E. J .
Haug. XXII, 700 pages. 1984. (out of print)
Vol. 10: Simulation and Model-BasedMethodologies: An IntegrativeView. Edited by T. I. Oren, B. P.
Zeigler, M. S. Elzas. XIII, 651 pages. 1984.
Vol. 11: Robotics and ArtificialIntelligence. Edited by M. Brady, L. A. Gerhardt, H. F. Davidson. XVII,
693 pages. 1984.
Vol. 12: CombinatorialAlgorithmson Words. Edited by A. Apostolico, Z. Galil. VIII, 361 pages. 1985.
Vol. 13: Logics and Models of Concurrent Systems. Edited by K. R. Apt. VIII, 498 pages. 1985.
Vol. 14: Control Flowand Data Flow: Concepts of Distributed Programming. Edited by M. Broy. VIII,
525 pages. 1985. Reprinted as Springer Study Edition 1986.
Vol. 15: ComputationalMathematical Programming. Editedby K.Schittkowski.VIII, 451 pages. 1985.
Vol. 16: New Systems and Architectures for Automatic Speech Recognition and Synthesis. Edited
by R. De Mori, C.Y. Suen. XIII, 630 pages. 1985. (out of print)
Vol. 17: Fundamental Algorithmsfor ComputerGraphics. Editedby R. A. Earnshaw. XVI, 1042 pages.
1985. Reprinted as Springer Study Edition 1991.
Vol. 18: Computer Architectures for Spatially Distributed Data. Edited by H. Freeman and G. G.
Pieroni. VIII , 391 pages. 1985. (out of print)
Vol. 19: Pictorial Information Systems in Medicine. Edited by K. H. Hahne. XII, 525 pages. 1986.
Vol. 20: Disordered Systems and Biological Organization. Edited by E. Bienenstock, F. Fogelman
Soulie, G. Weisbuch. XXI, 405 pages.1986
Vol. 21: Intelligent Decision Support in Process Environments. Edited by E. Hollnagel, G. Mancini,
D. D. Woods. XV, 524 pages. 1986.
Vol. 22: Software System Design Methods. The Challenge of Advanced Computing Technology.
Edited by J. K. Skwirzynski. XIII, 747 pages. 1986. (out of print)
Vol. 23: Designing Computer-Based Learning Materials. Edited by H. Weinstock and A. Bork. IX, 285
Vol. 24: Database Machines. Modern Trends and Applications. Edited by A. K. Sood and A. H.
Qureshi. VIII, 570 pages. 1986.
NATO ASI Series F

Including Special Programmeson SensorySystemsfor Robotic Control (ROB) and on
Vol. 25 : Pyramidal Systems for Com puter Vision. Edited byV. Cantoni and S. Levialdi. VIII, 392 pages .
1986. (ROB)
Vol. 26 : Modelling and Analysis in Arms Control. Edited by R. Avenhaus, R. K. Huber and J. D. Kettelle .
VIII, 488 pages . 1986 . (out of print)
Vol. 27: Computer Aided Opt imal Design: Structural and Mechanical Systems. Edited by C. A. Mota
Soares. XIII, 1029 page s. 1987.
Vol. 28: Distributed Operating Systems. Theory und Practice. Edited by Y. Paker, J.-P. Banatre and
M. Bozyigit. X, 379 pages. 1987 .
Vol. 29 : Language s for Sensor-Based Control in Robotics. Edited by U. Rembold and K. Hormann .
IX, 625 page s. 1987. (ROB)
Vol. 30 : Pattern Recognit ion Theory and Applications. Edited by P. A. Devijver and J. Kittler. XI, 543
pages . 1987 .
Vol. 31: Decision Support Systems: Theory and Application. Edited by C. W. Holsapple and A. B.
Whin ston . X, 500 pages. 1987 .
Vol. 32 : Information System s: Failure Analysis. Edited by J. A. Wise and A. Debons . XV, 338 page s.
1987 .
Vol. 33 : Machine Intelligence and Knowledge Engineering for Robotic Applications. Edited by A. K.
C. Wong and A. Pugh. XIV, 486 pages . 1987. (ROB)
Vol. 34 : Modelling , Robustness and Sensitivity Reduction in Control Systems. Edited by R.F. Curtain.
IX, 492 pages . 1987.
Vol. 35 : Expert Judgment and Expert Systems. Edited by J. L. Mumpower, L. D. Phillips, O. Renn and
V. R. R. Uppuluri. VIII, 36 1 pages . 1987 .
Vol. 36 : Logic of Programming and Calculi of Discrete Design. Edited by M. Broy . VII, 415 pages.
1987 .
Vol. 37: Dynamics of Infinite Dimensional Systems. Edited by S.-N . Chow and J. K. Hale. IX. 514
pag es. 1987.
Vol. 38: Flow Control of Cong ested Networks. Edited by A. R. Odoni , L. Bianco and G. Szeqo . XII,
355 page s. 1987.
Vol. 39 : Mathematics and Computer Science in Medical Imaging. Edited by M. A. Viergever and A.
Todd-Pokropek. VIII, 546 pages. 1988 .
Vol. 40 : Theoretical Foundation s of Computer Graphic s and CAD. Edited by R. A. Earnshaw. XX,
1246 pag es. 1988 . (out of print)
Vol. 41: Neural Computers. Edited by R. Eckmiller and Ch . v. d. Malsburg . XIII , 566 pages . 1988.
Reprinted as Springer Study Edition 1989, 1990.
Vol. 42 : Real-Time Object Measurement and Classification . Edited by A. K. Jain. VIII, 407 pages .
1988. (ROB)
Vol. 43 : Sensors and Sensory System s for Advanced Robots. Edited by P. Dario. XI, 597 pages .
1988. (ROB)
Vol. 44 : Signal Processing and Pattern Recognition in Nondestructive Evaluation of Materials. Edited
by C. H. Chen . VIII,344 pages . 1988. (ROB)
Vol. 45 : Syntactic and Structural Pattern Recognition. Edited by G. Ferrate,T. Pavlidis, A. Sanfeliu and
H. Bunke . XVI, 467 pages . 1988. (ROB)
NATO ASJ Series F

Including SpecialProgrammes on SensorySystems for RoboticControl (ROB) and on
Advanced Educational Technology (AE7)
Vol. 46: RecentAdvancesin Speech Understandingand DialogSystems. Edited by H. Niemann,M.
Lang and G. Sagerer. X, 521 pages. 1988.
Vol. 47: Advanced Computing Concepts and Techniques in Control Engineering. Edited by M. J.
Denham and A. J. Laub. XI, 518 pages. 1988.
Vol. 48: Mathematical Models for Decision Support. Edited by G. Mitra. IX, 762 pages. 1988.
Vol. 49: Computer Integrated Manufacturing. Edited by I. B. Turksen. VIII, 568 pages. 1988.
Vol. 50: CAD Based Programmingfor Sensory Robots. Edited by B. Ravani. IX, 565 pages. 1988.
(ROB)
Vol. 51: Algorithmsand Model Formulations in MathematicalProgramming. Editedby S. W. Wallace.

IX, 190 pages. 1989.
Vol. 52: Sensor Devicesand Systemsfor Robotics. Edited by A. Casals. IX, 362 pages. 1989. (ROB)
Vol. 53: Advanced InformationTechnologiesfor IndustrialMaterialFlowSystems.Edited by S. Y. Nof
and C. L. Moodie. IX, 710 pages. 1989.
Vol. 54: A Reappraisal of the Efficiencyof Financial Markets. Edited by R. M. C. Guimaraes, B. G.
Kingsman and S. J. Taylor. X, 804 pages. 1989.
Vol. 55: Constructive Methods in Computing Science. Edited by M. Broy. VII, 478 pages. 1989.
Vol. 56: Multiple Criteria Decision Making and Risk Analysis Using Microcomputers. Edited by
B. Karpak and S. Zionts. VII, 399 pages. 1989.
Vol. 57: Kinematics and Dynamic Issues in Sensor Based Control. Edited by G. E. Taylor. XI, 456
pages. 1990. (ROB)
Vol. 58: HighlyRedundantSensingin Robotic Systems.Edited by J. T. Tou and J. G. Balchen. X, 322
pages. 1990. (ROB)
Vol. 59: Superconducting Electronics. Edited by H. Weinstock and M. Nisenoff. X, 441 pages. 1989.
Vol. 60: 3D Imagingin Medicine. Algorithms, Systems,Applications. Edited by K. H. Hahne, H. Fuchs
and S. M. Pizer. IX, 460 pages. 1990. (out of print)
Vol. 61: Knowledge, Dataand Computer-AssistedDecisions. Editedby M. Schaderand W. Gaul.VIII,
421 pages. 1990.
Vol. 62: Supercomputing. Edited by J. S. Kowalik. X, 425 pages. 1990.
Vol. 63: Traditionaland Non-Traditional Robotic Sensors. EditedbyT. C. Henderson. VIII, 468 pages.
1990 . (ROB)
Vol. 64: Sensory Robotics for the Handlingof Limp Materials. Edited by P. M. Taylor. IX,343 pages.
1990 . (ROB)
Vol. 65: Mapping and Spatial Modelling for Navigation. Edited by L. F. Pau. VIII, 357 pages. 1990.
(ROB)
Vol. 66: Sensor-BasedRobots: Algorithmsand Architectures. Edited by C. S. G. Lee. X, 285 pages.

1991 . (ROB)
Vol. 67: DesigningHypermediafor Learning. Editedby D. H.Jonassenand H. Mandl.XXV, 457 pages.

1990 . (AET)
Vol. 68: Neurocomputing. Algorithms,Architecturesand Applications. Edited by F. Fogelman Soulie

and J . Herault. XI, 455 pages. 1990.
Vol. 69: Reai-Time IntegrationMethods for MechanicalSystem Simulation. Edited by E. J. Haug and
R. C. Deyo. VIII, 352 pages. 1991.
NATO ASI Series F

Including Special Prog rammes on Sensory Systems for Robotic Control (ROB) and on
Advanced Educational Technology (AElJ
Vol. 70 : Numerical Linear Algebra, Digital Signal Processing and Parallel Algorithm s. Edited by
G. H. Golub and P. Van Dooren. XIII, 729 pages. 1991,
Vol. 71: Expert Systems and Robot ics. Edited by T. Jordanides and B.Torby . XII, 744 pages . 1991 .
Vol. 72: High-Capacity Local and Metropolitan Area Networks. Architecture and Performance Issues .
Edited by G. PUjolle. X, 536 pages . 1991,
Vol. 73: Automat ion and Systems Issues in Air Traffic Control. Edited by J. A. Wise, V. D. Hopkin and
M. L. Smith. XIX, 594 pages. 1991.
Vol. 74 : Picture Archiving and Communication Systems (PACS) in Medicine . Edited by H. K. Huang,
O. Ratib , A. R. Bakker and G. Witte . XI, 438 pages. 1991.
Vol. 75: Speech Recognition and Understanding . Recent Advances, Trends and Applications. Edited
by P. Laface and Renato De Mori. XI, 559 pages. 1991.
Vol. 76 : Multimedia Interface Design in Education . Edited by A. D. N. Edwards and S. Holland . XIV,
216 pages. 1992. (AET)
Vol. 77: Computer Algor ithms for Solving Linear Algebraic Equations. The State of the Art. Edited by
E. Spedicato. VIII, 352 pages . 1991 .
Vol. 78 : Integrating Advanced Technology into Technology Education. Edited by M. Hacker,
A. Gordon and M. de Vries. VIII, 185 pages. 1991 . (AET)
Vol. 79: Logic, Algebra, and Computation. Edited by F. L. Bauer. VII, 485 pages. 1991 .
Vol. 80 : Intelligent Tutoring Systems for Foreign Language Learning. Edited by M. L. Swartz and
M. Yazdani. IX, 347 pages . 1992. (AET)
Vol. 81 : Cognitive Tools for Learning. Edited by P. A. M. Komrners , D. H. Jonassen, and J. T. Mayes .
X, 278 pages. 1992 . (AET)
Vol. 82 : Combinatorial Optirnization . New Frontiers in Theory and Practice . Edited by M. AkgQI, H. W .
Hamacher , and S. TUfekg. XI, 334 pages. 1992.
Vol. 83 : Active Perception and Robot Vision. Edited by A. K. Sood and H. Wechsler. IX, 756 pages .
1992 .
Vol. 84 : Computer-Based Learning Environments and Problern Solving. Edited by E. De Corte, M. C.
Linn, H. Mandl, and L. Verschaffel. XVI, 488 pages. 1992. (AET)
Vol. 85 : Adaptive Learning Environments. Foundations and Frontiers. Edited by M. Jones and P. H.
Winne. VIII, 408 pages. 1992 . (AET)
Vol. 86 : Intelligent Learning Environrnents and Knowledge Acquisition in Physics. Edited by
A. Tiberghi en and H. Mandl. VIII, 285 pages. 1992. (AET)
Vol. 87: Cognitive Modell ing and Interactive Environrnents. With demo diskettes (Apple and IBM
compatible). Edited by F. L. Engel, D. G. Bouwhuis, T. Basser, and G. d'Ydewalie. IX, 311 pages .
1992. (AET)
Vol. 88: Prograrnming and Mathematical Method . Edited by M. Broy. VIII, 428 pages . 1992.
Vol. 89 : Mathernatical Problern Solving and New Information Technologies. Edited by J. P. Ponte ,
J. F. Matos , J. M. Matos, and D. Fernandes. XV, 346 pages. 1992. (AET)
Vol. 90 : Collabor ative Learning Through CornputerConferencing. Edited by A. R. Kaye. X, 260 pages .
1992. (AET)
Vol. 91 : New Direction s for Intelligent Tutoring Systerns. Edited by E. Costa . X, 296 pages. 1992 .
(AET)
NATO ASI Series F

IncludingSpecialProgrammes on SensorySystemsfor RoboticControl (ROB) and on
Advanced Educational Technology (AE7)
Vol. 92 : Hypermedia Courseware: Structures of Communication and Intelligent Help. Edited by
A. Oliveira. X, 241 page s. 1992 . (AET)
Vol. 93 : Interactive Multimedia Learning Environments. Human Factors and Techn ical Con siderati ons
on Design Issues . Edited by M . Giard ina. VIII, 254 page s. 1992 . (AET)
Vol. 94: Logic and Algebra of Specifi cation. Edited by F. L. Bauer , W . Brauer , and H. Schwichtenberg.
VII, 442 pag es. 1993 .
Vol. 95 : Comprehensive Systems Design : A New Educational Technology. Edited by C . M. Reigeluth,
B. H. Banathy, and J. R. Olson. IX, 437 pages. 1993 . (AET)
Vol. 96 : New Directions in Educat ional Technology. Edited by E. Scanlon and T. O'Shea. VIII, 251
pag es. 1992. (AET)
Vol. 97 : Advanced Models of Cogn ition for Medical Training and Pract ice. Edited by D. A. Evans and
V. L. Patel. XI, 372 pages . 1992. (AET)
Vol. 98 : Medical Images: Formation, Handling and Evaluation . Edited by A. E. Todd-Pokropek and
M . A. Viergeve r. IX, 700 pages. 1992 .
Vol. 99 : Multisensor Fusion for Computer Vision . Edited by J . K. Aggarwal. XI, 456 pages. 1993 . (ROB)
Vol. 100: Communication from an Artificial Intelligence Perspective. Theoreti cal and Appl ied Issues .
Edited by A. Ortony, J. Slack and O. Stock. XII, 260 pages. 1992.
Vol. 101 : Recent Developments in Decision Support Systems. Edited by C. W. Holsapple and A. B .
Wh inston . XI, 618 page s. 1993 .
Vol. 102 : Robots and Biological Systems : Towards a New Bion ics? Edited by P. Darto, G. Sandini and
P. Aebis che r. XII, 786 pages. 1993 .
Vol. 103 : Parallel Computing on Distr ibuted Memory Multiprocessors. Edited by F. Ozquner and
F. En;al. VIII, 332 pages . 1993 .
Vol. 104 : Instru ctional Models in Computer-Based Learning Environments. Edited by S. Dijkstra,
H. P. M . Krammer and J. J. G. van Merrle nboer. X, 510 page s. 1993 . (AET)
Vol. 105 : Design ing Environments for Constructive Learning . Edited by T. M. Duffy, J. Lowyck and
D. H. Jonassen. VIII, 374 pages . 1993 . (AET)
Vol. 106: Software for Parallel Computation. Edited by J. S. Kowalik and L. Grand inetti. IX, 363 pages.
199 3 .
Vol. 107 : Advanced Educational Technologies for Mathematics and Science . Edited by D. L.
Ferguson . XII, 749 pages . 1993 . (AET)
Vol. 108: Concurrent Engineering : Tools and Technologies for Mechanical System Design . Edited by
E. J. Haug. XIII, 998 pages. 1993.
Vol. 109 : Advanced Educational Technology in Technology Education . Edited by A. Gordon ,
M. Hacker and M . de Vries. VIII, 253 pages . 1993 . (AET)
Vol. 110 : Verification and Validation of Complex Systems: Human Factors Issues. Edited by J. A.
Wise, V. D. Hopkin and P. Stager. XIII, 704 pages . 1993 .
Vol. 111 : Cognitive Model s and Intelligent Environments for Learning Programming. Edited by
E. Lemut, B . du Boulay and G. Dettori. VIII, 305 pages . 1993. (AET)
Vol. 112 : Item Banking: Interactive Testing and Self-Assessment. Edited by D. A. Leclercq and J. E.
Bruno. VIII, 261 pages. 1993 . (AET)
Vol. 113: Interactive Learning Techn ology for the Deaf. Edited by B . A. G. Elsendoorn and F. Conin x.
XIII, 285 pages . 1993. (AET)
NATO ASI Series F

Including Special Programmes on Sensory Systems for Robotic Control (ROB) and on
Vol. 114: Intelligent Systems : Safety, Reliability and Maintainability Issues. Edited by O. Kaynak,
G. Honderd and E. Grant. XI, 340 pages. 1993 .
Vol. 115 : Learning Electric ity and Electronics with Advanced Educational Technology. Edited by
M. Caillot. VII, 329 pages . 1993. (AE7)
Vol. 116: Control Technology in Elementary Education. Edited byB. Denis. IX,311 pages. 1993 . (AE7)
Vol. 117 : Intelligent Learning Environments : The Case of Geometry . Edited by J.-M . Laborde. VIII, 267
pages . 1996. (AE7)
Vol. 118: Program Design Calculi. Edited by M. Broy. VIII, 409 pages. 1993.
Vol. 119 : Automating Instructional Design, Development , and Delivery. Edited by. R. D. Tennyson .
VIII, 266 pages. 1994 . (AE7)
Vol. 120 : Reliability and Safety Assessment of Dynamic Process Systems . Edited by T. Aldemir ,
N. O. Siu, A. Mosleh , P. C. Cacciabue and B. G. Goktepe. X, 242 pages . 1994 .
Vol. 121: Learning from Computers: Mathematics Education and Technology. Edited by C. Keitel and
K. Ruthven. XIII, 332 pages . 1993. (AE7)
Vol. 122: Simulation-Based Experiential Learning. Edited by D. M. Towne, T. de Jong and H. Spada .
XIV, 274 pages. 1993. (AE7)
Vol. 123 : User-Centred Requirements for Software Engineering Environments. Edited by D. J.
Gilmore, R. L. Winder and F. Detienne. VII, 377 pages . 1994 .
Vol. 124: Fundamentals in Handwriting Recognition . Edited by S. Impedovo. IX, 496 pages . 1994 .
Vol. 125: Student Modelling : The Key to Individualized Knowledge-Based Instruction. Edited by J. E.
Greer and G. I. McCalla . X, 383 pages. 1994. (AE7)
Vol. 126 : Shape in Picture. Mathematical Description of Shape in Grey-level Images. Edited by
Y.-L. 0 , A. Toet , D. Foster, H. J. A. M. Heijmans and P. Meer. XI, 676 pages . 1994.
Vol. 127: Real Time Computing. Edited by W. A. Halang and A. D. Stoyenko. XXII, 762 pages . 1994 .
Vol. 128: Computer Supported Collaborative Learning. Edited by C. O'Malley. X, 303 pages . 1994 .
(AE7)
Vol. 129: Human-Machine Comm unication for Educational Systems Design. Edited by M. D.
Brouwer-Janse and T. L. Harrington . X, 342 pages . 1994 . (AE7)
Vol. 130 : Advances in Object-Oriented Database Systems . Edited by A. Dogac , M. T. Ozsu, A. Biliris
and T. Sellis. XI, 515 pages . 1994 .
Vol. 131: Constraint Programming . Edited by B. Mayoh, E. Tyugu and J. Penjam. VII, 452 pages.
1994 .
Vol. 132: Mathematical Modelling Courses for Engineering Education . Edited by Y. Ersoy and A. O.
Moscardini. X, 246 pages . 1994 . (AE7)
Vol. 133 : Collaborative Dialogue Technologies in Distance Learning. Edited by M. F. Verdejo and
S. A. Cerri. XIV, 296 pages. 1994 . (AE7)
Vol. 134: Computer Integrated Production Systems and Organizations . The Human-Centred
Approach. Edited by F. Schmid, S. Evans, A. W. S. Ainger and R. J. Grieve. X, 347 pages . 1994 .
Vol. 135 : Technology Education in School and Industry . Emerging Didactics for Human Resource
Development. Edited by D. Blandow and M. J. Dyrenfurth. XI, 367 pages . 1994. (AE7)
Vol. 136 : From Statistics to Neural Networks. Theory and Pattem Recognition Applications. Edited
by V. Cherkassky, J. H. Friedman and H. Wechsler. XII, 394 pages . 1994 .
NATO ASI Series F

IncludingSpecialProgrammes on SensorySystemsfor Robotic Control (ROB) and on
Vol. 137 : Technology-Based Learning Environments. Psychological and Educat ional Foundations.
Edited by S. Vosniadou , E. De Corte and H. Mandl. X, 302 pages . 1994. (AET)
Vol. 138 : Exploiting Mental Imagery with Computers in Mathematics Education. Edited by
R. Sutherland and J. Mason . VIII, 326 pages . 1995. (AET)
Vol. 139 : Proof and Computation. Edited by H. Schwichtenberg. VII, 470 pages . 1995.
Vol. 140: Automating Instructional Design: Computer-Based Development and DeliveryTools. Edited
by R. D. Tennyson and A. E. Barron . IX, 618 pages . 1995 . (AET)
Vol. 141 : Organizat ional Learning and Technological Change. Edited by C. Zucchermaglio, S.
Bagnara and S. U. Stucky . X, 368 pages . 1995 . (AET)
Vol. 142: Dialogue and Instruction . Modeling Interaction in Intelligent Tutoring Systems . Edited by
R.-J. Beun, M. Baker and M. Reiner. IX, 368 pages . 1995. (AET)
Vol. 143: Batch Processing Systems Engineering. Fundamenta ls of Chemical Engineering. Edited by
G. V. Reklaitis, A. K. Sunol, D. W. T. Rippin, and O. Hortacsu , XIV, 868 pages . 1996 .
Vol. 144 : The Biology and Technology of Intelligent Autonomous Agents. Edited by Luc Steels. VIII,
517 pages. 1995 .
Vol. 145: Advanced Educational Technology : Research Issues and Future Potent ial. Edited by T. T.
Liao. VIII, 219 pages . 1996 . (AET)
Vol. 146: Computers and Exploratory Learning. Edited by A. A. diSessa, C. Hoyles and R. Noss .
VIII, 482 pages. 1995. (AET)
Vol. 147: Speech Recognition and Coding. New Advances and Trend s. Edited by A. J. Rubio Ayuso
and J. M. Lopez Soler. XI, 505 pages . 1995.
Vol. 148: Knowledge Acquisition, Organization, and Use in Biology. Edited by K. M. Fisher and M. R.
Kibby. X, 246 pages. 1996. (AET)
Vol. 149 : Emergent Computing Methods in Engineering Design. Applications of Genetic Algorithms
and Neural Networks. Edited by D.E. Grierson and P. Hajela. VIII, 350 pages. 1996 .
Vol. 150 : Speechreading by Humans and Machines. Edited by D. G. Stork and M. E. Hennecke. XV,
686 pages . 1996.
Vol. 151 : Computational and Conversational Discourse . Burning Issues - An Interdisciplinary

Account. Edited by E. H. Hovy and D. R. Scott. XII, 202 pages. 1996 .
Vol. 152 : Deduct ive Program Design. Edited by M. Broy . IX, 467 pages . 1996 .
Vol. 153: Identification, Adaptation, Learning. Edited by S. Bittanti and G. Plcci. XIV, 553 pages. 1996 .
Vol. 154 : Reliability and Maintenance of Complex Systems . Edited by S. Ozekici, XI, 589 pages. 1996 .
Vol. 155 : Cooperation: Game-Theoretic Approaches. Edited by S. Hart and A. Mas-Colell. VIII, 328
pages . 1997 .
Vol. 156 : Microcomputer-Based Labs: Educational Research and Standards. Edited by R.F. Tinker .
XIV, 398 pages . 1996. (AET)
Vol. 157 : Logic of Computation. Edited by H. Schwichtenberg . VII, 396 pages . 1997 .
Vol. 158 : Mathematical Methods in Program Development. Edited by M. Broy and B. Schieder . VIII,
528 pages . 1997 .
Vol. 160 : Discourse, Tools, and Reasoning: Essays on Situated Cognition . Edited by L.B . Resnick,
R. Salj6, C. Pontecorvo and B. Bunge. XII, 474 pages . 1997 (AET)
Springer
and the
environ ent
At ringer we firmly believe that an
intern ational science publisher has a
special obligation to the environme nt,
and our corp orate policies consistently
reflect this conviction.
We also expect our business partners paper mills, printers, packaging
manufacturers, etc. - to commit
themselves to using materials and
production processes that do not harm
the environment. The paper in this
book is made from low- or no-chlorine
pulp and is acid free, in conformance
with international standards for paper
permanency.
Springer

Pontecorvo Clotilde

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pontecorvo Clotilde

Uploaded by

Copyright:

Available Formats

NATO ASI Series

Advanced Science Institutes Series

Plenum Publishing Corporation

C Mathematical and Physical Sciences

Kluwer Academic Publishers

Computer and Systems Sciences

Partnership Sub -Series

The Partnership Sub-Series incorporates activities undertaken in collaboration with

Series F: Computer and Systems Sciences, Vol. 160

The NATO ASI Series F Special Programme on

Springer-Verlag Berlin Heidelberg GmbH

Proceedings of the NATO Advanced Research Workshop on Discourse,

LIbrary of Congress CatalogIng-In-PublIcatIon Data

Discourse. too ls. and reason ing : essays on situated cognition /

CR Subject Classification (1991): K.3, K.4

Typesett ing: Camera-ready by editors

List of Contributors........... ............ ...................................................

Chapter 2 Centers of Coordination: A Case and Some Themes

Chapter 3 Animated Texts: Selective Renditions of News Stories

Chapter 4 To Resolve a Technical Problem Through Conversation

Chapter 5 The Blackness of Black: Color Categories as Situated Practice

Chapter 14 Strategy-Specific Information Access in Knowledge

Accountable Talk: Learning to Reason

Chapter 15 Talking About Reasoning: How Important Is the Peer

Chapter 17 Learning to Argue in Family-Shared Discourse:

Chapter 18 Discourse in the Adult Classroom :

Kerstin Bergqvist, Department of Communication Studies, Linkoping University,

Elinor Ochs, Department of Applied Linguistics, UCLA, CA, USA

Discourse, Tools, and Reasoning:

In 1990, we were asked by the Scientific Affairs Division of NATO I to organize a

'Our invitation came from the Panel on Advanced Educational Technology.

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

Discourse, Tools, and Reasoning

conceive of thought as situated activity. Discourse (and, therefore, all of

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

profoundly sociocultural in the Vygotskian spirit (Wertsch, 1991a, 1991b) . With

Discourse, Tools, and Reasoning

socioindividual account, if you wilt-could then be constructed by compiling

Lauren B. Resnick, Clotilde Pontecorvo , and Roger Saljo

Part 1. Distributed Cognition: Tools, Discourse,

Discourse, Tools, and Reasoning

The chapters in this section generally replace the concept of coordinated

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Sliljo

laboratory. There an apprentice chemist is learning to prepare a fiber to be used as

Discourse, Tools, and Reasoning

challenge the basic interpretation . A related problem is one of sampling . In each

Part 2. Negotiating Identities: The Construction

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

A community of practice is not just a collection of individuals who happen to be

Discourse, Tools, and Reasoning

surroundings. In syncretic literacy, as Duranti and Ochs (chapter 7, this volume)

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

Part 3. Learning in Practice: How People and Tools

Discourse, Tools, and Reasoning

newcomers to a particular environment both adapt to and modify the environment.

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

Discourse, Tools, and Reasoning

Part 4. Accountable Talk: Learning to Reason

Lauren B. Resnick, Clotilde Pontecorvo, and Roger Saljo

a conclusion by challenging the veracity of a premise, even an unstated premise

Discourse, Tools, and Reasoning

practitioner. Perception itself--the "blackness ofblack"--is shaped by the discourse

Lauren B. Resnick, Clotilde Pontecorvo , and Roger Saljo

Discourse, Tools, and Reasoning