Bertrand Russell and The Origins of The Set-Theoretic 'Paradoxes'. Alejandro R. Garciadiego

Bertrand Russell In 1907
Alejandro R. Garciadiego
Bertrand Russell
and the Origins of the Set-theoretic
'Paradoxes'
1992
Birkhauser Verlag
Basel Boston Berlin
Author's address
Dr. A.R. Garcladlego
Umversldad Naclonal Aut6noma de MexIco
Cubiculo 016
Departamento de Matematlcas
Facultad de Clenclas
Cludad Umversltarla
04510 Mexico, D.F. (MexIco)
Library of Congress Cataloging in Publication Data

Garciadiego Dantan, Alejandro Ricardo.
Bertrand Russell and the origins ofthe set-theoretic 'paradoxes' / Alejandro R. Garcladlego.
p.cm.
Includes bibliographical references and index.
ISBN-13 978-3-0348-7404-5
e-ISBN-13 978-3-0348-7402-1
001 10 1007/978-3-0348-7402-1
1. Logic, Symbolic and mathematical. 2. Set theory. 3. Paradox. 4. Russell, Bertrand,

1872-1970 Principles of mathematiCs.
I. Title.
QA9.G24 1992
511.3 - dc20
Deutsche Bibliothek Cataloging-inPublication Data
Garciadiego, Alejandro R.:
Bertrand Russell and the origins of the set theoretic 'paradoxes' / Alejandro R. Garcladlego. Basel; Boston; Berlin: Birkhauser, 1992
ISBN-13 978-3-0348-7404-5
This work IS subject to cOPYright. All rights are reserved, whether the whole or part ofthe material
IS concerned, speCifically those of translation, reprinting, re-use of Illustrations, broadcasting,
reproduction by photocopying machine or similar means, and storage In data banks. Under 54
of the German COPYright Law, where copies are made for other than private use a fee is payable to
Verwertungsgesellschaft Wort, Munich.
1992 Birkhauser Verlag Basel
Softcover reprint of the hardcover 1st edition 1992
Cover deSign: Albert Gomm

ISBN-13 978-3-0348-7404-5
I've have put in so many

enigmas and puzzles that
it will keep the professors
busy for centuries arguing
over what I meant, and that's
the only way of ensuring one's
immortality.
James Joyce
TO LUPITA, GABRIELA & ALEJANDRO
Contents
Prologue by Ivor Grattan-Guinness
ix
Preface
1. The goal of the book
2. The sources
xi
xx
Acknowledgements
xxvii
Chapter I. Antecedents
1.1 Introduction
1.2 Cantor's transfinite numbers
1. 3 Some reactions
1
2
13
Chapter II. A standard interpretation

2.1 Introduction
2.2 The Burali-Forti 'paradox'
2.3 Cantor's 'paradox'
2.4 Epilog
19
21
32
40
Chapter III. The philosophical and mathematical background to The Principles

of Mathematics, 1872-1900
3.1 Introduction
3.2 Russell's childhood and adolescence
3.3 Russell as a student at Cambridge
3.4 Russell's An Essay on the Foundations of Geometry
3.5 Russell's early attempts to write a book on the
principles of arithmetic
41
43
49
56
61
viii
Contents
Chapter IV. Russell's discovery of the 'paradoxes'

4.1 Introduction
4.2 The First International Congress of Philosophy, Paris 1900
4.3 Russell's writing of the 'first final' draft of The Principles of
Mathematics, November - December 1900
4.4 The 'paradoxes'
4.5 A last effort to finish The Principles of Mathematics, April 1902 May 1903
Chapter V. The 'semantic paradoxes'
5.1 Introduction
5.2 Berry's argument
5.3 The Konig - Zermelo proposition
5.4 Richard's contradiction
5.5 Polemics within the London Mathematical Society over the
Well-Ordering Theorem
5.6 Konig's and Dixon's Arguments
81
83
87
100
117
131
134
137
140
143
146
Conclusions
151
Appendix. Correspondence
6.1 Alys Russell
6.2 G. E. Moore
6.3 David Hilbert
6.4 Cesare Burali-Forti
6.5 G. G. Berry
6.6 Alfred N. Whitehead
6.7 G. H. Hardy
6.8 E. H. Moore
155
160
162
164
166
185
188
205
List of tables
207
List of illustrations
209
Bibliography
211
Index
257
Prologue
The history of the paradoxes of set theory is well known. Cantor

found the one concerning the greatest cardinal in the 1890s, and soon
afterwards Burali-Forti discovered the corresponding for the ordinals.
Then around 1900 Russell showed that the set of all sets which do not
belong to themselves led to a paradox. This finding caused a crisis, and
mathematicians, philosophers and logicians struggled to free mathematics
froql these infections. However, other paradoxes came to light, such as
those of finite definability and of the visiting card.
Like all well-known stories, this one is better known than studied.
While some historians have revealed a more complicated story, Dr.
Garciadiego has gone to new levels of detail to reveal the complications
of the true record. They involve not only more information of the same
kind but especially many corrections of our understanding. My own
favourites are these: neither Cantor nor Burali-Forti thought of their
results as paradoxes; Russell was their principal founder and first librarian, in that he both collected them and interpreted them as paradoxes; the
status of such result as paradoxes was entwined in a melange of poorly
distinguished paradoxes, contradictions, antinomies, reductio ad absurdum arguments, and strange objects like the axioms of choice which one
had to accept or at least tolerate; hardly any major figure was drawn to
set theory by the paradoxes; and the community of mathematicians in
general gave the matter little thought.
Dr. Garciadiego covers the period from the mid 1890s to the late
1900s. He begins his tale with the final form of Cantor's set theory and
transfinite numbers, and he steers it through Russell's own discoveries
Prologue
and reinterpretations to the emergence of the 'semantic' paradoxes (as

they were normally not distinguished at the time). He adjoins a substantial quantity of unpublished correspondence, mostly with Russell, in
which the perplexities and some disagreements of the time are vividly
exhibited. His book is unusually readable for an academic work and
deserves to gather a substantial measure of attention-more than arose at
the time, indeed, when the full consequences of the paradoxes were not
yet clear.
Dr. Ivor Grattan-Guinness,
Middlesex Polytechnic
The love of power is terribly strong

in me-l can't help reflecting that all these
math'al philosophers have different thoughts from
what they wd. have had if I had not existed.
Bertrand Russell.
Preface
1. The goal of the book.- The main goal of this book is to
reconstruct and reinterpret the role played by the British philosopher and
mathematician Bertrand Russell (1872-1970) in the origins of the set-theoretic 'paradoxes'. 1
My book is addressed to all students interested in the history and
philosophy of ideas, not only to those who specialize in mathematics. The
professional historian, philosopher and mathematician will find a new
historical interpretation of the events surrounding the development of the
paradoxes, in turn provoking new avenues of discussion.
1. The term paradox might indicate an argument that derives self-contradictory conclusions by apparently valid deductions from apparently acceptable premises. A paradox could also be "a statement
contrary to receive opinion" [The American Heritage Dictionary 1982a, 9(0). Paradoxes might involve
single inconsistencies (e.g., showing that two different numbers are equal) or double inconsistencies
(e.g., the barber: In a village there is a barber who "shaves all and only those men in the village who
do not shave themselves" [Quine 1966a, 2). Now, does the barber shave himself?) But paradoxes, in
any case, usually invite solution. The word contradiction, on the other hand, means "a statement
containing propositions one of which denies or is logically at variance with the other" [The Oxford
English Dictionary 1989a III, 838). However, even if one could put forward definitions satisfying the
criteria of contemporary mathematicians and philosophers, these criteria would not necessarily correspond to those current at the beginning of the century. The meaning and connotation of the words
paradox, contradiction and allfinomy (see: page 62 note 1) seem to have changed in the interim. Most
of the intellectuals involved in the discovery of such arguments (in the 18908 and 1900s) selected and
consistently used one (or some) of these terms when referring to the inconsistencies. For example,
Russell used all of the terms given above, but with different conceptions in mind. In fact, a close
analysis of the way in which Russell used these terms clarifies the diverse philosophical influences that
he underwent while he was writing his book The Prillciples of Mathematics. In order to remain
impartial, and at the same time to keep a historical perspective, I will use neutral terms (e.g., inconsistency, argument, proposition, 'contradiction' and 'paradox').
Preface
Xll
Russell's published works include more than sixty books, several

unpublished manuscripts, many hundreds of articles, dozens of radio and
TV interviews and films, covering a wide spectrum of knowledge. His
writings embrace discussions and analysis of such diverse topics as social
sciences, foundations of mathematics, philosophy of physics, philosophy
in general, religion, moral sciences, education, pacifism, natural sciences
(including biology and physics), linguistics, statistics, probability, economic theory, history, politics, international affairs and other topics. He
corresponded with a large and diverse group of colleagues including both
prominent and obscure figures in politics, the arts, humanities and sciences. Russell's communication with his colleagues began in the late nineteenth century and was especially active through much of the twentieth
century. In spite of being one of the most controversial public personalities of his day (let us not forget that he went to prison twice, was dismissed from Cambridge University and was prevented from teaching at
the College of the City of New York), his merits have been recognized
and appreciated. He was awarded many medals, diplomas and honors,
including the Nobel Prize for Literature in 1950. On considering the vast
corpus of his work, the people who surrounded him, the events that he
experienced, and the polemics and disputes that he provoked and in
which he participated, it seems almost an unforgivable sin to devote a
whole book to such a small idea, as the discovery of some inconsistencies. However, Russell himself conceived this idea as one of the most
fundamental of his work. Now, according to Russell:
Three passions, simple but overwhelmingly strong, have governed my life: the longing for love, the search for knowledge, and unbearable pity for the suffering of mankind. These passions, like great winds, have blown me hither and thither, in a wayward course, over a deep ocean of anguish, reaching to the very verge of despair.
These three passions that governed his life-which extended for

almost ninety-eight years-were carved, molded and deeply rooted in his
mind by approximately 1901 and 1902, when he lost his love for his first
wife, witnessed the pains suffered by a close friend and discovered the
1. Russell 1967a, 13.
Russell & the 'paradoxes'
Xlll
StnkA~~~
4 ~\( f~ A.AA-~!.
,.1. ~ ttdL~ ~
daM ( +~ ~
(l;...
bl, .....t- ,.~
i6 ~~-I-f.
~~ k.~Jj,aA1.
"l,..,. ~ 'M\- ~ ~L
f~~ " ~ ..,. ci
i ~. .-'aA( ( ~ I- i6
ta..........w..
h.4'i- A.. ~.6c....
1 =..~ : "'~~ ~ ~
il-".
~~~
P"....~ ~ fII~c~ ~ .. /o~
Plate 2. Russell drafted this version of his 'contradiction' at a

mature age, showing that he still regarded his argument as
important.
xiv
Preface
inconsistencies of the theory of transfinite cardinal and ordinal numbers. I

Intellectually speaking, it was the discovery of an argument he called The
Contradiction that forced him to stress the importance of the existence of
these elements and their solution for a correct understanding of the principles of mathematics. This proposition is known today as the argument
involving the class (set) of all classes (sets) which are not members of
themselves, and it might be popularized as follows: think for a moment
of a special library where the librarians do not catalogue the books in
cards but in notebooks; i. e., the catalogue itself could be taken as another
book. Some librarians list the book ~s an element of the system and
catalogue the catalogue-but some others do not. What would happen if
the chief librarian in charge of the central library decides to compile a
catalogue of all catalogues that do not catalogue themselves? Should this
general catalogue be included in the list (or catalogue of catalogues) or
not? If it is listed, then it is not one of those catalogues that should be
contained in the catalogue; but, if it is not, then it should be. That is to
say, from each answer its opposite follows.
Near the end of his life, Russell believed he had fulfilled some but
not all of his early academic goals. One of the most relevant for the
study of philosophy and the foundations of mathematics is the one the
young British thinker devised in the 1890s as two series of books-one
going from pure mathematics to physiology and the other dealing with
social questions, hoping to find a final synthesis at once scientific and
practical. To some extent, as he claims, he carried out his plan. But, in
retrospect, in the early 1900s, mathematics was in a doubtful state because of the inconsistencies, therefore, it would be impossible to show
that other disciplines were logically derived from pure mathematics and
were internally consistent. For Russell, it was like carrying a small stone
in one of his shoes which bothered him constantly while he has walking
through the path of life and knowledge. I believe that it is not an exaggeration to assert that Russell's discovery of the inconsistencies-in particular the one associated with the class of all classes which are not members of themselves-was the most influential event for his later contributions to philosophy and the foundations of mathematics.
1. Grattan-Guinness 1977a, 159-160.
xv
When asked about his greatest contribution to philosophical

thought, Russell responded by mentioning Principia Mathematica. The
main thesis of this book-now known as the Logicist thesis-was to show
that pure mathematics was deducible from logic. 1 This thesis was developed in the same year of 1901 (approximately) and its origins are narrowly related with those of the inconsistencies. The discovery of such
inconsistencies, intrinsically associated with the principles of traditional
logic and/or set theory, showed Russell that his goals were not as direct
and easy to fulfill as he had originally thought.
In order to carry out the main task of this book, I have divided it
into five chronological and sometimes overlapping sections. The goals
and contents of these sections are the following:
Chapter 1. Antecedents.- Some sources in the secondary literature
may have led the reader to believe that there was a sudden interest in the
study of the foundations of mathematics at the turn of the century resulting from the discovery of the set-theoretic 'paradoxes'. I would like to
present an alternative historical reconstruction by describing Georg Cantor's (1845-1918) theory of transfinite numbers as an example of an
attempt to clarify the concept of number. On one hand, according to most
historians, philosophers and mathematicians, this theory contains all
elements needed to develop the inconsistencies that provoked a sudden
interest on the foundations of mathematics at the turn of the century. At
the same time, it provides the framework to understand how the 'paradoxes' emerged and how they influenced the subsequent development of the
foundations of mathematics. On the other hand, Cantor's theory is a clear
example of a research project concerned with showing that the foundations of arithmetic have their original roots in a precise mathematical
problem.
1. According to Sacks, Russell had other motives for writing this book. "There were difficulties, but
not paradoxes; there was no need to write Principia Mathematica to straighten them out" [Sacks 1975a,
523]. One should not forget that, In 1903, Russell promised his readers to present a second volume of
The Principles of Mathematics containing the technical work of how the deduction of all of pure
mathematics from symbolic or formal logic was possible. This second volume, with Alfred N. Whitehead's (1861-1947) collaboration, turned out to be Principia Mathematica.
XVI
Preface
Other scholars who were interested in constructing a more solid

mathematics, for different reasons, were completely unaware of the
existence of some of the inconsistencies. Gottlob Frege (1848-1925) and
Giuseppe Peano (1858-1932), among others, also played a fundamental
role in the clarification of the concept of number in the last few decades
of the nineteenth century and they had originally started their works
unaware of the existence of the 'paradoxes'.
In the end, Russell modified three different conceptualizations
underlying the principles of the natural numbers
{I, 2, 3, ... , n, ... }

to create a new understanding that emphasized the inconsistencies. This
notion resulted from a final breakdown of a general synthesis in which,
after he constrasted various apparently contradictory dialectical analyses,
a new comprehensive thesis originated. This introductory chapter will
provide the technical knowledge required to understand how Russell's
ideas evolved and how he discovered the inconsistencies.
Chapter II. A Standard Interpretation.- It is of fundamental importance to describe what I call a standard interpretation of the origins of
the set-theoretic 'paradoxes' in order to fully comprehend the historiographical consequences of the present work, and to analyze how it affects
our understanding of the history of mathematics. Some of the main elements of the standard interpretation are: (a) that Cesare Burali-Forti
(1861-1931) discovered an inconsistency in 1897, as a result of his criticism of Cantor's ideas; (b) that Cantor came upon another 'paradox' in
1899; (c) that Russell discovered his own Contradiction-Russell's
term-in 1901 after analyzing the previous two.
Of course, there are several discrepancies in the accounts put forward by several historians, but none of them has provoked a major historiographical reexamination of the basic history. For example, some researchers have suggested that there was an immediate response of the
mathematical community because of the need for a close scrutiny of the
foundations of set theory. Some others, on the contrary, have attempted
Russell & the paradoxes ,
xvii
to explain the lack of such a reaction. Some have dated Cantor's discovery as early as 1883, 1895, 1896, and others as late as 1899. I
There is evidence that Burali-Forti's and Cantor's original contributions have been misinterpreted, both mathematically and historically. I
do not agree that there was an immediate response of the mathematical
community, as implied by the standard interpretation. I also refute the
alternative account suggesting that there was no reaction at all. The reader will not find, in this chapter, a historical reconstruction of events such
as the narration of the discovery of the inconsistency of the greatest
cardinal number-incorrecty called 'Cantor's paradox', and the argument
of the class of all classes which are not members of themselves (Russell's
contradiction). These issues will be discussed in chapter IV. The discovery of the 'paradox' of the greatest ordinal number (incorrectly attributed
to Burali-Forti) has already been correctly described in another study.2
Chapter III. The philosophical and mathematical background to The
Principles of Mathematics, 1872-1900.- It is difficult to establish either a
precise date or a reason to explain Russell's original interest in the foundations of mathematics. I will first endeavor to outline the most significant events of Russell's childhood and adolescence before he went to
Cambridge University in 1890, with respect to the development of this
book.
Then I will narrate how, in his fourth year at Cambridge, after
having dedicated three years to the study of mathematics, he studied the
works of Immanuel Kant (1724-1804), Francis H. Bradley (1846-1924)
and Bernard Bosanquet (1848-1923), among other philosophers. It is
important to stress that at this time (ca. 1893-1898), Russell thought of
all sciences as being inherently defective, therefore, mathematics was not
expected to be a fully consistent science. The 'contradictions' found in
mathematics would be solved in a subsequent science following a dialectical process. In the last few years of the nineteenth century, Russell studied Cantor's theory of transfinite numbers and made an in-depth study of
1. Complete references will be provided in the appropiate places (see: Chapter II, 3, page 33).
2. See: Moore & Garciadiego 1981a.
xviii
Preface
the philosophy of Gottfried Wilhelm Leibniz (1646-1716). At this time,

approximately 1898-1899, he turned away from Georg Wilhelm Hegel
(1770-1831) and developed intellectual associatidIli with Whitehead and
especially with G. E. Moore (1873-1958), among others. I try to show
that all these developments are related and I will show that Russell turned
away from the philosophical antinomies to logical or mathematical 'contradictions' after studying Cantor's Mengenlehre. In short, it will be
explained how Russell came to believe that mathematics was a consistent
science implying that previous philosophical antinomies related to mathematics were simply apparent 'paradoxes' and not real 'contradictions'.
Finally, in this chapter, I will attempt to determine when the early
drafts of The Principles of Mathematics were written in order to illustrate
the intellectual background existing before Russell met Peano in 1900.
An important feature of the last draft, written before July 1900, entitled
simply "Principles of Mathematics", is the absence of any discussion of
real 'contradietions'.
Chapter IV. Russell's discovery of the 'paradoxes'.- Once it has
been expounded how and why Russell studied inconsistencies related to
philosophical systems, it will then be possible to explore his discovery of
the mathematical (or logical) 'contradictions'. The best way to carry out
such a task will be a detailed analysis of the manner in which he wrote
The Principles of Mathematics in the form that it was published. This
will necessitate the study of the way Peano's influence modified Russell's
understanding of the relations between logic and pure mathematics, which
Russell used to explain the principles of arithmetic. It is important to
emphasize, once again, that Russell's book focussed primarily on the
proposition that pure mathematics was a branch of logic. It was less
concerned with the discovery or solution of the 'paradoxes'.
In the last decade of the nineteenth century, Russell started corre-
sponding with the French philosopher Louis Couturat (1868-1914).

Through their correspondence they discussed a great variety of topics
such as Cantor's theory of transfinite numbers, Burali-Forti's papers and
their respective books on Leibniz's logic. I believe that the disclosure and
explanation of some of the issues discussed in their correspondence illus-
XIX
trates my argument of ho~ Russell created the 'paradox' of the greatest

cardinal number and his own argument of the class of all classes which
are not members of themselves. Fundamental changes related to the
inconsistencies were apparently made in galley as the book was being
printed, and I will examine these changes.
Chapter V. The 'Semantic Paradoxes'. - It became a fundamental
task for Russell-in his effort to show that pure mathematics was a
branch of logic-to discover all possible inconsistencies in order to study
and solve them. There were others who were also interested in finding
'paradoxes'. Some were probably dissatisfied with Russell's thesis and
wanted to discredit it. Today, these 'paradoxes' might be divided in two
categories: logical (those involving only logical or mathematical terms)
and semantic 'paradoxes' (those containing some reference to thought,
language or symbolism). 1 The goal of this fifth chapter is to discuss
why, how and when the semantic 'paradoxes' were discovered. Here, it
is important to consider the role played by the Well-Ordering Theorem, 2
and its associated polemics in popularizing the logical 'paradoxes' and the
discovery of the semantic ones. I will illustrate that there was not one
simple factor that dominated the whole outlook. I will show that some of
the semantic 'paradoxes' arose from misunderstandings or misconceptions. Others arose from the logical inconsistencies, and some arose
independently of any polemic.
Appendix 1. Correspondence.- In this appendix I have included
some of Russell's letters that" reveal different aspects of his personality
and thought. Although hundreds of Russell's letters have been published
already [see: Russell 1967a; 1968a, 1969a, 1969b, Kennedy 1975a,
Wright 1974a, & Grattan-Guinness 1977a, among others], there remain
thousands more (at least at The Bertrand Russell Archives) that show the
diversity and richness of Russell's personality and knowledge. This appendix contains some (in fact, only a very few) items of Russell's corre-
1. Ramsey I926a, 353.
2. The assertion that any set, ordered or not, can be well-ordered (see: Chapter I, 2, P 11-13).
xx
Preface
spondence. 1 Some are included because they reflect Russell's emotional

health at the time of the writing of The Principles of Mathematics. Others
exemplify the growth and development of his ideas. But all of them
provide insight into the events surrounding the development of set theory
and mathematical logic at the beginning of the present century. Finally, I
have limited myself to transcribing the correspondence. The richness of
the letters is so vast that it would be necessary to write another essay
explaining names, dates, topics, symbology, and so on. I have included
very sporadic footnotes to clarify some very basic issues.
2. The sources.-2 Throughout his life, Russell made several attempts to write his autobiography, beginning as early as 1901 when he
was only twenty nine years 01d. 3 He made at least one other attempt in
1948 at age seventy six. 4 Nevertheless, the first volume of Russell's
autobiography was published almost twenty years later in March 1967,
when he was almost ninety-five years old.
Although comments of friends and relatives describe Russell as
being extremely fit, both mentally and physically even near the end of his
life, it is reasonable to suppose that there might be numerous, potentially
important flaws in his reminiscences, especially for events occurring
around the years 1900-1901, over fifty years before the first recollections
were published. Russell may have consulted records of events occurring
1. I also transcribed a letter from E. H. Moore (1862-1932) to Georg Cantor. The pertinence of the
inclusion is justified by the topic discussed. This missive is a very clear example that 'paradoxes' were
in the mind of the working mathematician of those days. The letter was written before Russell's The
Principles of Mathematics were published-although, unfortunately, the surviving records from Cantor
seem to indicate that he did not respond to Moore. This is another clear example of someone thinking
in the same lines as Russell, as well as Ernst Zermelo (1871-1953), did. I must express my deepest
gratitude to Ivor Grattan-Guinness for sharing his knowledge of this particular item. The original
version of this letter is kept at the Institut Mittag-Leffler (Stockholm, Sweden).
2. Professional historians may skip this section.

3. Unfortunately, the manuscript in Alys Russell's (nee Alys Pearsall Smith (1868-1949 handwriting
is now lost.
4. Some sources suggest that Russell made probably a previous attempt in about 1931. On the other
hand, perhaps, Russell may have started writing his 1948 attempt in 1931.
xxi
in 1901, assuming that they were available.! The draft of 1948 was supplemented and revised constantly prior to its publication in the 196Os.
Some passages, written almost twenty years previously, began to show
the effects of time. Russell himself mentions an example:
I wrote the above passage in 1931 [of an event of 1919], and in 1949 I showed it to
Colette [Lady Constance Malleson (1895-1975), Russell's mistress]. Colette wrote to
me, enclosing two letters that I had written to her in 1919, which showed me how
much I had forgotten. 2
The disadvantages of relying on the use of autobiographical sources

are too well known to be discussed here in detail, but it is important to
keep some of them in mind. In most cases, the flaws found in an autobiographical source are closely related to the precise dating of events of
actual matters, but some are also connected to the development of ideas.
As an example of the first case one finds Russell's small discrepancies in
the dating of the discovery of his inconsistency. As another example of
how his memory of previous experiences betrayed him on some occasions, let me refer to Russell's account of how he wrote a book. After
describing his mature methodology, Russell used his book Our knowledge
of the External World as a Field for Scientific Method in Philosophy
[1914a] to illustrate this procedure. He narrated how he spent a whole
year thinking about the problem. He was always finding errors, objections and difficulties on all his possible solutions. Then, in one day, he
dictated almost the entire book, to a stenographer without a moment's
hesitation. However, his memory betrayed him on this occasion. This is
not the book he wrote in barely a single day. He probably confused the
writing of his book with the writing of an article called "The relation of
sense data to physics."3
1. As far as I know, there are no clues indicating whether Russell ever finished or even took seriously
that early attempt at writing his autobiography. If, in fact, he did, which is very doubtful, there are no
hints he ever used it when he wrote later recollections. Nevertheless, there are other early autobiographical sources (e.g . Russell 1983a, 3-21 & 41-67, Russell 1902g), but they contain few references
to his technical work, most of them being observations of a very personal character (e.g., he describes
his feelings towards his first wife Alys after he discovered that he did not love her any more; he also
characterizes his friends, personal reactions to his education, and so on).
2. Russell 1968a, 96. See, also: Ibid., 151.

3. See Russell 1956a, 195-196. Cj., Blackwell 1973a, 13.
xxii
Preface
Allan Wood, one of Russell's best known biographers, provided an

example of the difficulties encountered when recalling the development of
ideas in an autobiography:
I am sometimes tempted to think that there is no single subject on which, from the
vast corpus, of RusseII's writings over the years, I could not produce two opposing
quotations. I
Other problems are related to Russell's very personal writing style.

For example, Philip Jourdain (1879-1919) complained that certain observations and remarks made by Russell in some of his works were "taken
literally and thus misunderstood by superficial and serious people".2
Wood structured a similar argument showing that Russell did not always
expect to be taken literally. How, then, does one know when to take him
at face value, and when to read ad pedem literae? Because there is not a
concrete and objective answer, I will take him literally on some occasions
but not others. I will try to provide some criteria to determine whether
Russell expected to be taken literally or not.
On another occasion, when Russell was asked to write an article on
the recent developments of mathematical philosophy at the tum of the
century, Jourdain also lamented that the editor of the magazine asked him
to write an article that was "as romantic as possible".3 Moreover, some
of Russell's arguments (and statements) were purposely modified by the
publishers.
Some confusion may result from the methods used to prepare a
book. This issue is of fundamental importance in understanding some of
the possible flaws contained in his writing-not only those autobiographical, but those related to The Principles of Mathematics. According to
Wood:
1. Wood 1957., 212. I wiII have the opportunity to show, later on, that it is even possible to find
'opposing quotations' in the same piece of work. See, for example, RusseII 1903., 105 & 528.
2. Jourdain 1919., 454.

3. [bid.
xxiii
There were successive periods of intense thinking, each of which culminated in a

book which, in the end, was written rapidly. Russell hardly ever revised anything he
had written, and almost never re-read a book after it had been published. I
These sentences accurately describe Russell's mature methodology, the

better known to the general public. Russell himself confirmed Wood's
observation in a questionnaire answered in 1925. Nevertheless, The Principles of Mathematics did not follow the scheme described by Russell and
Wood. This book was written over a long period of time-even when one
only considers Russell's final attempt to produce it. (By 1898, Russell
already had a book in mind with similar content and the same goal: to
examine the foundations of arithmetic).2 Later, I will discuss (see: Chapter IV, 3, 4 and 5) that The Principles of Mathematics was not written
in its order of final presentation. It will be argued that parts I and II were
originally written after parts III, IV, V and VI. Furthermore, part VII
was written at least a~ear before the other sections. The appendices were
added when the book was at the printing house, and fundamental changes related to the inconsistencies were apparently made in galleys. There
is some evidence that Russell did not revise earlier sections of the book
when these last components were added; the material contained in the
additions contradicted some of the earlier work.
Another possible source of inaccuracies that deserves special emphasis resulted from Russell's oversimplification of complicated events.
Russell was aware, when writing his reminiscences, that his work would
be read both by the specialist and lay-people. Unfortunately, in order to
describe some ideas or events in a straightforward manner, it was necessary for Russell to introduce some distortion, perhaps unwillingly. A
detailed analysis of the course of the events involved in the preparation of
The Principles of Mathematics may have required a lengthy complex
essay. Russell may have prefered to oversimplify his description of these
events and to say that the book was written over a (short) period of time.
Or perhaps, in his old age, Russell was misled by his own false historical
clues.
1. Wood 1959a, 203.
2. Cf Blackwell 1984a, 275-277.
XXIV
Preface
Biographers of Russell have pointed out other problems in interpreting his work, such as his egotism and use of exaggeration, among
other things. 1
In order to eliminate most of the possible inaccuracies and inconsistencies found in Russell's memoirs and, possibly, some of his original
research, I have tried to verify all information from Russell with his own
unpublished material and with other sources. The sources include: unpublished manuscripts; personal correspondence with his family, wives,
mistresses, friends and colleagues; diaries, and other material published
at the time. Because, it is sometimes difficult to determine whether or not
to take Russell literally, I cannot claim that I have been exhaustive in my
task.
There are other points that the reader must keep in mind constantly. First, sometimes it is impossible to expand or explain ideas expressed
by Russell in greater detail, because the actual status of the manuscripts
is too incomplete. Often very limited information is available (sometimes
only one phrase) in which Russell expressed significant thoughts for the
development of my book. More importantly, the main goal of this book
is to explain why, how and when Russell discovered the 'paradoxes' (and
how others were discovered); its main purpose is not to present a detailed
historical reconstruction of how The Principles of Mathematics was written.
Finally, I would like to mention some points in relation to the
.footnotes and the bibliography. Footnotes, located at the bottom of each
page, have been used sparingly and are concise enough to avoid destroying the continuity of the book. I have tried, as much as possible, to
merely cite those references (in the footnotes) corresponding to explicit
quotations. As I mentioned before, the book is addressed to the general
public and not only to the specialist. Those readers interested in checking
all specific references should consult the original dissertation. 2 I have
1. See: Clark 1975a, 307; 1981a, 46; and, Wood 1957a, 76. Beware, the reader should not get an
incorrect impression. At a great extent, Russell's autobiographical writings are accurate and objective.
But, as I said before, in some cases, he did not expect to be taken literally.
2. See: Garciadiego 1983a.
xxv
also avoided references to complete bibliographical sources in the footnotes for the reason of continuity and to eliminate, as much as possible,
a false sense of academicism. The bibliographical sources contained in
brackets in a footnote are exclusively related to the information described
in that particular footnote and not to the narrative of the text.
The entire bibliography is contained at the back of the book. Each
source is uniquely determined by the name of the author and the date of
publication which always contains (even in the case when there is only
one publication) an alphabetical character. A Roman numeral following
the year of publication designates the volume, if the book was originally
published in multiple volumes. In a few cases, mostly related to some of
Russell's unpublished manuscripts, I have included more specific information (e.g., the part of the book orland chapter) because the pagination
of the folios begins again at "1" on several occasions. In the bibliography
the character "*" means that the particular source listed remains unpublished. 1 The following abbreviations are also used: "Ms" = Manuscript, "Ts" = Typescript and "BRA" = The Bertrand Russell Archives.
I have tried, when mentioning a personal letter, to indicate if it has
already been published. I have also attempted to rely on the original
sources but refered to English translations when they are available, to
assist the non-specialist reader. The bibliography contains both the original sources and the translation used, if any specific reference is needed.
Finally, on the few occasions that I provide a personal interpretation that
differs from the viewpoint of other authors I have also tried to summarize
the other researcher's position to give the reader a perspective from
which to choose between the two ideas.
1. However, it it important to keep in mind that an edition of Russell's collected works is in preparation and most of the unpublished essays will be available in due come. Thus, some of the sources of
my book will be readily available for the general public.
Acknowledgements
This book is based on a doctoral dissertation defended at the Institute for the History and Philosophy of Science and Technology of the
University of Toronto (Ontario, Canada) in January 1983. Since then, the
work has undergone substantial modification. Although the major thesis
of the book remains the same, the explanation of how, when and why
Russell discovered the set-theoretic paradoxes, changes have been made
on the basis of new historical data and novel analyses presented by several historians. A general introduction has been added to explain the origin
of some of the concepts used by Russell. Many stylistic changes were
made in the final revision; in fact, I believe that not one single sentence
from the original dissertation remained untouched. I would like to acknowledge the assistance of several institutions and colleagues, which has
been of tremendous value in the preparation of this work.
After completing my dissertation, subsequent research was generously supported by the Department of Mathematics of the Faculty of
Sciences of the U.N.A.M. (Universidad Nacional Aut6noma de Mexico)
and by Conacyt (Consejo Nacional de Ciencia y Tecnologia (Mexico)).
The Bertrand Russell Estate and The Bertrand Russell Archives allowed
me to obtain the material necessary to pursue my work. The copyright on
Russell's unpublished letters and writings rests with Res.-Lib. Ltd.,
McMaster University (Hamilton, Ontario, Canada), and the material has
been quoted with permission from The Bertrand Russell Peace Foundation and The Bertrand Russell Publications Committee. At The Bertrand
Russell Archives, I was assisted by Kenneth Blackwell and Carl Spadoni,
whose expert knowledge in the vast field of Russelliana proved helpful in
developing the ideas expressed in this book.
XXVlll
Acknowledgements
Numerous professionals facilitated my work through different

periods of its development. The original dissertation was developed under
the guidance of Charles V. Jones. During its first period of composition,
Tom Archibald, Ed Barbeau, Philip Enros, Jamie Whelan and Bob Findlay read different sections. Joe Dauben read a completely modified version of the book and recommended it to the editorial staff of Birkhauser.
Mrs. Elvira Pelayo provided invaluable support by helping to transcribe
some of Russell's original letters and Jaime-Oscar Falc6n greatly contributed by translating Hilbert's postcards into English. My colleague Elisa
Viso translated myoid computer files into the new PC compatibles. My
students Rodrigo Cambray and Ana Sacristan, after reading the entire
typescript, presented long lists of typographical errors. I would like to
express my deepest gratitute to Walter Purkert who made important and
profound technical commentaries on the entire typescript. Ken Blackwell,
Joe Dauben, Janis Langins, Albert C. Lewis and Sheila Turcon read and
criticized specific sections of the final typescript. Nevertheless, Ivor
Grattan-Guinness deserves a very special mention. His support was a
tremendous driving force towards the completion of the project. Also, I
greatly appreciate the immense patience showed by Erwin N. Hiebert,
Hans Wussing, Benno Zimmerman and Doris Womer waiting for the
final version of the book.
Most especially, I would like to acknowledge my deepest debt to
Nick White, who read the complete final version of the typescript and
made innumerable suggestions to improve my English style and grammar. The remaining mistakes are due to my impatience and stubbornness.
For permission to publish a modified version of my essay entitled
"The Emergence of Some of the Nonlogical Paradoxes of the Theory of
Sets, 1903-1908" (Historia Mathematica 12 (1985) 337-351), I am grateful to Academic Press. For permission to publish: Russell's letters to
Alys, Russell's missives to Moore, Hilbert's postcards to Russell, Russell missives to Hardy as well as most of the photographs I am grateful to
The Bertrand Russell Editorial Committee and The Bertrand Russell
Peace Foundation. Russell's photograph at age 13 [see: page 46] has
been published with the kind permission of 'The Principal and Fellows of
Newnham College, Cambridge (Great Britain)'. For permission to pub-
xxix
lish: Whitehead's missives to Russell, Hardy's letters to Russell and E.

H. Moore's missive to Cantor, I am grateful to Mrs. North Whitehead,
The London Mathematical Society (Great Britain) and the Institut MittagLeffler (Stockholm, Sweden), respectively. Unfortunately, up to now, it
has being impossible to locate the owners of the copyrights of BuraliForti's papers and Berry's works.
Finally, and most significantly, I would like to express my appreciation for the encouragement and support that I have received from my
wife through all these years. It is unfair and redundant to dedicate this
book to Lupita when it already belongs to her. It is, however, necessary.
The time will come when these

things which are now hidden from
you will be brought into the light.
Georg Cantor"
Chapter I. Antecedents
1.1 Introduction. - The main goal of this introductory chapter is

to describe some of the mathematical and philosophical ideas that will
allow the reader to understand the significance of Russell's contributions.
Contrary to the general concensus, mathematicians were well aware, in
the last few decades of the nineteenth century, of important issues on the
principles of mathematics. They were especially aware of problems concerning the foundations of the concept of number. In this chapter, my
analysis will not be historical, but will provide a logical foundation for
material to be presented later. Considering that the set-theoretic 'paradoxes' were discussed as such in Russell's The Principles of Mathematics
(published in May 1903) for the first time and that some of Russell's
ideas on the concept of number rely on Cantor's concept of transfinite
numbers, then it is natural to focus the discussion on Cantor's contributions to the theory of transfinite numbers. In particular, the discussion
will be centered upon those ideas presented by Cantor in his articles of
1895 and 1897, which played a fundamental role in Russell's final presentation of his concept of number. I will try to very briefly narrate the
reactions of some mathematicians and philosophers to Cantor's ideas.
Among the scholars who influenced Russell in the early stages of
his intellectual career, there are two authors who, for the purpose of this
1. Cantor took this quotation from the Epistle to the Corinthians as an epIgraph for his 1895 article.
The original Latin reads as follows: "Veniellempus, quo isla qUiZ nunc lalent, in lucem dies exlrahal
ellongioris iZvi diligentia" [Epistle 4:5, as quoted in Cantor 1932a, 282; 1915a, 85. See, also: Dauben
1979a, 239].
Antecedents
book, deserve special mention. These were not the only people who
influenced Russell's general intellectual development, nor necessarily the
most important. Many individuals influenced Russell, depending on the
area of interest (e.g., economics, philosophy, non-Euclidean geometries,
religion, etc.). With respect to the foundations of arithmetic, in particular
the concept of natural number, Cantor and Peano, in some way or another, put forward concepts, ideas, methods and approaches that Russell
would integrate to form the basis for his own original ideas, later known
as logicism. He also borrowed some elements from other sources and
fields. In fact, Russell constructed his Logicist Thesis by adapting some
mathematical concepts and by rejecting earlier profound philosophical
influences. Moreover, before he met Peano at the First International
Congress of Philosophy in Paris in 1900, his renunciation of his previous
neo-Hegelian beliefs compelled him to adopt new mathematical ideas to a
greater extent than the influence of Cantor and Peano.
1.2 Cantor's transfinite numbers.- The story of how Cantor

arrived at his ideas on transfinite cardinal and ordinal numbers has already been exquisitely narrated in other secondary sources. 1 Dauben, in
particular, has carefully depicted how Cantor turned his attention from
his original mathematical interests on number theory to the study of
trigonometric series, following a question suggested by Edward Heine
(1821-1881). Under Heine's advise in approximately 1869, Cantor attempted to clarify the uniqueness of the representation of an arbitrary
function by a trigonometric series. His research necessitated that he
studied the properties of specific sets of points. In some cases, these sets
contained an infinite number of points. Eventually, the study of properties of these sets, and their implications for the foundations of mathematics, became the focus of Cantor's mathematical career.
Already by 1883, Cantor was completely convinced that a new
understanding of the concept of number was required in order to continue
his research on the properties of infinite sets. He apologized to his colleagues because he knew he was challenging a well established conven1. See: Dauben 1971a; 1974a; and 1979a. See, also: Purkert 1986a, 1989a and Purkert & I1gauds
1987a.
tion. But Cantor did not present an ordered and clear development of his
ideas on transfinite cardinal and ordinal numbers until 1895. He began
this paper by defining what he meant by the concept of set.
Bya [set] we are to understand any collection into a whole M of definite and
separate objects m of our intuition or our thought. These objects are called the
"elements" of M.I
He then continued to define what he meant by the power or cardinal

number of a set. Nevertheless, the careful reader will notice immediately,
that in fact Cantor does not define what a cardinal number is; instead he
described how one obtains the cardinal number of any set. This 'power'
or 'cardinal number' is the result of a double process of abstraction.
First, one abstracts the nature of the elements, that is to say, one takes
for granted what the elements are. Second, one abstracts their order, that
is to say, one overlooks the arrangement in which the elements originally
appeared. In fact, one could substitute each and every element of the
primary set for a 'unit' indicating that there was an element originally
there. Cantor denoted the cardinal number of a set M by,
M.
As I said, the careful reader will notice that Cantor does not define the
notion of cardinal number. Instead he describes how we obtain the cardinal number of a set. According to Cantor,
since every single element m, if we abstract from its nature, becomes a "unit", the
cardinal number M is a definite [set] composed of units, and this number has existence in our mind as an intellectual image or projection of the given [set] M.2
So, if one considers the set {a,b,c}, one could think of its cardinal number as the set {laolb,Icl or simply {I,I,I}, represented by the number
113
11
1. Cantor 1915a, 85. Dauben's alternative translation reads: "By a 'set' we mean any collection M into
a whole of definite, distinct objects m (which are called the 'elements' of M) of our perception [Anschauung] or of our thought" [Dauben 1979a, 170].
2. Cantor, 1915a, 89.
Antecedents
Immediately following this discussion, Cantor defined what he

meant by an 'equivalence relation' among sets and some basic operations
among cardinal numbers (addition, multiplication and exponentiation). At
the tum of the century, mathematicians and philosophers, Russell among
one of them, were impressed by encountering definitions of these concepts and their operations without a single reference to the finiteness or
infiniteness of the sets involved.
In the fifth section of the paper, Cantor defmed the concept of
finite cardinal number. Cantor, relying on our mental capacities, assumed
that one could think of one single item eo, which is reclassified or renamed as a set Eo = {eo}. By carrying out the double process of abstraction, which removes from the set the nature and order of its elements,
one gets its cardinal number. The cardinal number of this set Eo is called
"one" and denoted by "1". Thus, one has
In the same way that one was able to conceive a unit eo, one might consider another unit el and obtain the union of the two, such that
Again one carries out the double process of abstraction to get the cardinal
number of this new set. It is called "two", which is denoted by "2":
One might add additional singular elements to this set and then obtain a
new series of sets:
~
= (E lt ~, E.J = (~,
~),
... ,
Proceeding as one did before, one will obtain the cardinal numbers (or
powers) of each one of the sets, denoting them by,
1. Cantor's notation (E." e.) indicates the union set of the set E.. with the singular element e
3,4,5, ... ,
Cantor pointed out that the notion of number (or the numbers themselves)
had not been used to define these finite cardinal numbers. Now, he
claims that if by "1'-1 is understood as the number immediately preceding
v in the above series", 1
and
then, by applying the definition of a sum of cardinal numbers, one obtains,
Ev
E v-1 + 1.
This property assigns the cardinal number one of their most important
characteristics: "every cardinal number, except 1, is the sum of the immediately preceding [cardinal number] and 1."2
Once one has been able to construct the cardinal numbers as the
following sequence,
1, 2, 3, ... , v, ... ,
one has to verify that this sequence really represents the cardinal numbers
as one knows them. First of all, one has to observe that all numbers of
the sequence are different from one another. Cantor also proves that all
numbers are greater than the preceding ones and less than ones that
follow. Another property of this sequence affirms that: "There are no
1. Cantor 1915a, 98.
2. Ibid., 99.
Antecedents
cardinal numbers, which, in magnitude, lie between two consecutive

numbers II and lI+ 1".1
The preceding sequence represents the cardinal numbers of fmite
sets. Each of the sets is built upon the previous one and all of them have
been constructed by the addition of the unity. Cantor has shown that all
these numbers represent cardinal numbers of finite sets.
But how does one get a transfinite cardinal number? One has to
conceive of an infinite set in order to obtain its transfinite cardinal number. Are there any examples of infinite sets in nature? To the best of our
knowledge there are none. Even the set of all bits of dust in our universe
is a finite set and its power or cardinal number is also finite. Perhaps,
personally, one cannot be able to count the elements of such set, but the
process of enumeration would cease some day. So, where does one get
an infinite set? Possibly, infinite sets could be created in thought processes like, for example, the set of all ideas. But there must be a more 'mathematical way' of doing it. The answer relies in the sequence of finite
cardinal numbers. Let one think for a moment that one already has the
completed sequence,
1, 2, 3, ... ,
lI, ...
That is to say, one is supposing that one already has all the finite cardinal numbers. One ignores its cardinal number, but let it be called Aleph
null-and denoted by No. So, one has that
No = {I, 2, 3, ... ,
lI, ... }
Now, one may question if this symbol (No) represents a finite or a transfinite cardinal number. 2 Without examples of infinite sets in nature, how
can one prove their existence? One way of doing it is by showing that the
1. Cantor 1915a, 99.
2. I will use the term 'transfinite', following Cantor's own terminology, while discussing cardiual
numbers, instead of the word 'infinite', because this last word has a negative connotation ('that which
is not finite'), while the first term has a positive meaning ('beyond finite'). Is it possible to use this
word 'transfinite' before the acceptance of an actual infinite?
whole sequence of finite cardinal numbers has properties that contradict

the proper nature of the finite ones. Let one think in the same sequence
plus one single element (eo) called one, and represented by "1". The
cardinal number of this new sequence is equal to the sum of the cardinal
number of the set with one element {I} and the cardinal number of the
whole series, that is to say,
~o
1.
But, it is easy to prove that this new sequence is equivalent to the previous one. In other words, it is possible to find a one-to-one correspondence among all elements of both sequences, in such a way that each and
every element of the first sequence will correspond to a unique element
of the second sequence and vice versa. "For we can think of this reciprocally univocal correspondence between them: to the element eo of the first
corresponds the element 1 of the second, and to the element JI of the first
corresponds the element JI + 1 of the other".1 According to previous
results of Cantorian set theory this means that the cardinal numbers of
both sequences are equal,
but this contradicts one of the properties already shown to characterize

the finite cardinal numbers, implying that this new cardinal ~o does not
represent the cardinal number of a finite set. In short, ~ is a transfinite
cardinal number. In this example, if the reader pays close attention, he
will notice that the second sequence is a subset of the first one. 2 Cantor
showed that this first transfinite cardinal number had very special properties. He demonstrated, for example, that
1. Cantor 1915a, 104.
2. Richard Dedekind (1831-1916) used this last property to define an infinite set [Dedekind 1893a, 63].
Later on, Dedekind probably became acquainted with this property through Charles s. Peirce (1839-1914), who sent him a copy of a paper [1881a] where Peirce used the same property to "characterize
the difference between finite and infinite sets" [Dauben 1977b, 127].
Antecedents
If one considers the two following sequences,
1, 3, 5, 7, ... , 211-1, .. ,
and,
2, 4, 6, 8, ..... , 211, ..... .

Then, it is easy to prove that each of these two series is equivalent to the
sequence of all finite cardinal numbers (in the first case associate the
211-1th number to the 11th number (f(lI) = 211-1) and in the second case
associate the 211th number with the 11th number (f(lI) = 211. In addition,
it can be shown that the union of these two sequences is also equivalent
to all finite numbers; therefore, the addition of their cardinal numbers (~
+ No) is also equal to the cardinal number of all fmite cardinal numbers
(No) [see: page 139 note 2]. After some other considerations, one may
also show that,
Once Cantor was able to exhibit infinite sets, then he proceeded to

prove that No was not a unique transfinite cardinal number-although, it is
unique relative to its definition. In fact, he then went on to show that No
was the smallest of all transfinite cardinal numbers. He not only affirmed
that there was an unlimited sequence of transfinite cardinal numbers
but showed that when this sequence was given in its totality, then it is
possible to think in terms of another sequence beginning with N"" and so
on without end. (Cantor promised to prove these assertions later on). In
fact, Cantor had previously shown that given any finite or infinite set, it
was always possible to show the existence of a greater set, in particular
the set of all its subsets. So, if one considers the existence of any infinite
set, then there is an infinity of infinite sets. In other words, once one
takes the first and smallest step in defming an infinite set, there is no way
to stop. Obviously, these properties are so unsound that the theory seems
to be the product of a charlatan. Cantor himself was so impressed that he
even wrote on one occasion, to Dedekind saying, "I see it, but I cannot
believe it." 1
Subsequently, Cantor then studied the ordinal numbers. First it was
necessary to define a simply ordered set. According to Cantor,
we call a [set] M "simply ordered" if a definite "order of precedence" rules over its
elements m, so that, of every two elements ml and m2' one takes the "lower" and the
other the "higher" rank, and so that, if of three elements ml , m2, and m3' m l , say, is
of lower rank then m2, and m2 is of lower rank than m3, then ml is of lower rank than
m3 2
It is obvious, although I will not discuss the issue in detail here,

that it is possible to find different rules or laws for simply ordering within the same set. For example, the set {I, 2, 3} might be simply ordered
following its natural order of precedence under the law defined by the
relation "-<", and one might find another order by using the relation ">-"
and listing the set as {3, 2, I}.
Now, every ordered set has an ordinal type and it is denoted with
a single bar above the letter (e.g., M). One gets this ordinal type by
abstracting the nature of the elements of a set and by retaining the order
of precedence among them. So, once again, as in the case of the cardinal
numbers, this ordinal type is a simply ordered set composed of units
which retain the same order of precedence as the original elements in the
set M. When one has a one-to-one correspondence among the elements of
two sets, these sets are called "equivalent" (M - N); if the elements of
two sets are such that under the one-to-once correspondence the corresponding elements maintain the same rank of order as they did in the
original sets, then the sets are called "similar" (M == N).
As it was mentioned before, in order to obtain the ordinal type of
a set one carries out one form of abstraction-the nature of the elements.
1. Cantor to Dedekind, June 29, 1877; Cavailles 1962a, 211 & Dauben 1974a, 105.
2. Cantor 1915a, 110.
10
Antecedents
Once one has obtained the ordinal type of a set, say M, then one also
abstracts the order of the elements to obtain the cardinal number (M) of
the same set. From here it follows that if one has,
then,
M == N, (or M = N)
M - N, (or M =
N).
That is to say, the equivalence of sets follows from their similarity, but
the converse is not necessarily true. Cantor was convinced that the finite
ordinal types were not of any special interest, because they were identical
with the finite cardinal numbers. But the transfinites offered a completely
different situation. For example, the set
and the set,
1, 2, 3, ... , ", ...

... , ", ... , 3, 2, 1
have exactly the same number of elements but with the inverse order of
precedence. Therefore, they have the same transfinite cardinal number.
But, it is obvious that they do not have the same order among their elements and, therefore, they must be of different transfinite ordinal types.
In fact, when a new relation of precedence is defined among all the
elements of a set M (with ordinal type "a") by inverting the "lower"
numbers in "higher" numbers and vice versa, the "higher" numbers to
"lower" numbers, one gets another simply ordered set denoted by *M
with ordinal type ,,*a" .
From here, Cantor went on to derme the addition and multiplication of ordinal types; showing that, contrary to the case of transfinite
cardinal numbers, the commutative law was generally not valid for transfinite ordinal types. Let us define by""," the ordinal type of the sequence,
1, 2, 3, ... ,
II, ,
such that" -< ,,+ 1
11
then, it is possible to show that 1 +w is different from w+ 1. Iffis a new

element of our sequence of the finite cardinal numbers, one has
1 +w = if, {n}} and w + 1 = ({n}, /).
The set {n} is similar to the set if, {n}). But the sequences {n} and ({ n} ,
f) cannot be similar-the first one has no last term, while the second one
has a highest element in rank.
After defining the multiplication of ordinal types, Cantor also
showed that the distributive law was generally not true. To conclude,
Cantor also studied the properties of two different ordinal types: "The
ordinal type 71 of the [set] R of all Rational Numbers which are Greater
than 0 and Smaller than 1, in their Natural Order of Precedence"! and
"The Ordinal Type 'Y of the Linear Continuum X".2
The second part of this article appeared two years later. 3 The central issue of discussion here is what Cantor called 'well-ordered [sets]'.
Cantor's definition (in 1897) reads as follows,
We call a simply ordered [set] F "well-ordered" if its elements I ascend in a
definite succession from a lowest II in such a way that:
I. There is in F an element II which is lowest in rank.
II. If F' is any [subset] of F and if F has one or many elements of higher
rank than all elements of F', then there is an element f' of F which follows immediately after the totality F', so that no elements in rank between f' and F' occur in F. 4
1. Cantor 1915a, 122
2. Ibid. 133. Cantor defined the 'linear continuum' as the set {x x is a Real number and 0
I}.
S;
S;
3. Some historians have tried to explain and justify why it took Cantor two more years to publish this
second part of the article, when it was completed only six months after the first part. I will discuss
these questions in chapter II.
4. Cantor 1915a, 137-138. Cantor assures us that this defmition, 'apart from the wording', is identical
to the one he had previously introduced in 1883 (see: page 27). I will discuss some {)f the oonsequences of the wording of this definition in relation to the origin and development of the 'paradoxes' in
chapter II. It is a fact that in the 1890s, eminent mathematicians confused and misinterpreted Cantor's
concept. In 1899, Cantor established that "a multiplicity is said to be well-ordered if it satifies the
condition that every submultiplicity of it has afirst element [see: Letter: Cantor to Dedekind, 28 July
1899; Cantor 1932a, 443 and translated in: van Heijenoort 1967b, 114].
12
Antecedents
Condition (I) assures the existence of a first or minimum element of the

whole set. Condition (II) says that if there is a sequence of elements
contained in the set in such a way that a subsequence does not exhaust
the original sequence, then there is a first or minimum element of those
elements immediately following the subsequence. For example, the following sets are well-ordered sets,
{ai, a2, ... , a" ... },
{ai, a2' ... , a" ... , b h b2, ... , bl" ... },
{a h a2, ... , a., ... , b h b2, ... , bl" ... , c l , C2, C:J},
where,
But the sequence of all finite cardinal numbers in their inverse
order of precedence { ... , 11, ... , 3, 2, I} is not a well-ordered set-the
set itself does not have a first element, or a lowest element in rank,
contrary to condition (I). Nevertheless, it is possible to well-order this set
by changing the original order of the elements.
On the other hand, now consider F = {x I 0 :::;;; x ::;;; I} and F' =
{ x I Ih < x < 0/3}. One cannot find an element r F satisfying condition (II). But recall that, Cantor already believed by 1883 that it was
possible to rearrange the order of any set to produce a well-ordered set.
In fact, a positive proof of this last principle known as the Well-Ordering
Theorem was fundamental for the sound development of the theory of
transfinite cardinal numbers.
After discovering a good number of theorems about well-ordered
sets and their segments,l Cantor defined ordinal numbers as the ordinal
types of well-ordered sets. Using his earlier theorems, Cantor was able to
prove the Tricotomy Law, that is if Ct and fJ are any two ordinal numbers,
then one and only one of the three following relations holds among them:
1. "Iff is any element of the well-ordered [set] F which is different from the initial element f., then
we will call the [set] A of all elements of F which precede f ..... 'the segment of F which is defined
by the elementf. , .. [Cantor 1915a, 141].
O!
= {:J or
O!
<
{:J or
O!
>
13
{:J.
Finally, Cantor was able to show that given any two cardinal numbers it
was possible to compare them and determine if they were equal, or if one
was greater or less than the other one.
Nevertheless, Cantor failed to resolve several questions that he
considered of major importance for the development of his theory.
Among other things, he was unable to prove his Well-Ordering Theorem
which asserted (as already mentioned) that any set could be well-ordered.
That is, given any set (ordered or not) it was possible to define a new
order among its elements in such a way that there was a first element and
every element had an immediate successor. For example, given the set of
all rational numbers R-Cantor's 1895 notation-it is possible to rearrange its elements [see: page 14]. In this case, the elements are arranged
in such a way that every subset (one may think of them as arrows or
columns) of R always has a first element and therefore is well-ordered.
Concluding, some historians have correctly stated that the final presentation of Cantor's ideas was "elegant but ultimately disappointing". 1
1.3 Some reactions.-
It has been postulated that criticisms of
Cantor's theory of transfinite numbers contributed to his mental illness.

Dauben and Grattan-Guinness have demonstrated that such criticism only
had a limited effect upon Cantor's health, and that the personality of his
father was not a negative one as some authors have suggested. 2 Grattan-Guinness has even gone further, suggesting that there was no connection
between the development of Cantor's ideas and the progressive stages of
his mental disfunction, with the possible exception of 1899 when Cantor
was corresponding with Dedekind.
It is also tempting to believe that there was a general negative
reaction to Cantor's work. Possibly, too much attention has been paid in
the secondary literature to the opposition of Leopold Kronecker (1823-1891) and Henri Poincare (1854-1912), that is to say, their opposition to
1. Dauben 1979a, 217.
2. See, for example: Dauben 1978a, 535-550; 1979a, 271-288 and Gratlan-Guinness 1971b, 355-374.
14
Antecedents
1 ..... 1
1
/'
,/
.t.
1 ..... 1
/'
,/
,/
,/
.......................................................
n
2
n
3
n
4
n
n
.......................................................
1. Reordering the Rational Numbers.
15
general set theory. The response of other mathematicians like Jacques

Hadamard (1865-1963) and Gosta M. Mittag-Leffler (1846-1927), among
others, has not been considered. Of course, there were differences in
understanding and interpretation of the basic ideas among those who
accepted the theory.
Frege, for instance, expressed his main criticisms in the early
1890s. He was in complete agreement with the basic hypothesis of the
theory (i. e., the acceptance of an actual infinity given in its totality-although Cantor had not claimed to have established any) and with
many of the results, but he had serious objections towards some of the
basic principles. Paradoxical as it sounds, Cantor had previously agreed
with Frege's critical arguments when he reviewed Frege's Die Grundlagen der Arithmetik.l Let me attempt, first, to very briefly describe the
background of Frege's objections.
Frege had been working on the foundations of arithmetic for at
least three reasons. In the first place, he was in complete disagreement
with a prevalent form of psychologism in philosophy and logic. He also
opposed the Empiricism of John S. Mill (1806-1873) and, finally, he
confronted Kant's conception of the principles of mathematics.
Some sorts of psychologism treated the categories of psychology as
proper categories in philosophical analysis. That is to say, some intellectuals, mainly in the XIX century, approached some philosophical problems from the standpoint of psychology, relying on our personal mental
capacities. For Frege this tendency was infected with subjectivity, as the
following quotation reflects.
From the prevailing logic I cannot hope for approval, [... ] for it seems to be thoroughly infected by psychology. If people consider, instead of things themselves, only
subjective representations of them, only their own mental images-then all the most
delicate distinctions in the things themselves are naturally lost and others appear
instead which are logically quite worthless. 2
1. Cantor 1932a, 440-441.

2. Frege 1893a, xiv; quoted and translated in Resnik 1980a, 25.
16
Antecedents
Perhaps, it would be easier to understand Frege's objection by analyzing

some of the Cantorian elements affected by psychologism. For example,
Cantor's definition of 'set' depends upon our personal and individual
capacity of abstraction [see page 3]. The reader immediately notices that
Cantor clearly uses the words intuition and thought. But do we all have
the same capacities of abstraction, intuition and thought?
The problems relating to psychologism have even more serious
implications when one realizes that Cantor's fundamental concept of
'cardinal number' or 'power of a set' requires not only one, but two
processes of abstraction. Let assume for the moment that one ignores
Cantor's definitions, and takes, for example, the set A = {a pen, a pencil, a piece of chalk}. Let one abstract first the nature of the elements
and then their order. Did one obtain its 'cardinal number' "3", or ...
something else? It is possible to understand now, following Frege's conception of the nature of arithmetical propositions, why he objected to
some of Cantor's definitions. Frege's logical character of arithmetical
propositions was highly objective, while Cantor's definitions were indeterminate due to their psychological origin.
Peano, perhaps best known in mathematics for his set of postulates
for the natural numbers,l was also deeply concerned with the foundations
of mathematics. Peano's concern was more technical and internal to mathematics than philosophical. According to him, most of the difficulties
found in the foundations of mathematics had their roots in the ambiguity
of language. For this reason, mathematics needed a symbolism capable of
expressing each and every formula without ambiguity. The creation of
this symbolism turned out to be Peano's most significant contribution to
the principles of mathematics.
Peano wrote to Cantor about "fundamental shortcomings in Cantor's introduction of the transfinite numbers and raised specific questions
about Cantor's treatment of the finite numbers".2 Unfortunately, the
Cantor-Peano correspondence has apparently been lost, and it is impossi1. See: Peano 1973a, 113.
2. Dauben 1979a, 227.
17
ble to elaborate upon what other historians have already discussed. Dauben asserts, based upon drafts of Cantor's letters to Peano, that Peano's
main criticism about the finite numbers was Cantor's implicit use of the
'induction principle'. Peano suggested that induction was an 'unproven
postulate', but Cantor insisted that the principle of induction was a direct
consequence of his own construction of the theory of the finite cardinal
numbers.
These criticisms are important for the future development of this
book. Since Russell adopted concepts similar to those of Frege and Peano, it seems reasonable to suppose that Russell would share their same
criticisms. On the other hand, these criticisms might also provide some
insight for understanding the thinking behind Russell's discovery of the
set-theoretic 'paradoxes'.
My theory stands as firm as a rock; every arrow directed

against it will return quickly to its archer. How do I know this?
Because I have studied it from all sides for many years; because
I have examined all objections which have ever been made against
the infinite numbers; and above all, because I have followed its
roots, so to speak, to the first infallible cause of all created things.
Georg Cantor.
Chapter II.
A standard interpretation
2.1 Introduction. - The popular interpretation of the origins of

the set-theoretic 'paradoxes' is now almost a classic story, and it is to
this account that I now turn along with some criticisms. The term popular is not used in a derogatory way: it means that this interpretation is the
best known and accepted by the general public.! This standard interpretation is simply a summary of the account given by mathematicians and
professional historians of mathematics, among others, in general encyclopedias, philosophical and mathematical encyclopedias and dictionaries,
textbooks of mathematics and textbooks on the history and philosophy of
mathematics. 2 The following quotations are representative of the prevailing chronicle:
There remained logical difficulties in the theory of transfinite numbers and paradoxes
appeared, such as those of Burali-Forti and Russell. This again led to different
schools of thought on the foundations of mathematics. 3
1. I would call the reader's attention to the fact that there is no unique interpretation of the origins and
development of the set-theoretic 'paradoxes'. Several historians may object that they do not propose the
same interpretation as their colleagues, because they do not agree with them in all details. In this
chapter, however, I have attempted to consolidate the various accounts, including the discrepancies, in
order to allow me to answer potential criticism of my own proposal. I will call this account: a standard
interpretation.
2. See, among others: Kleene & Feferman 1982a XI, 630-639; van Heijenoort 1967a Y, 45-51;
Iyanoga & Kawada 1977a II, 982; Curry 1977a, 3-9; Eves 1983a, 473-477; Kline 1972a, 1183-1185.
3. Stroik 1989a, 163.
20
A Standard Interpretation
The crisis was brought about by the discovery of paradoxes or antinomies in the
fringe of Cantor's general theory of sets. Since so much of mathematics is permeated
with set concepts and, for that matter, can actually be made to rest upon set theory as
a foundation, the discovery of paradoxes in set theory naturally cast [sic] into doubt
the validity of the whole foundational structure of mathematics.
Most historians agree, at least, on the following points: First, that

Burali-Forti discovered his inconsistency in 1897; second, that Cantor
communicated a similar proposition to Dedekind by a letter in 1899; and,
third, that Russell discovered his own inconsistency in 1901 and informed
Frege, also by letter, in June of 1902. As I have mentioned earlier, there
have been points of conflict among historians researching this area. Some
do not agree with others on specific dates, 2 others disagree on immediate
causes. 3 Nevertheless, the basic generalities and consequences of the
account remain unchanged, despite the conflicting interpretations. That is,
they all conclude that a critical attitude towards Cantor's ideas resulted in
the discovery of the set-theoretic 'paradoxes'; in turn, these arguments
revealed the need for a reexamination of the foundations of mathematics.
Consequently, the inconsistencies directly stimulated the creation of three
philosophical schools-Logicism, Formalism and Intuitionism-in mathematics. Moreover, it could also be argued that the semantic 'paradoxes'
were a direct product of the logical ones. 4
The objections that I will raise to this interpretation will cast doubt
on some issues and point the way to a more logical and accurate interpretation of events. If this new account seems more plausible, then some
basic features of the traditional story would have to be rejected. In particular, I will argue that the 'paradoxes' did not appear as a necessary consequence of a disapprobation of Cantorian set theory and, that they did
not stimulate by themselves the origin of the philosophical schools. In
addition, I will then be able to argue the possibility of the semantic paradoxes having a plurality of sources or origins.
1. Eves 1983a, 474.

2. This is due, in part, to the fact that the original discoverer has proposed alternative dates. Cf,
Russell 1944a, 13; 1959a, 58; and, 1967a, 147.
3.
Cf.
van Heijenoort 1967a, 46; Coffa 1979a, 32 & 37 and Garciadiego 1983a, 151-167.
4. See: Eves 1983a, 475.
21
2.2 The Burali-Forti 'paradox'.- Some historians propose

that the set-theoretic inconsistencies originated from criticism of Cantor's
theory.! When one considers the nature of Cantor's departure from traditional mathematics, it is not surprising that there was intense opposition.
Like most innovators, Cantor had to confront a formidable opposition
bent on preserving the status quo. Those attempts to impugn Cantor's
work came from different quarters and with varied aims. One of his
contemporanies called him a "corrupter of youth'? in an attempt to stop
the publication of his articles. Others protested against the methods that
Cantor used. So, it is not surprising if some historians point Burali-Forti
as a detractor of Cantor.
If one was to believe that Burali-Forti discovered an inconsistency
in Cantor's set theory in 1897, as the traditional interpretation would
have it, then it is reasonable to suppose that he was attempting to find
flaws in Cantor's theory. Indeed, as I mentioned above, recent studies
argue that Burali-Forti published his finding as a consequence of his
negative attitude towards Cantor's new theory on transfinite cardinal and
ordinal numbers. That this is implausible is shown by Burali-Forti's
background and later contributions.
Burali-Forti's mathematical interests developed around Peano's
program. 3 One of Peano's main goals was to construct a symbolism capable of expressing any mathematical proposition without ambiguity. It is
logical to suppose, if there is no explicit evidence to the contrary, that
Peano's students would follow their mentor in most essentials. Burali-Forti's published work at the time satisfies this criterion: his publications
show him as Peano's student, as well as an admirer and follower of Cantor. If, in fact, Burali-Forti had discovered an inconsistency laying at the
roots of Cantorian set theory (which was not the case, as I will argue) it
was not because of his opposition, but rather as a consequence of his
acceptance of Cantor's basic principles. 4 Peano, as well as Frege, object1. See, for example, Copi 1958a, 283 note 13; Kennedy 1970a, 593; among others.
2. Schoenflies 1927a, 2 (referring to Kronecker).

3. Kennedy 1980a, 85-86 & 187.
4. This also applies accurately to Russell's discovery, as we will see later on.
22
A Standard
Interpretation
ed to Cantor's construction of the finite cardinal numbers, but both accepted the general principles and results of the theory of transfinite numbers. Burali-Forti, being a follower of Peano, probably accepted the same
basic principles. Burali-Forti's published work shows that it is not just a
conjecture, but a fact.
By 1894, only three years before Burali-Forti had supposedly
discovered the inconsistency, he had already published a small book
entitled Logica Matematica. This volume was the product of a course on
mathematical logic given at the University of Turin, where he was an
assistant to Peano. That same year, Burali-Forti published an article
"expressing Cantor's theory of transfinite numbers in Peanoesque symbolism" illustrating once more the close association between Cantor's and
Peano's mathematical interests. I That was not all. Only two years later,
in 1896, Burali-Forti was still working on Cantorian set theory, trying to
solve one of Cantor's most important problems: whether the Trichotomy
Law-which allow us to compare any two numbers and to establish if
they are equal, or one is greater or less than the other one- holds between any two transfinite cardinal numbers. Finally, not much later, on
March 27, 1897, Burali-Forti read an article at the meeting of the Circolo
Matematico di Palermo. This paper, according to most historians and
mathematicians, contains the first published 'paradox' related to the
foundations of set theory. 2 The argument has been described in the secondary literature in the following terms.
The series of ordinal numbers, which is well-ordered, must have the greatest of all
ordinal numbers as its order type. Yet the type of the above series of ordinal numbers, when followed by its type, must be a greater ordinal number, for (3 + 1 is
greater than (3. Therefore, a well-ordered series of ordinal numbers containing all
ordinals numbers itself defines a new ordinal number not included in the original
series. 3
There are, at least, three alternative enunciations of this 'paradox' in the

mathematical literature. The one above has the great advantage of not
1. Moore & Garciadiego 1981a, 321.
2. Bell 1945a, 279; Bourbaki 1960a, 49; Fraenkel et ai1958a, 8; Hobson 1905a, 170; Kattsoff 1948a,
88; Poincare 1905a, 822-823; among others. Cf. Moore 1978a, 308.
3. Cajori 1919a, 401.
23
having to construct a fJ + 1 ordinal. A second version arises in the following way. One could
assign the ordinal number 0 to the series of all ordinals, and then note that since 0
must be an element of that series, it must be the order type of the [segment] of all
ordinals [less than] O. Here we have the result that the series of all ordinal numbers
is similar to one of its own [segments], which is impossible. I
Burali-Forti's supposedly original statement (and third version of the

same argument) might be described in the following terms: First, it is
shown that the class of all ordinal numbers (No), in their natural order of
precedence, is 'perfectly-ordered' [see: pages 27-29], thus it has an ordinal number O. Therefore, 0 is an element of No and a ::::;; 0, for all a
contained in No. One also notes that a < a+ 1. "If we now substitute
0+ 1 for a in the former statement, and 0 for a in the latter, we obtain:
0+1 ::::;; 0 and 0 < 0 + 1
which is an obvious contradiction."2
I was arguing that Burali-Fortis's printed record shows that he was
not opposed to set theory, but that he was a member of the mathematical
community dedicated to the growth of set theory and mathematical logic.
Did Burali-Forti modify his conception of Cantor's theory by March
1897 when he published his now famous article? It is difficult to believe
so, if one reads the original paper with an impartial mind. Burali-Forti's
1897 article was intended as another positive contribution to the body of
Cantor's ideas, although Cantor did not believe that Burali-Forti's reasoning supported his point of view. The goal of the paper was, as Burali-Forti's put it, "to prove that there actually exist transfinite numbers (or
order types) a and b such that a is not equal to b, not smaller than b, and
not larger than b.,,3 In other words, Burali-Forti's main concern was to
prove that the Trichotomy Law did not hold for all pairs of transfinite
ordinal numbers. Burali-Forti defined No as the class of all order types of
1. Copi 1958a, 284.
2. Ibid., 281.
3. Burali-Forti 1897a, 154; translated by van Heijenoort 1967b, 105.
24
A Standard
Interpretation
'perfectly-ordered' classes. Supposing, for the purpose of a reductio ad

absurdum argument, that trichotomy held for this class (No), he then
proved that No was itself a perfectly-ordered class and consequently that
its order type 0 was a member of No, as well as 0+ 1. But No was defined as the class of all order types, and "if O! f No, then O! < 0; hence
both
o < 0+1 and 0+1
< 0
held, a contradiction."l Burali-Forti consequently inferred that No violated the trichotomy law, contrary to the hypothesis of the proof; thus, in
conformity to the indirect procedure, he concluded that it was impossible
to order the order types in general and the ordinal numbers in particular.
A slight alteration of the original argument could have produced
the "Burali-Forti paradox" as it is known today. But at the time, without
any opposition to set theory or any intention of impugning it, he did not
claim to have discovered any inconsistency within the foundations of this
theory. For him, the contradiction appeared only as a step in the procedure of reductio ad absurdum. Here a question of historical interpretation
is raised: should one talk of inconsistencies in mathematics simply because the word contradiction is mentioned in an argument? Many historians and mathematicians erroneously have thought that because Burali-Forti mentioned the word contradiction in his paper this explicitely
implies that he claimed to have discovered one. But this is not the case.
Burali-Forti was simply trying to prove by reductio ad absurdum that the
hypothesis was false and this method required supposing the hypothesis
true and arriving at a contradiction. The employment of the hypothesis,
as an initial premise, generated the inconsistency. But once the hypothesis is seen to imply a contradiction it is thereby proved to be false. And
this was the case, according to Burali-Forti.
There are some items in the literature of the history of mathematics
stating that scholars reacted immediately to Burali-Forti's paper. It has
been claimed that dozens of papers dealt with it, forcing the need to
25
re-examine the foundations of set theory. 1 Some writers even state that a
third crisis was provoked by the discovery. 2 It is not clear, however,
whether they claimed that research in mathematics would collapse as a
result of the inconsistencies or, whether those were times of insecurity or
difficulty in certain mathematical circles. 3
In terms of time, what should one understand by the word 'immediate'? A month? A year? Scholars supporting this interpretation have
failed to mention even one publication between the years 1897 and 1902
that was concerned with this topic. Even more significantly, three years
after the original publication of the paper, the review published by Giulio
Vivanti (1859-1949) describing Burali-Forti's attainments failed to mention the existence of any contradiction or paradox. Perhaps, in 1900,
before the legend was born, Vivanti was not biased and was not yet
reading what was not there.
On the other hand, there is a group of historians who have commented on the absence of discussion of the 'paradoxes' in the years immediately after the publication of Burali-Forti's paper. For example,
Bochenski claims that scholars did not pay too much attention to the
'paradox' at the beginning. 4 Bunn asserts that not only Burali-Forti' s
inconsistency went unnoticed but two others as well: one related to the
set of all alephs, and the other with the set of all powers.5 Kennedy
mentions that Burali-Forti's paradox "went almost unnoticed until Bertrand Russell published a similar antinomy in 1903."6 In another essay,
1. van Heijenoort 1967b, 104.
2. Eves 1983a, 474; Fraenkel 1947a, 17.

3. Although their viewpoint has been recently questioned, historians discussed an early crisis that
originated with the discovery of incommensurable quantities in the Vth century B.C. [Tannery 1887a,
98; Heath 1956a II, 112-113. Cf Knorr 1975a, 306-308. See, also: Dauben 1984a, 84-89]. It has been
argued that another crisis took place in the XVII century as a result of George Berkeley's (1685-1753)
criticisms to the nascent Differential Calculus of Isaac Newton (1642-1727) [Boyer 1949a, 224; Eves
1983a, 474].
4. Bochenski 1961a, 388.
5. Bunn 1980a, 237. Unfortunately, Bunn is not specific on the names and places of publication
relating to these two paradoxes , .
6. Kennedy 1980a, 86.
26
Kennedy had presetlJed a slightly modified version of his interpretation. 1

But, in all the above cases, historians assumed that Burali-Forti's had
already discovered the inconsistency by the time Russell publicly revealed
his own. Copi has provided the most detailed explanation of why there
was not an immediate response to the 'paradox' of Burali-Forti. 2 Although Copi's arguments could had supported those of others who also
claimed that there was not an immediate reaction to Burali-Forti's papers,
his explanation, however, has not attracted too much attention and his
arguments were not adopted by others-but as often happens, the exception confirms the rule. 3
But Copi has failed to explain how, why and when Burali-Forti's
'contradiction' became well known by the mathematical community at
large. Regardless of which interpretation one accepts (that Burali-Forti
did not describe a paradox in the paper or that the influence of the paper
was not immediate), the paper was to become more influential later. By
1903, Russell had publicly identified Burali-Forti's argument as a possible difficulty. The argument was soon transformed from a reductio ad
absurdum procedure into a real contradiction in its own right. On the
other hand, if the statement was considered to be a contradiction already,
then it suddenly captured the attention of the mathematical community. In
either case, it is necessary to elucidate how the transformation took
place-although, I hope to have already made clear the unsatisfactory
character of the traditional interpretation with respect to both explanations.
Russell is probably to blame because he was the one who first
credited Burali-Forti with the discovery of an inconsistency in the theory
1. Kennedy 1963a, 262.

2. Copi mentions the following possible reasons, among others: first, he incorrectly states that perhaps
the Rendiconti was not well known outside Italy before 1900; second, that the journal was published in
Italian and perhaps not many mathematicians outside Italy spoke Italian; third, the journal made heavy
usage of Peano's newly created symbolism and even fewer mathematicians could had followed the
technical development of the essay; fourth, and finally, Copi argues that few mathematicians knew any
philosophy and, that few philosophers were capable of following any contemporaneous mathematics at
the time [Copi 1958a, 281-283].
3. Curry 1977a, 5.
27
of transfinite numbers.' Jourdain exacerbated matters by disseminating

this understanding among British and German mathematicians in his
paraphrase of Russell's writings. 2 Burali-Forti may have contributed to
the disorder himself. Being confused by Cantor's definition of well-ordered sets, Burali-Forti later introduced his own conceptualization which in
turned forced the introduction of some later explanations.
In 1883, Cantor had defined the notion of a well-ordered set using
the following conditions: an ordered set is well-ordered if:
i) it contains a first element;
ii) any element with a successor has an immediate successor; and,
iii) any finite or infinite set of elements which has a successor has
an immediate successor. 3
The complexity of Cantor's wording is easily grasped when one
notices, as we did before, that these three conditions might be expressed
by saying that a set is well-ordered if any subset (different from the null
set-although this was not a set for Cantor) has a first element [see: page
28 note 2]. The complexity of the definition makes it easier to understand
why Burali-Forti, Hadamard, and possibly others, misunderstood Cantor's definition by keeping only the first two conditions to define well-ordered sets.4 Obviously, before Burali-Forti realized his mistake, late in
1897, he did not have the same understanding of the concept of well-ordered set as Cantor. In the now famous paper, Burali-Forti introduced
another concept of his own, that of a perfectly-ordered class. This new
idea, however, still differed from Cantor's concept. Burali-Forti defined
this new concept by taking Cantor's first two original criteria and adding
a third requirement different from Cantor's which limited the concept.
Without Cantor's third stipulation the concept was too loose, so Burali-Forti added a new one of his own. This reads,
1. Russell 1903a, 323.

2. Jourdain 1904b, 64-65; Moore & Garciadiego 1981a, 334.
3. Cantor 1932a, 168.
4. Burali-Forti 1897a, as translated in van Heijenoort. 1967b, 107; Hadamard 1897a, 201.
28
for every x in A, if x has an immediate predecessor then there is some predecessor y
of x such that y has no immediate predecessor and that only finitely many elements of
A are between y and x. I
That not all perfectly-ordered classes are well ordered sets follows from
the condition in which y has no immediate predecessor, meaning that
there is at least one subset with no first element. But the reader should
not forget that Burali-Forti misread Cantor's definition and their interpretations of a well-ordered are different. According to Burali-Forti's understanding of well-ordered sets, all perfectly-ordered classes are well-ordered sets, but not vice versa. That is to say, Burali-Forti's concept of
a well-ordered set would be a particular case of the one defined by Cantor. But, as I have emphasized, Burali-Forti was only taking part of Cantor's original definition.
Some sets which are not well ordered, in Cantor's sense, are those
sets containing subsets of order types w [see: page 10]. For example, the
set
{I, 2,3, ... ,
P, , , P,
,3,2, I},
with its natural order of precedence, is not well-ordered in Cantor's

sense. 2 It is clear that there are perfectly-ordered classes that are not
Cantor's well-ordered sets, because of Burali-Forti's new way of defining
well-ordered sets and because some of Burali-Forti's perfectly-ordered
classes are generated by classes containing subclasses of order
1. Burali-Forti 1897a, 157; translated in: Moore & Garciadiego 1981a, 323.
2. It is also easier to characterize a well-ordered set following Jourdain's proposal that "in order that a
simply ordered [set] M should be well-ordered, it is necessary and sufficient that M should contain no
([subset]) of type 'w" [Jourdain 1904b, 65-66]. Jourdain was likely unaware that a analogous condition
had already been suggested by David Hilbert (1862-1943), who probably was acquainted with such
condition through his own correspondence with Cantor [see: page 11 note 4]. Hilbert's new definition
of a well-oredered set reads as follows: "a well-ordered [set] ... is characterized ... [by the condition]
that not only in the [set] itself but also in every [subset] there exists a first element" [Hilbert 1900b,
446].
3. For a detailed discussion of a set being perfectly-ordered see: Young & Young 1928a, 102.
29
Nevertheless, in early 1897, while Burali-Forti was misunderstanding

Cantor's definition, he would have been correct making the opposite
claim, that is, that all perfectly-ordered classes are well-ordered sets. It
did not take long before Burali-Forti changed his mind. In late 1897
Burali-Forti read Cantor's second part of the "Beitdige zur Begriindung
der transfiniten Mengenlehre". Here Cantor positively proved the trichotomy law for any two order types of well-ordered sets. Obviously, this
does not mean that both articles are proving opposite results. Keep in
mind that Cantor is using his well-ordered set concept, while Burali-Forti
is basing his proof upon his peifectly-ordered class idea, and that both
these concepts are different. In October 1897, Burali-Forti published a
brief note apologizing for having misunderstood Cantor's definition.
Nevertheless, as he also mentioned, this mistake did not force him to
modify his previous result in any respect. Through the development of
his proof of the trichotomy law, he did not depend upon Cantor's definition. I Burali-Forti now correctly concluded that every well-ordered set,
in Cantor's sense, was also perfectly-ordered. The converse, however,
was not true.
Unfortunately, Burali-Forti's brief note did not clarify all of the
confusion associated with these two concepts. Copi argues that the article
in which Burali-Forti explained his mistake "passed unnoticed by many
investigators".2 He points out that in 1914, a mathematician devoted a
whole article to explain that Burali-Forti's concept was different from
Cantor's idea. 3 Even by 1928, more than thirty years after Burali-Forti's
original date of publication, William H. Young (1863-1942) and his wife
Grace Chisholm (1868-1944) called the attention of the mathematical
community to the errors and misunderstanding associated with these
articles. 4
1. Cf Grattan-Guinness 1977a, 27 note 2, who asserts that "[Burali-Forti] acknowledged the excessive
narrowness of his definition, while stressing that his paradox was not affected by the change. "
2. Copi 1958a, 281.
3. See: Hagstrom 1914a.
4. Young & Young 1928a.
30
It is easy to understand why mathematicians confused the definition

of a well-ordered set at the turn of the century. Burali-Forti himself does
not appear to have smoothly recognized and acknowledged the difference
between his definition and Cantor's. The question was complicated further when Burali-Forti claimed in a letter to Couturat "that there was a
printer's error in [my second 1897] article, interchanging instances of
well-ordered and perfectly-ordered" . 1
By now, in 1906, Burali-Forti believed, unlike his previous interpretation, that every perfectly-ordered class was a well-ordered set but
not vice versa. This unfortunately is false. Burali-Forti's conceptualization of the relation between the concepts of a perfectly ordered class and
a well-ordered set went through three stages. During the first stage, in
early 1897, he was partially correct; that is to say, he incorrectly defined
a well-ordered set, although under his incorrect conceptualization, he
correctly asserted that every perfectly-ordered class was a well-ordered
set. In the second stage, late 1897, he was fully correct, after fixing his
mistaken definition. In the last stage, in approximately 1906, he was
undoubtedly mistaken!
Another objection to the traditional interpretation rests on the following argument. If, in fact, Burali-Forti thought he had found an inconsistency in the foundations of the theory of transfinite numbers, why did
he not claim so in his second article of 1897? Even if there was no inconsistency in the original paper, Burali-Forti could have produced one in
the subsequent paper if he had wanted to do so. Once Cantor had published the second part of his seminal article on the foundations of the
theory of transfinite numbers in late 1897, Burali-Forti could have demonstrated an inconsistency in two different ways. The first way would
have been accomplished by slightly modifying the thesis of his theorem
and then showing that both propositions, his and Cantor's, contradicted
each other. The second would have been by altering the original procedure of reductio ad absurdum and then showing that an inconsistency
could be generated if the class of all ordinal numbers was considered.
But Burali-Forti did not resort to either of these two tactics. As men1. Moore & Garciadiego 1981a, 341.
31
tioned above, he simply apologized and argued that there was no need to
modify his concept of a well-ordered class because he did not use Cantor's definition of a well-ordered set in his argument. 1
Even after the publication of Russell's The Principles of Mathematics in May 1903, Burali-Forti did not participate in the polemics resulting
from the discovery of the 'contradictions', especially in the polemics
which followed Zermelo's presentation of the proof of the Well Ordering
Theorem in September 1904. 2 Burali-Forti's lack of interest in the discussion surrounding these inconsistencies provides additional evidence
that he did not consider himself responsible for the discovery of the
'contradiction' .
If Burali-Forti ever thought that he had discovered a 'contradiction'
in his first article in 1897, it did not lie in his argument by itself, but in
the apparent inconsistency between his finding and Cantor's theorem on
the comparision of ordinal numbers. Nevertheless, Burali-Forti explained
to Couturat that if either he or Cantor had succeeded in proving or disproving the trichotomy law, they did so to support different underlying
conceptualizations of well-ordered sets. But once this difference (in well-ordered sets) was recognized, there was no contradiction. 3 So, Burali-Forti intended to deny the existence of the 'contradiction' between his
and Cantor's findings. But one should not misunderstand Burali-Forti's
motives: he was not intimidated by a public debate. Later he was to take
part in the dispute precipitated by his proposition of a "unified system of
vectorial notation". 4 The vehicle of this polemic was the journal L 'Enseignement Mathematique. Some of the most important members of the
mathematical community took part in it: for example, Peano and Felix
Klein (1849-1925). Neither was Burali-Forti afraid of proposing and
defending radical or unpopular ideas: he rejected, for example, the theory
1. In fact, in the same previously mentioned letter to Couturat, Burali-Forti explicitly denied to have
discovered any inconsistency [Couturat 1906a, 228-229]. But, according to Poincare, "even after this
correction, Burali-Forti's explanation is not satisfactory" [Poincare 1906b, 304].
2. See: Moore 1982a, 92-141.
3. Couturat 1906a, 228-229; Moore & Garciadiego 1981a, 341.
4. Kennedy 1970a, 594.
32
of relativity in the 1920s. Despite the contradictory evidence, proponents

of the standard interpretation have concluded that "to Burali-Forti must
go the distinction of having been the first to publish a modem instance of
the logical paradoxes". 1
2.3 Cantor's 'paradox'.- The standard interpretation implied
that, although Burali-Forti was the first mathematician to publish an

inconsistency in the theory of sets, he was not the first person to discover
one. 2 It is generally thought that the first example of a 'paradox' was
discovered by Georg Cantor, the creator of the theory of transfinite
numbers. But historians are still discussing Cantor's role in the origins of
these inconsistencies. Some historians have related Cantor's original
discovery of the 'paradox' to a universal set-or set of all sets;3 others to
the set of all finite cardinal numbers. 4
A modem version of Cantor's paradox-probably one of the best
known-could be described in the following terms: it is always possible
to show the existence of a set greater than any given set-in particular,
its power set (the set of all its subsets including the null set). That is to
say, Cantor showed that the cardinal number of any given set is always
less than the cardinal number of its power set. In symbols,
=
<
P(M).
Now, let one consider the 'universal set U', the set of all sets. Then
P(U), being a set, has to be a subset of the universal set and, therefore,
its cardinal number has to be less or equal to that of U,
But this contradicts Cantor's theorem.

1. Copi 1958a, 283 (my emphasis).
2. Ibid., 283.
3. Dauben 1979a, 242; Dou 1970a, 65.
4. Clark 1975a, 80.
Cf.
van Heijenoort 1967b, 113.
33
Most scholars claim that Cantor described this inconsistency to

Dedekind in a letter dated July 28, 1899. 1 But even the date of the letter
is now in doubt. Grattan-Guinness asserts that Zermelo did a poor job of
editing Cantor's papers. Grattan-Guinness showed that there was not just
one letter from Cantor to Dedekind, as implied in the original German2
and the English translation,3 but two letters. 4 Others have dated the discovery of the 'paradox' as early as 1883,5 or 1895,6 or 1896. 7 So, at
least, the date of the discovery is disputed.
I will return shortly to the question of the date of Cantor's discovery. First, let me discuss how and why Cantor, according to other historians, came upon his discoveries of the 'paradoxes'. Recently, Dauben
has studied this question. According to him,
[Cantor) had come upon [the paradoxes) in the course of trying to establIsh his
comparabilIty theorem for transfinite cardmal numbers, and in the attempt to deal
with the related questions of whether every transfinite power was necessarily an
aleph, and whether every set could in fact be well-ordered. 8
Bunn's explanation is very similar.

Now my guess is that [Cantor) thought that any power must at least be comparable
with the alephs [ ... ) and, therefore, that two incomparable powers must both be
greater than any aleph. I speculate further that it was in pursumg this line of thought
that he came upon the antinomies and subsequently made his distinction between
consistent and inconsistent muitiplIcities.9
The careful examination of these two passages shows that both

scholars refer to the 'paradoxes' (antinomies, according to Bunn) in the
plural, implying that Cantor came upon not just one inconsistency, but
1. Dou 1970a, 65; van Heijenoort 1967b, 113.
2. Cantor 1932a, 443-447.
3. van Heijenoort 1967b, 113-117.
4. Grattan-Guinness 1974b, 126-128.
5. Paukert 1989a, 56.
6. Dauben 1979a, 192; Fraenkel 1967a, 21; Zlot 1957a, 100.
7. Grattan-Guinness 1971b, 365; 1980a, 75; 1981a, 15.
8. Dauben 1979a, 192.
9. Bunn 1980a, 251.
34
two or more. According to the traditional interpretation, one of the 'paradoxes' is expressed in terms of the set of all cardinal numbers and the
other in terms of all ordinal numbers. This would mean, if correct, that
Cantor anticipated Burali-Porti's discovery.
So, according to the standard interpretation, Cantor probably came
upon the 'paradoxes' as early as 1883, 1895 or 1896; and the discovery
perhaps resulted from an attempt to prove the Well-Ordering Theorem or
some other results associated with it, while he was attempting to write a
precise and clear presentation of his theory of transfinite cardinal and
ordinal numbers. The evidence for this view rests in the letters that Cantor wrote to Dedekind in 1899 and to Jourdain in 1903. Contrary to the
claims of historians who support the standard interpretation, however, the
content of these sources was not focussed on the discovery of any 'paradoxes'.!
One should stress that Cantor's desired greatly during his lifetime
to find a positive proof of his Continuum Hypothesis and his Well Ordering Theorem. 2 Cantor made several attempts to create 'proofs' of this
last theorem in particular. I believe one of these 'proofs' was the argument that he sent to Dedekind in the summer of 1899. 3 In the exposition
of his proof, Cantor demonstrated the impossibility of conceiving all
mUltiplicities or totalities as unities. There were some multiplicities such
that the assumption that all of their elements 'were together', implied a
contradiction. Cantor called these multiplicities 'absolutely infinite' or
'inconsistent multiplicities' and, as an example he mentioned "the totality
of everything thinkable".4 Cantor divided the multiplicities into two
classes: consistent and inconsistent. He called the consistent multiplicities
1. Moore 1978a, 308-309.
2. In his famous list of problems, Hilbert described Cantor's Continuum Hypothesis in the following
terms: "[e]very system of infinitely many real numbers i.e., every assemblage of numbers (or points),
is either equivalent to the assemblage of natural integers 1,2,3, ... or to the assemblage of all real
numbers and therefore to the continuum, that is, to the points of a line: as regards equivalence there
are, therefore, only two assembloges of numbers, the countable assembloge and the continuum [...J
[Hilbert 1900b, 446].
3. Moore 1978a, 309.
4. Cantor 1932a, 443; [see, also, van Heijenoort 1967b, 114].
35
'sets'. The concept of inconsistent mUltiplicity was used by Cantor in his

'proof' of the Well-Ordering Theorem.
The first step in his proof, as communicated to Dedekind, was to show that the
collection 0 of all ordinals is inconsistent or absolutely infinite. For if 0 were
consistent, then it would have an ordinal number a greater than every member of 0;
but a was a member of 0 and so a < a, an absurdity [ ... ].
Cantor continued his argument by establishing that every transfinite cardinal
was an aleph: If some infinite set V had no aleph as its cardinal, then V would have
a subcollection V' in one-one correspondence to O. Since 0 was inconsistent, so was
V' and hence Vas well. But by definition only sets (i.e., consistent multiplicities)
possessed cardinal numbers, and hence V had no cardinal. Consequently every transfirute cardinal was an aleph, and every set could be well-ordered. 1
Once again, as in Burali-Forti's case, Cantor's did not consider his absurdity to be either a contradiction in itself, or an argument supporting
the existence of a 'paradox'. This absurdity is only a step in the procedure termed reductio ad absurdum. The false assumption used as a starting point was to suppose that O-the collection of all ordinals-was consistent. But this implies nothing apart from the reductio proof.
On the other hand, it is possible that by 1890 or 1902 Cantor had
visualized the impossibility of considering as a set 'the totality of everything thinkable' once he had proven his theorem that the cardinal number
of the power set is always greater that the cardinal number of the original
set. 2
I am assuming that Cantor conceived of the existence of inconsistencies within his theory: his own distinction between consistent and
inconsistent multiplicities shows that he was aware of that possibility. In
his letters to Dedekind, Cantor mentions the need to distinguish between
consistent and inconsistent multiplicities in order to avoid falling into
contradictions. When and how did he become aware of the need to make
this distinction? Cantor himself asserts that he clearly saw the need for
2. As we will later on, Russell affirms that he followed a similar path in the discovery of his own
'contradiction' of the set of all sets which are not members of themselves [Russell 1903a, 362 second
footnote]. Moreover, in his The Principles o/Mathematics Russell accidently and inadvertently attributed the 'contradiction' of the greatest cardinal number to Cantor. Perhaps, Russell thought that this
'contradiction' was to obvious and that Cantor could not had missed it [Ibid., 362].
36
this distinction by 1883, but does not provide any specific reason why"
It is possible that Cantor had intuitively felt the existence of the 'para-
doxes' since 1883-as he claims, but we lack sources supporting the

assertion that Cantor had formally found the 'paradoxes' by then. 2
Even if Cantor did not formally discovered the 'paradoxes' by
1883, it is reasonable to suppose that there were other motives-some
even non mathematical-to strengthen his intuition. For example, it has
been convincingly argued that Cantor's theological believes could had
allowed him to transform his belief in the theological absolute into the
mathematical absolute of collections too large to be sets. 3
Here, I believe, paraphrasing Russell, 'the master has been guilty
of a very subtle fallacy' and Cantor's recollections might confuse the
general reader. If Cantor had in fact visualized the distinction between
consistent and inconsistent multiplicities by 1883 and he had arrived, as
his letter to Dedekind indicates, at the conclusion that the multiplicity of
all cardinal numbers was an inconsistent one, then he could had not
defined aleph-zero (No) as the cardinal number of the set of all finite
cardinal numbers, in his seminal essay of 1895. According to his conceptualization, all inconsistent multipliticies could not have cardinal numbers. Following Grattan-Guinness's interpretation, I am inclined to believe that Cantor formally discovered the 'Burali-Forti paradox' in 1896,
while he was trying to develop a clear and concise final presentation of
his ideas.
In order to support the claim that Cantor did not attempt to disclose
the existence of any 'paradoxes' in his correspondence with Dedekind
and to show that the 'paradoxes' were first publicly reported independently of Cantor in 1903, then one should try to find Cantor's comments or
1. See: Letter: Cantor to Grace Young, March 9, 1907; quoted in Moore & Garciadiego 1981a, 342.
See, also, letters from Cantor to Hilbert of 26 September 1897 and IS November 1899, contained in:
Purkert & Ilgauds 1987a, ISO & IS4; and Purkert 1986a, 320-321.
2. See: Letter Cantor to Jourdain, 4 November 1903, quoted in Grattan-Guinness 1971a, 116-117 and
Purkert & Ilgauds 1987a. On this occasion, Cantor used the words 'intuitiv erkannt' [intuitive
knowledge].
3. Dauben 1979a, 294-297.
37
reactions to these arguments in sources written after 1903. 1 In reading

Cantor's seminal papers of 1895 and 1897, one recognizes (without the
aid of historical interpretation) the absence of any discussion of the nature and effect of the 'paradoxes'. Without considering posthumous publications, Cantor's last original contribution in mathematics was printed in
1897. Is there any other evidence or source that should be used? Cantor
delivered a lecture at the Jahresversammlung at Keisel in September
1903. Unfortunately, the surviving records of the meeting printed at that
time mention only the title of the discourse: "Bemerkungen zur Mengenlehre". 2 This is too broad to suggest any unquestionable conjecture on
the precise content of the lecture.
There is, nevertheless, a written reminiscence of Cantor's lecture
published several years later. In 1911, Arthur Schoenflies (1853-1928)
described Cantor's defence of his theory of sets against the objections of
the French philosophers in his 1903 lecture. 3 Schoenflies also recalled
that the validity of the Principle of Excluded Middle was the focus of
interest of the participants in the discussion which followed Cantor's
presentation. 4 Schoenflies went on to say that this discussion influenced
him in the writing of his own paper on the 'paradoxes' published in
1906. But there is no reason to conclude, from Schoenflies comments,
that the topic of Cantor's lecture was the 'paradoxes'. If, in fact, the
'paradoxes' were the central theme of the talk, then I would argue that it
would be more reasonable to assume that Cantor explained, perhaps
unhopefully, how the distinction between consistent and inconsistent
multiplicities helped to solve the question of the 'paradoxes'. This supposition does not invalidate the possibility that the discussion following the
lecture did not center around the 'paradoxes'.
1. It is clear, from Cantor's correspondence, that he considered the need to distinguish between sets
and inconsistent multiplicities as a positive indication of the consistency of the theory of transfinite
numbers. Nevertheless, after 1903, once Russell's The Principles of Mathematics, Cantor had to face
the understanding and interpretation of others.
2. "Remarks on set theory". See: Krazer 1903a, 519.

3. Schoenflies 1911a, 251 note 2.
4. This principle (also known as tertium non datur) asserts that either a proposition is true, or the
contradictory of the proposition is true; but no other outcome is possible. This principle is the basis of
the reductio ad absurdum proofs.
38
A few years after the lecture, once Schoenflies had published some
of his ideas on the 'paradoxes', Cantor was still convinced that his distinction between these two concepts of consistent and inconsistent multiplicities was the key to the whole affair. 1 At about the same time as the
talk, September 24, 1903, scholars were discussing the inconsistencies of
Russell and Cantor, 2 but apparently, the French philosophers ignored
these discussions. Perhaps by the time Schoenflies wrote his recollections, he had already been persuaded by the polemics surrounding Zermelo's proof of the Well-Ordering Theorem to incorporate the French
philosophers into the context of Cantor's lecture. But there is not, as the
current historical evidence now indicates, any generalized objection to
Cantorian set theory before 1904 among the French philosophers. Notwithstanding that some of them have already questioned the pertinence of
publishing the French translations of Cantor's work in Acta Mathematica
edited by Mittag-Leffler.
If one accepts the argument that Cantor sent Dedekind evidence of
a 'contradiction', several unanswered questions remain. Suppose that
Cantor did in fact discover the inconsistencies, without solving them.
According to most historians, such a discovery could have ocurred in
either: 1895, 1896 and 1899. The main argument in favor of the first
date was that Cantor published his fundamental article of 1895, in which
the concept of set was limited. As Dauben argued, the "definition of set
in part I of the "Beitrage ... " might well have been calculated to exclude
cases where paradoxes might arise".3 Before 1895, however, Cantor had
expressed his understanding of the concept of set in terms similar to
those used in the "Beitrage ... ", although the wording of the definition
was not exactly the same. 4
Dauben maintained that Cantor regarded the existence of the inconsistencies as constructive additions to set theory, because these arguments
could be used to prove other mathematical propositions (e.g., that every
1. See: Moore & Garciadiego 1981a, 342.
2. [Hardy] 1903a, 263; Frege 1903a, 253.
3. Dauben 1979a, 226.

4. Cantor 1932a, 244 note 1.
39
transfinite power was necessarily an 'aleph' or cardinal number). Dauben

further claims that Cantor "was perfectly willing to assert that his theory
was completely free of contradictions, and this he regarded as an established fact". 1 Dauben bases his argument on a letter from Cantor to
Albert Eulenburg (1840-1917), dated February 28, 1886. But according
to Dauben's interpretation, Cantor discovered the paradoxes in 1895-nine years after he expressed his thoughts to Eulenburg. Perhaps, by the
time Cantor wrote to Eulenburg, he had not yet discovered the 'paradoxes'; and, on the other hand, it is also possible that Cantor elucidated a
solution for them as soon as they emerged.
Other historians have claimed that Cantor discovered the 'paradoxes' in 1896 "doubtless in connection with preparation of the 1895-1897
paper" .2 This is perhaps the most plausible interpretation. Sources that
support this thesis include a letter from Cantor to Jourdain; and the fact
that once Cantor had discovered the 'paradoxes' and had made the distinction between consistent and inconsistent multiplicities, he could have
not defined No-as he did in the first part of his 1895 paper-as the cardinal number of the set of all finite cardinal numbers. This last multiplicity
was inconsistent and could not have had a cardinal number associated
with it. The long interval, (two years) between the publication of the first
and second parts of the "Beitdige ... " also supports this conjecture. In
other words, it is believed that Cantor delayed the publication of the
second part of the article while he was searching for a solution to the
inconsistencies. Nevertheless, Cantor finally published the second part,
without any reference to the possible existence of the 'paradoxes' and
their solutions.
A third group of historians argue that Cantor discovered his 'paradox' in 1899,3 when formulating his thoughts to Dedekind in his correspondence. I think, however, that the goal of this correspondence was not
to discuss the existence of the 'paradoxes' and their solutions, although
the 'paradoxes' were mentioned with a proposed solution. As I said
1. Dauben 1979a, 246.
2. Grattan-Guinness 1980a, 75.

3. Dou 1970a, 65; van Heijenoort 1967b, 113.
40
A Standard
Interpretation
before, I believe that it was only another unsatisfactory effort to prove

the Well-Ordering Theorem.
2.4 Epilog.- Cantor anticipated Burali-Forti in his discovery of

the 'paradoxes'. As I have argued throughout this chapter, however
neither of them was really interested in the 'paradoxes'. Nevertheless,
according to the standard interpretation, Cantor communicated his discoveries to some of his colleagues through personal correspondence. Some
of these letters became known publicly only in 1932 when Zermelo edited Cantor's papers. But there are references to "Cantor's paradox" in the
mathematical literature prior to 1932. Since none of these writers knew
the contents of the correspondence at the time of their writing, it must be
argued that someone else made the same discovery and attributed it to
Cantor. This person made these findings public before 1932.
To be consistent, I will have to say that Russell anticipated Cantor
in the publication of the inconsistency. However, the evidence from the
literature supporting this point is rare and subtle. I Russell, who was the
first mathematician to publish the inconsistencies attributed (although
perhaps unconsciously) one of the inconsistencies to Cantor, believing
that the inconsistency was too obvious to have passed unnoticed. 2 Russell
became intrigued by this argument, no later than the end of 1900. When
he wrote an article on the recent developments of the foundations of
mathematics shortly after [1901e], he called attention to a 'mistake' or
'fallacy' in Cantor's ideas. But it was not until May 1903, when The
Principles of Mathematics was available at the bookstores, that mathematicians could have learned that Russell's understanding had shifted and
now he was describing Cantor's argument as a contradiction. But, I will
discuss in detail how this 'paradox' (as well as others) emerged in chapter IV, sections 3, 4 & 5.
1. Grattan-Guinness is 8 good exception [see: Grattan-Guinness 1978a, 129).
2. Russell 19038, 362.
I was especially attracted to problems concerning the foundations

of mathematics. I wished to believe that some knowledge is certain and I
thought that the best hope offinding certain knowledge was in mathematics.
[ .. .J. But it took me nearly twenty years to find all the justification
that seemed possible and even that fell far short of my youthful years
Bertrand Russell.
Chapter III.
The philosophical
and
mathematical background
to
The Principles of Mathematics,
1872 - 1900.
3.1 Introduction. - The chief purpose of this third chapter is to
discuss those aspects of Russell's philosophical and mathematical background which led to the ideas expressed in The Principles of Mathematics
published in May, 1903. 1 Emphasis will be given to events ocurring
before July 1900, when Russell met Peano.
In order to understand his philosophical and mathematical background, it is necessary to divide Russell's intellectual development into
different stages:
Russell's early education: in this section we discuss some events
that took place during Russell's childhood and adolescence that Russell
himself thought to be significant for his later understanding of the foundations of mathematics.
1. See, also: Grattan-Guinness 1976a, Griffin 1991a, Griffm & Lewis 1990a, Rodriguez 1988a &
1991a.
42
Russell's Philosophical & Mathematical Background
Russell's studies at Cambridge: I will outline the reasons why

Russell's interests shifted from mathematics to philosophy. (Two important consequences of his sojourn at Trinity College have already been
examined by others in detail: First of all, Russell's acceptance of what he
called Idealism, German Idealism, Hegelianism or neo-Hegelianism, with
his later incomplete rejection of it. 1 Secondly, some aspects of his fellowship dissertation on the foundations of geometry which resulted in a
polemic with Poincare, have been discussed elsewhere).2
The immediate consequences of Russell's indoctrination into neoHegelianism (ca. 1893-1898): in this period Russell developed his interests in the foundations of physics, which were based on earlier related
investigations on the foundations of geometry.3 By 1897, Russell considered the relation between the categories of number and quantity as not
only the most important problem in the foundations of mathematics, but
also the most difficult to solve. The study of this period would help us
understand why Russell came to consider mathematics to be a defective
science.
Russell's efforts to write a book on the principles of mathematics:
In 1898, while under the neo-Hegelian influence, Russell attempted to
write his 'big book' on the principles of arithmetic for the first time.
Russell always credited Peano's influence as the turning point in his
intellectual development in 1900. I will attempt to show, however, that
Russell slowly adopted Cantor's ideas. These ideas permitted Russell,
against his own inclinations, to transform the philosophical antinomies
inherent in the philosophical idea of an infinite number into (logical)
contradictions related to the mathematical concept of an actual infinite
set. This was possible once he had abandoned the neo-Hegelian doctrines.
Contrary to Russell's own reminiscences and comments,4 I will show
that Cantor's work, in particular his theory of transfinite cardinal and
1. See: Spadoni 1977a, Griffin 1991a & Rodriguez 1991a.
2. See: Gross 1974a, Richards 1988a, Griffin & Lewis 1990a.

3. Wood holds the contrary opinion that "the easiest approach is to explain how Russell followed a
chain of thought which began with a problem in Dynamics, and led him in turn to Geometry, Analysis,
Symbolic Logic and Grammar" [Wood 19S9a, 191, my emphasis].
4. See: Letter: Russell to Iourdain, April IS, 1910; printed in Grattan-Guinness 1977a, 132-134.
43
ordinal numbers, played a profound role in the development of Russell's

own formulation. Cantor's influence on Russell will be even clearer in
the next chapter, once I have analysed how parts of The Principles of
Mathematics were written after Russell had studied Cantor's seminal
articles of 1895 and 1897.
3.2 Russell's childhood and adolescence.- Ravenscroft

Estate, Russell's original place of birth, has recently been restored, in an
attempt to keep it as close as possible to its 1870s settings. Russell was
born on Friday May 18, 1872 at 17:45 hours in Trelleck, Monmouthshire, Wales. He was a big, fat, ugly, vigorous and strong newborn. In
his early childhood, a profoundly grow up in a liberal environment,
conducive to his intellectual development, was provided by his parents:
John Russell (1842-1876), known as Viscount Amberley and Viscountess
Amberley (nee Katherine Stanley (1842-1874. In addition, Mill was his
godfather and Helen Taylor (?-1907) (Mill's step daughter) was his godmother. The Amberleys both supported and practised such advanced
ideas as free thinking in religion, women's suffrage, extramarital relations and birth control.
Unfortunately, everything changed by 1876, when Russell was four
years old. First, his mother (and his sister Rachel (1868 died of diphtheria in 1874. Less than two years later, his father also passed away.
After threatening a lawsuit, Russell's paternal grandparents won custody
of the children over the two tutors that were originally intended by Russell's parents to assume guardianship of Bertrand and his older brother
Frank (1865-1931). The tutors assigned were Douglas A. Spalding (ca.
1840-1877) and Thomas J. Cobden-Sanderson (1840-1922). Spalding, a
close friend of Charles Darwin (1808-1882) and Mill, attempted to prove
that behavior was regulated through Darwinian evolution. I In fact, some
of his experiments and observations on instinct were carried out at the
Amberley's residence, having Lady Amberley's personal assistance.
1. See: Haldane 1954a, Garciadiego 1990a, Gray 1962a & 1967a.
44
Cobden-Sanderson is known today as an excellent bookdesigner and

binder.'
In 1876, the orphan Bertie, as close friends and relatives called
him, moved to Pembroke Lodge in Richmond Park. His paternal grandfather, Lord John Russell (1792-1878), who served twice as England's
Prime Minister, died only two years later. Consequently, the responsability of Russell's education rested on his grandmother, Lady Russell (1815-1898), (nee Lady Frances Anna Maria Elliot), who was known by
close friends of the family as 'the deadly nightshade'. This woman exerted a most powerful influence on Russell's early education. Her approach
was completely the opposite to that of his parents. She believed in discipline to build character, and she attempted to inculcate her standards in
her grandchild.
Plate 3. Lord Russell.
Plate 4. Lady Russell.
[Russell's] early life was one of great ease. Besides himself there were in
Pembroke Lodge only his grandmother, his Aunt Agatha, and his Uncle Rollo. Frank
was there only for school vacations. To look after the Russells there were eleven
1. See: Hawskwell-Briggs 1972a.
45
servants, so no Russell ever had to do any ordinary household chore. In addition,

there was a succession of governess (in the early years) and tutors (in the later
years), nearly all foreigners, who looked after Bertrand's intellectual development.
None of these stayed very long because Lady Russell did not want Bertrand unduly
influenced by anyone but herself and her children. When there were signs that
Bertrand and a tutor had developed a personal as distinct from a formal relationship,
the tutor was dismissed often without even being allowed to bid Bertrand goodbye. At
lesson time there would simply be a new tutor. I
When reminiscing in his older years, Russell complained not of

having an unhappy childhood, but a solitary one. Russell began his study
of geometry when he was eleven years old. For some unknown reason it
was his older brother Frank who provided his first lessons, rather than
one of the governess or tutors.
Mathematics was what I hked. My first lesson in mathematics I had from my
brother Frank, who started me on Euclid, and I thought it the most lovely stuff I'd
ever seen in my life. I didn't know there was anything so nice in the world. But I
remember that the first lesson was a dissapointment because he said: "Now, we start
with axioms." I said: "What are they?" and he said: 'Oh, they're things you've got
to admit although we can't prove them." So I said: "Why should I admit them if you
can't prove them?" and he said: "Well, If you won't we can't go on. " And I wanted
to see how it went on, so I admitted them pro tern. 2
Although Russell was terribly disappointed by the fact that one had
to accept the axioms as if they were true without any proof, he was fascinated by geometric proofs themselves. It is not really known whether the
study of Euclid was careful, deep or complete. Russell did not comment
on the definitions and common notions, as well as more advanced and
difficult matters. It is well known that mathematicians in the past were
dissatisfied with the uneasy text of the Fifth Postulate-the one on parallelism. But they accepted without question that the other axioms could
not be proved. There was a need for a starting point. Several scholars,
including Russell, have pointed out that one should not treat this anecdote
about Russell at age eleven as unimportant, because he was concerned
with the validity of the foundations of geometry from then until he was
thirty-eight years old.
1. Slater 1982a, 1.
2. Russell 1959c, 65.
46
Plate 5. Russell at age eleven, approximately.
47
Unfortunately, at Pembroke Lodge, Russell did not find the free

intellectual environment necessary to express his ideas and raise his
doubts and questions. There were no liberal and young thinkers whith
whom to communicate his thoughts. In his adolescence Russell developed
his interests in sex, religion and mathematics. He expressed his thoughts
in a sort of diary written in Greek characters, so the family would not
pay much attention to it. At the age of sixteen he described his difficulties in coping with his sexual passions and recorded some thoughts of
suicide. Fortunately, for our story, he decided not to kill himself because
he wanted to know more about mathematics.
In this intellectual diary, young Russell also exhibits his constant
dissatisfaction with religious teaching. In particular, he expressed doubts
on immortality, an idea that he would reject even after his indoctrination
into German Idealism. The brief diary also shows that Russell was a
devoted believer in Darwinian evolution: not only did he consider evolution to be possible, he stated that evolution was practically a proven fact.
At this time, Russell also held a mechanistic approach to the nature of
Man. He pictured the human species, ignoring the writings of Rene Descartes (1596-1650), as an 'ingenious machine'.
Possibly, Russell's first reading in philosophy was The Common
Sense of the Exact Sciences written by William K. Clifford (1845-1879)
and published posthumously in 1885. When Russell was fifteen years old,
his tutor recommended this book. Clifford had recently died at an early
age. Today, he is known, in the mathematical community, for his contributions to the field of linear associative algebras. He is also known for
his children literature, having written The little people, a collection of
tales. But, most importantly, at least for Russell, he wrote extensively on
the foundations of geometry. Unfortunately, Russell's notes from the time
did not survive, if there were any. It is only possible to analyze the
immediate effect of Clifford's book on Russell's intellectual development
from Russell's later recollections. Unfortunately, such recollections, as I
mentioned in the preface, are not always completely accurate.
In this book Russell found solutions to old problems and plans for
future studies. His reading of Clifford helped him to comprehend how
48
non-Euclidean geometries are logically consistent, yet independent from

the Euclidean geometry. Clifford also developed Euclidean geometry
from a different set of axioms. Russell could not accept those Euclidean
propositions that were proven by superimposing the figures, in particular
Prop 1-4. 1 Clifford set forth this proposition as one of the axioms, substituing for the fifth postulate, and eliminated these proofs.
Not all of Clifford's arguments were assimilated by Russell. For
example, Clifford believed that the nature of number was independent of
the order of counting; this premise does not seem to have elicited an
immediate response from Russell. Moreover, it was not until 1903, when
Russell's The Principles of Mathematics was published, that Russell
proposed a definition of the concept of number independent of the process of counting.
Another diary, written around 1890, contains allusions to Russell's
work on mathematics and mechanics. Russell mentions his studies on
conics and newtonian physics. He used E. J. Routh's (1831-1907) The
elementary part of the treatise of the dynamics of a system of rigid bodies
as his text in physics. It is reasonable to suppose Russell had an above
average education for his age. As a direct descendent of a British Prime
Minister, he would be expected to pursue a career in public service or
politics, perhaps having a major impact on England's future destiny.
Russell also received instruction in history, British literature, science and languages. At the same time, Russell became convinced of the
possibility of gradual social progress based on rational knowledge, perhaps, as a consequence of his study of the role of the Whigs in history.
Mathematics thought to be the best example of rational knowledge and,
most importantly, was the basis of any kind of complete rational knowledge. Russell perceived that Clifford supported such principles.
1. Here Euclid proves that "if two triangles have two sides equal to two sides respectively, and have
the angles contained by the equal straight lines equal, [ ... ] , the triangle will be equal to the triangle,
[ ... ]" [Heath 19S6a I, 247].
49
Shortly before he went to Cambridge in 1890, Russell read Mill's

autobiography for the first time; this book had a deep and lasting influence on him. 1 For a while, Russell considered Mill as his philosophical
god, before accepting the neo-Hegelian philosophy. Although Russell
admitted the profound influence of Mill's philosophy, he did not adopt all
of Mill's views. In particular, Russell was never interested in the positivism of Auguste Comte (1789-1857), even as a young man. Nevertheless,
Mill revived, as Clifford had donde before, Russell's old dream of finding knowledge that one could consider certain. Once again, Russell
thought that mathematics was more likely to be certain than any other
form of general knowledge. He also observed an idea shared by Clifford
and Mill: these two British thinkers had tried to formulate philosophical
systems that would contribute to a gradual social progress, individual
freedom and human happiness, which included the advancement of science. Unfortunately, Mill's ideas on mathematics contained in his autobiography are brief and scarce. The allusions to mathematical ideas are
concerned mainly with what he learned and was taught in his own childhood. But he does not pursue more sophisticated work relevant to the
principles of arithmetic and geometry. 2
3.3 Russell as a student at Cambridge.- In December 1889,

Russell presented his examination to enter Trinity College at Cambridge
University. Whitehead, who would become later on his professor and
colleague, was very impressed by Russell's mathematical background.
According to Russell's recollections, everything went well in Cambridge
from the moment he arrived in October 1890. There, he found the free
intellectual environment that he had missed for many years. The period
of extreme solitude and introverted thinking was over. Now, he would be
able to express his ideas and compare them with those of his classmates
and professors. This noisy and bothersome kid, according to the elders,
would become the rebel and passionate sceptic, admired and criticized by
1. He seems to have read it for a second time in September 1891.
2. According to his notebook What Shall I read? [see: Russell 1983a, 345], Russell read Mill's Three
Essays on Religion in April 1891, and Mill's Utilitarianism in July of that same year. According to
Russell [1956a, 116], he read Mill's A system of logic at age eighteen. In this book, Mill treated the
principles of mathematics in detail, emphasizing the foundations of arithmetic, algebra and geometry.
50
his colleagues. He remained convinced that the only way to acquire any
certain knowledge was through the study of mathematics. Philosophy also
fascinated him because he thought that it was perhaps the only possible
means to a solution of his religious doubts. The intellectual environment
at Cambridge dizzied him. He could express his wildest ideas without
any sign of rejection or prejudgement. Intellectually, he was respected for
the first time.
But the bureaucracy involved in obtaining the bacca laureale in
mathematics made him sick. The school attempted to foster a great degree of competition among students. This eagerness for competition was
so great that the members of the staff (in a calendar of the school) recommended that students who were antipathetic to competition not come to
Cambridge. Almost all activities at the college, including the acquiring of
scholarships, were determined by examinations.
According to the same calendar, the syllabus of mathematics during
the first year included such topics as: Euclid, algebra, trigonometry,
geometrical conics, plane coordinate geometry, differential and integral
calculus, elementary mechanics, equation theory and the first section (or
possibly the first three sections) of Newton's Principia. The topics to be
covered during the second year were: solid geometry, differential equations, Fourier's theorems, variational calculus, statics and attractions,
electricity and magnetism, hydrostatics, optics, dynamics of a particle
and parts of systems of dynamics. In his third year, Russell probably
studied: advanced material from the preceding subjects, formal and physical astronomy, hydrodynamics, elementary elliptic functions and Legendre and Bessel functions. 1
The academic staff perceived the examinations to be long and
tedious. Some of the courses may also have been tedious. The following
description of a typical lecture suggested that students spent "the whole
hour [of the lecture] in writing answers to written questions, or in solving
1. These are the topics covered by the new curriculum established in 1892 and it is not clear if the
'considerable changes' in the system of examinations included the contents of those courses. For the
origin and evolution of the Mathematical Tripos see, among others: Ball 1880a and Gascoigne 1984a.
51
problems" . 1 Russell was extremely dissatisfied with the teaching at the

school. The only exception was the classes taught by Whitehead. Whitehead taught him statics in 1890 and Russell's appraisal could not be
higher: "[Whitehead] was never repressive or sarcastic, or superior, or
any of the things that inferior teachers like to be."2
Russell thought that the mathematical demonstrations taught by his
professors were often fallacious. Shortly after he sat for the Mathematical
Tripos in June 1893, he decided to abandon mathematics. By terminating
his mathematical training in his third year, Russell closed the door to a
possible specialization in mathematics, according to the requirements of
the school. Some years later, Russell would complain that he never came
in contact with modern mathematics at Cambridge. He claims to have
never heard the name of Karl Weierstrass (1815-1897).3 But if Russell
selected this name just to make a point, his choice proved to be poor,
because Weierstrass's name was included in the reading list for the preparation of the final examination. Nevertheless, Russell's opinion of the
low standards and lack of a positive environment for mathematics at
Cambridge in the 1890s should not be questioned. Grace Young, who
studied at Cambridge around the same time, had exactly the same negative opinion. 4
In his fourth year at Cambridge, Russell modified his program by
studying philosophy. His election, in February 1892, to The Apostles,
composed by the intellectual elite of the university, had allowed him to
keep in close contact with those interested in philosophy. In his fourth
academic year (1893-1894), Russell studied: Ethics (Henry Sidgwick
(1838-1900)); Elements of Philosophy (Sidgwick); Metaphysics (James
Ward (1843-1925)); History of Philosophy (Ward); History of Philosophy
(George F. Stout (1860-1944)). The syllabi of some of the courses along
with the items listed in his small notebook What shall I read? indicate
that Russell read Mill, Bradley, Leibniz, David Hume (1711-1776),
1. Student's Guide 1888a, 74 & Student's Guide 1893a, 71.
2. Russell 1967a, 130; 1956a, 97.
3. See, for example, Russell 1926a, 242.
4. See: Grattan-Guinness 1976a, 158.
52
Kant, Benedict Spinoza (1632-1697), Descartes, Herbert Spencer (1820-1903), Darwin and Plato (428-348 B.C.). In some of his reminiscences
Russell claims that during this fourth year he came in contact with Cantor's "Mannigfaltigkeitslehre" ["General theory of invariants"] and Frege's Begriffsschrift [Conceptography] although he did not read them at
that time. 1
During his fourth year at Cambridge, Russell abandoned Mill's
ideas and adopted the British revival of the German Idealism known as
neo-Hegelianism. In most of his recollections, Russell described this
period of his life in succintly negative terms, leaving one with the impression that he was almost forced to accept neo-Hegelianism and, perhaps that it was a simple and superficial flirtation of an inmature adolescent philosopher. In most of his later recollections Russell narrated his
adoption of neo-Hegelian as an 'indoctrination', perhaps not in terms of
being impregnated with such opinions, but in the sense that he was forced
to embrace such principles. The unqualified adoption of Russell's own
views have influenced historians to stress that Russell's neo-Hegelian
period "was brief and soon repudiated". 2 Yet, Spadoni has shown that
Russell came to acept neo-Hegelianism only after serious consideration,
proving that he was deeply involved in this philosophy for almost six
years (ca. 1893 to 1898, approximately). Contrary to Russell's own
recollections, Spadoni noted that Russell's later rejection of neo-Hegelianism was not entirely satisfactory.
By checking Russell's reading book list during his fourth year at
Cambridge, one immediately gets the impression that Hegel's early influence was not direct, but ocurred only, possibly, through other books,
colleagues and professors. At the end of the nineteenth century, the ideas
of Bradley-and indirectly those of Thomas H. Green (1836-1882),
which emerged from Hegel's philosophy, dominated Cambridge philo so1. In a letter to Philip Jourdain dated September 11, 1917, Russell acknowledged having come across
Cantor's work when he reviewed a book by Arthur Hannequin (1856-1905) for Mind in 1896, and not
in 1894 as he claimed in his autobiography. In a earlier letter to Jourdain, Russell stated that he was
led to Frege by two independent sources: a review by Peano on Frege, and by Ward, who gave him an
old copy of the Begriffsschrift [see: Grattan-Guinness 1977a, 143-144 & 132].
2. Bowne 1966a, 20 note 1.
53
phy. By the end of 1893, when Russell became deeply involved with
these ideas, there were other British philosophers holding diverse principles of neo-Hegelian character: Bosanquet and Ellis McTaggart (1866-1925), among others. Although there are differences between Bradley
and Bosanquet, Russell considered both systems as essentially the same.
On the other hand, Russell thought that McTaggart's system focused on
the belief in immortality of souls or substances united by love.
There are several reasons that might account for Russell's accepting neo-Hegelian principles in his fourth year at Cambridge. First of all,
most of his philosophy professors accepted some form of the doctrine. A
second factor stemmed from Russell's desire to keep alive some form of
religious belief. Perhaps he thought that some of the inherent contradictions attached to the basic principles of religion could be resolved by a
satisfying synthesis. A third incentive was the failure of other systems to
explain the nature of mathematical propositions; in particular, Russell had
been greatly disillusioned with Mill's explanation. Finally, the immediate
cause of Russell's adoption was his acceptance of the ontological argument as a proof of the Absolute. 1 Among those immediately surrounding
Russell, it was McTaggart who played the most influential role in Russell's indoctrination; nevertheless, within a few months, Russell rejected
McTaggart's central metaphysics thesis.
It is worth mentioning that Russell also read some of Kant's works
in his fourth year at Cambridge, 2 and was greatly impressed by Kant's

discussion of the foundations of mathematics, in particular the principles
of arithmetic and geometry. A clear example of this influence is provided
by Russell's arguments in which he rejected Mill's conceptualization of
the principles of mathematics. Although, Russell was heavily influenced
by Mill for a certain time, he claims to have believed all of Mill's thinking on mathematics (not only those on arithmetic) to be inadequate. Nev-
1. A good general background reference to these arguments could be found in: Spadoni 1976a, 1977a
& 1978a.
2. According to What shall I read?, Russell studIed Kant's Prolegomena (Sept 1893) and the Grundlegung zur Metaphysik der Sitlen (Oct 1893) [Russell 1983a, 352].
54
ertheless, at age eighteen, Russell did not have a better answer for Mill's
empirical nature of mathematical propositions.
In his System of Logic, Mill proposed that induction was the only
valid form of inference; there were no exceptions, not even in mathematics. Contrary to the way in which Russell had learned elementary geometry from his brother Frank, deduction was never the source of new knowledge. Mill also rejected mathematical propositions as being necessary
truths; rather they were generalizations from experience, implying a
complete rejection of the a priori character of mathematics. In fact, Mill
sustained that geometrical axioms were not hypothetical but experiential
truths; that is to say, they were generalizations from events that one has
always experienced. He further held that no intuition was necessary. Not
only was geometry based on experience, Mill argued, but also arithmetic
and algebra. Russell could not accept Mill's empiricism. Even when
Russell had been enchanted with Mill's system, he was absolutely determined to reject the idea that simple additions of the form "two plus two
are four" express empirical generalizations. In particular, Mill asserted
that numbers were derived by abstraction from numerable objects. But,
according to Russell, numbers were neither contained in any object nor in
their totality. To count presupposed each one of the things as one. For
Russell, number was a free creation of one's thought. He wrote that
[numbers] arise from the gathering of a plurality into a unity. Until we apprehend a
plurality as a unity. we cannot apprehend it as a plurality. Successive acts of gathering of this sort give the series of natural numbers. Thus number is completely a
priori: it springs from a spontaneous activity of a purely formal nature. I
In this rejection of Mill's conception of number, Russell posited that both

number and counting presupposed each other. Here, Russell's argument
is resting on Kant's view of the synthetic a priori character of mathematical propositions; that is to say, that mathematics is neither empirical nor
definitional, but necessarily dependent on our cognitive processes. The
Kantian influence is even more obvious in his fellowship dissertation in
1. Russell 1983a. 129.
55
which Russell was concerned with the impact of non-Euclidean geometries on Kant's trascendental aesthetic-which studies the a priori forms
of sensation.
At the time (ca. 1893), Russell did not pay attention to the possible
points of antagonism between neo-Hegelianism and Kantianism; there is
no doubt, however, that neo-Hegelianism "overwhelmingly dominated
Russell's philosophical thinking." 1
There are several important philosophical and historical consequences of Russell's indoctrination into neo-Hegelianism and Kant's philosophy of mathematics. First, these doctrines did not allow the idea of
infinite number. Bosanquet, for example, argued that this idea was self-contradictory and unreal. Bradley also denied the idea because it contradicted a quality of the concept of 'actual number'. Russell, who was also
following Kant in matters relating to the principles of arithmetic, explained, that the series of natural numbers resulted from some sort of
successive acts of gathering. Under this Kantian premise, it would be
impossible to accept an infinite number, because it is not possible to
form, in/act, an infinite succession. Secondly, neo-Hegelians were strongly opposed to what they called 'equational logic', which developed into
algebraic logic or Boolean algebra. These concepts are outlined in the
works of William Hamilton (1788-1856), George Boole (1815-1864),
Augustus De Morgan (1806-1871), and W. S. Jevons (1835-1882),
among others. According to Bradley, equational logic had a tremendous
handicap: its mathematical nature. This characteristic limited logic "to
those problems which accommodate themselves to numerical reasoning.,,2 Bosanquet and Bradley conceived that equational logic dealt with
some subsidiary problems, and they both had considered that the goal of
logic (as a philosophical discipline) was to give a more precise description of reasoning in general. Probably, their criticisms had roots in their
incomplete appreciation of the nature of mathematics and equational logic
itself, as indicated by their view that mathematical developments were
irrelevant to metaphysics.
1. Spadoni 1977a, 105.
2. Bradley 1883a I, 370.
56
3.4 An Essay on the Foundations of Geometry. -
In the
meantime, Russell had madly fallen in love with a young American Quaker. Her name was Alys Pearsall Smith and she was five years older than
him. Her age, nationality and customs provided sufficient reasons for
Lady Russell to speculate that she was not a good candidate for her
grandson. Lady Russell vigorously attempted to discourage him from
marrying. The family even attempted to separate them by sending him to
France. But everything failed. Alys and Bertie got married on December
13, 1894. They spent together some time in Berlin, where they worked
on economic and social questions. Their effort directly resulted in the
publication of Russell's first book [German Social Democracy (1896a)],
but he "took no great interest in [the book], as [he] had determined to
devote [himself] to mathematical philosophy".1 After finishing his courses at Cambridge, Russell had not decided the topic of his thesis: mathematics or economics. In his detailed analysis of the evolution of Russell's
ideas from Imperialism to Free Trade, Rempel stressed Russell's deep
knowledge of economic theories, which would have allowed him to
become a distinguished economist, if he had wanted to do SO.2 Under
pressure from relatives and friends, Russell finally decided to write his
dissertation on the foundations of geometry.
It was in Berlin, in 1895, while taking a walk, that Russell decided
to write two series of books: one of these would go from mathematics to
physiology, the second would discuss social questions. Following a Hegelian method, he hoped to find a final synthesis, which would be both
scientific and practical. In Russell's view, each and every science contained contradictions, that is to say, each science was inherently defective. As he said in retrospective, he had "accepted the Hegelian view that
none of the sciences is quite true, since all depend upon some abstraction, and every abstraction leads, sooner or later to contradictions". 3
These defects or contradictions were to be solved dy dialectical transitions from one science to another. There were two possible means to
effect such transitions: either by the completion of an idea by its supple1. Russell 1967a, 130.
2. See: Rempel 1979a.
3. Russell 1959a, 42.
Plate 6. Alys Russell (December 1893).
57
58
ment, or by replacing the contradictory ideas by other noncontradictory

ones. It is important to obtain the least possible number of contradictions
that will have to be "dissolved by passing dialectically to another science".1 This is essentially the way in which Russell decided to write his
fellowship dissertation at Cambridge University on the foundations of
geometry.
The lost dissertation was, perhaps, originally entitled "The Metaphysical Basis of Geometry'? which was written during the spring and
summer of 1895, but it was not the form in which it was finally published. He rewrote his dissertation in 1896, the revision was published a
year later by Cambridge University Press, now under the title An essay
on the foundations of geometry. Unfortunately, it is not clear how Russell's ideas evolved from the original version of 1895 to the revision of
1897, once after receiving the criticisms, opinions and suggestions from
his professors, in particular, Whitehead. As it was mentioned above, the
original dissertation is now lost, although some mss is extant.
The book was addressed more to philosophers than to mathematicians. The main focus of the book was epistemology, and one recognizes
that he did not produce any original mathematical work. He studied the
mathematics related to the topic from a philosophical point of view. In
order to understand what he thought was his original contribution to
philosophy, it was necessary for the reader to be acquainted with Kant
and modern logic, by which he meant the works of Rudolf Hermann
Lotze (1817-1881), Bradley, Sidgwick and Bosanquet, and not those of
Boole, De Morgan, Jevons, Ernst Schroder (1841-1902), Peirce, Frege
and Peano, among others, which were unknown to Russell by 1897.
Russell's goal was to establish what geometrical knowledge was
logically required for a science of space, and that such knowledge was
logically necessary for the possibility of experience. The book was divided into five parts:
i) an introduction stating the problem;
1. Spadoni 1977a, 164. See, also, Russell 1959a, 52-53.
2. Russell 1983a, xxii.
59
ii) a chapter on the history of the rise and development of nonEuclidean geometries;
iii) a critical analysis of certain theories-those of Bernhard Riemann (1826-1866), Hermann von Helmholtz (1821-1894), among others-of the philosophy of geometry;
iv) an expository chapter describing Russell's own constructive
theory of geometry; and,
v) in the final chapter, a discussion of the philosophical consequences of Russell's own theory.
Russell wrote his dissertation under a neo-Hegelian influence: he
expected each science to be defective in itself, and those unsolved contradictions were to be dialectically corrected by other sciences. Taking these
principles into consideration, it is understandable why Russell would be
discussing contradictions as something natural and innate in each science.
As Russell pointed out in retrospect, neo-Hegelians were indifferent to
contradictions, because their systems were nourished by them.
In explaining the philosophical consequences of this new approach
for the foundations of geometry, Russell encountered contradictions that
he related to the "relativity, infinite divisibility, and unbounded extension
of space".1 The most prominent contradictions in geometry included,
first of all, the concept of a point. Because "a point must be spatial,
otherwise it would not fulfil the function of a spatial element; but again it
must contain no space, for any finite extension is capable of further
analysis" .2 A second contradiction resulted from an attempt to define
positions in space, because they are always relative: that is to say, "positions can only be defined by their relations", 3 but these relations are
defined by the positions they define. A third contradiction arose because
Russell regarded spatial figures as relations which are necessarily indivisible, while the figures themselves are necessarily divisible ad infinitum.
1. Russell 1897a, 177.
2. Ibid. 189.
3. Ibid., 188.
60
Although it may be interesting to understand and describe these

contradictions in detail, it is even more essential to grasp Russell's reaction to them. At this time, ca. 1895-1897, Russell thought that mathematics contained inescapable contradictions. That is to say, he considered
mathematics to be a defective science and therefore inherently imperfect.
Hence, the contradictions were inevitable, and therefore they did not
present a difficulty for the central argument of his book. To solve the
contradictions between the ideas of 'point' and 'empty space' it was
necessary to introduce the concept of 'matter'. The concept belonged not
to geometry but to kinematics, which leads one into the field of dynamics
and physics. Russell, then, expected to solve such contradictions by a
dialectical process, following a Hegelian method, in the field of a subsequent science. Russell even thought that he had solved all problems
related to the foundations of geometry. He had reasons to be proud. Most
reviews were very favorable, and Russell's old dream of being congratulated by an unacquainted foreign expert came true. He also had the opportunity for his first public debate with one of the leading mathematicians of all times: Poincare. Now, Russell believed, it was necessary to
solve all problems related to the physical sciences. 1
After finishing his An essay on the foundations of geometry, Russell "proceeded at once to what [he] intended as a similar treatment of
the foundations of physics".2 According to him, there were two main
problems closely connected with the principles of kinematics that needed
resolution. First there was the question of the ultimate nature of matter.
The second question was related to the problem of absolute versus relative motion. Russell desired to write a book on the philosophy of physics
again following a neo-Hegelian method. The general structure of his
book was going to be similar to his previous essay on the foundations of
1. Cf. Wood 1959a, 259. Wood claims that the natural way to study Russell's The Principles 0/
Mathenuuics is in the reverse order of its presentation; that is to say, he suggests that Russell first
developed his ideas on the principles of dynamics, then those of geometry, and, finally, his views on
the concept of number. This is true for those ideas contained in The Principles o/Mathematics. But one
should conclude from Russell's recently published manuscripts of 1896 and 1897 relating to the
physical sciences that his general intellectual path went from dynamics to geometry. One should not
forget that Russell wrote his essay on the foundations of geometry [1897a] before he wrote his papers
on the foundations of physics.
2. Russell 1959a, 41.
61
geometry. First, he would describe a critical history of dynamics. Then,

in a second part, he would elaborate a constructive theory of the same
field.
But although Russell worked on the foundations of physics for two
years (ca. 1896-1898), he began to comprehend that dynamics was not an
a priori science, but that its questions were mainly empirical. His later
rejection of neo-Hegelianism was so extreme that he subsequently
claimed that all that he had wrote on the foundations of physics during
this period was "complete nonsense", I and that it was difficult for him
to comprehend how he came to accept such ideas in the past.
Published [see Russell 1897c and 1899c] and unpublished material
[see Russell 1897h, 1897i and 1897j] kept at The Bertrand Russell Archives shows that he was also working on the foundations of mathematics
at the time. Subsequently, he decided to turn his attention exclusively to
questions related to the foundations of arithmetic, a topic that he began to
pursue mainly as a result of his first trip to the United States in 1896.
3.5 Russell's early attempts to write a book on the principles of arithmetic.- In the last months of 1896, Russell travelled to
the United States. According to his later reminiscences, it was there
where he became acquainted with German mathematics. He soon adopted
the view that German scholarship was superior to British scholarship, not
only in mathematics but in almost all academic matters. Back in England,
Russell was asked to review a book by Hannequin entitled L 'hypothese
des atomes (1895). Here, Russell found, perhaps for the first time, references to Cantor's theory of transfinite numbers. The book was very
critical of the German mathematician, and Russell thought the criticism
was justified, as he was still a follower of Kant. Hannequin had argued
that Cantor's attempt to treat number as a continuous quantity gave rise
1. Russell 1959a, 43.
62
to Kant's second antinomy} In the review, it was not appropiate for

Russell to discuss his own understanding and criticism of Cantor's principles. But, at least, he publicly stated that he could not accept Cantor's
second number class.
For Cantor's second number class of numbers [w, w+l, w+2, ... ], by which
[Cantor] hopes to exhaust Continua, begins with the first number [w] larger than any
of the first class; but as the first class (the ordinary natural numbers) has no upper
limit, it is hard to see how the second class is ever to begin. 2
Through the references of Hannequin's book, Russell learned of

the French translations of Cantor's work published in Acta Mathematica.
Russell read Cantor's articles for the first time in March 1896 at his
in-laws' residence, while he was lecturing at the London School of Economics. Russell even wrote an essay [see: Russell 1896d] in which he
discussed and analyzed his reasons for rejecting Cantor's results.
In this manuscript, Russell claimed that the philosophical antinomies had their counterparts in mathematical fallacies, like those of the
infinitesimal calculus and others associated with Cantor's collections.
Russell denied the validity of the concept of mathematical infinity, as
expected from a neo-Hegelian, because infinity would have to be "a
quantity larger than any assignable quantity".3 There was no doubt that
Cantor's methods were ingenious, but Russell questioned their philosophical validity. ,If the succession of natural numbers
1, 2, 3, ... , n, ...
was unlimited, how then it was possible for someone to talk of the first
number [w] following all these numbers? There was not even a chance
1. In Kant's system, an antinomy is the contradiction resulting from the inferences correctly drawn
from two equally valid principles (usually a thesis and an antithesis). In particular, Kant's second
antinomy read as follows: "Thesis: Every composite substance in the world is made up of simple parts,
and nothing anywhere exists save the simpler or what is composed of the simple. Antithesis: No
composite thing in the world is made up of simple parts, and there nowhere exists in the world
anything simple" [Kant 1787a, A434-435, B462-463].
2. Russell 1896a, 412.
3. Russell 199Oa, SO.
63
for 'w' to ever arise. Russell's conclusion was simple and direct: "even
the humblest philosopher [ ... ] must grow indignant at Cantor's 'W"'.1
Russell's main argument rested on the point that if it was said that the
natural numbers had no limit, then it could not be possible to demand, in
the next step, that one had a number larger than any natural number. On
the basis of these views, his conclusion was evident: "Cantor's transfinite
numbers, then, are impossible and self-contradictory.,,2 To categorize
them otherwise would mean "to admit the truth of hopeless and irresoluble contradictions."3 Some contradictions may be helpful; but others,
like those generated by Cantor's ideas, were insolvable.
To conclude, Russell's first encounter with the theory of transfinite
numbers gave him a negative impression of Cantor's ideas. His reading
of the French translations, instead of the original German articles, may
have contributed to Russell's early rejection of Cantor's ideas. Cantor
had published an article [Cantor 1932a, 165-208] discussing in detail the
distinction between the potential infinite and the actual infinite. This
distinction was fundamental in comprehending how the principles of
generation of transfinite numbers were introduced. One of these principles, the second principle of generation, allowed Cantor to introduce a
number starting a new sequence of numbers, even though the previous
sequence had no upper limit. In the French translation [Cantor 1883i],
however, Cantor approved only the inclusion of those sections directly
concerned with mathematics. All sections dedicated to the philosophical
discussion of the validity of the actual infinite and the transfinite numbers
were omitted. It was, perhaps, this lack of philosophical discussion that
made Russell 'indignant' towards Cantor's theory. Thus Russell rejected
the theory completely, including its mathematical aspects. According to
Couturat, Russell had fallen into two of the most natural mistakes made
by philosophers when discussing the infinite. They talked of an infinite
number-a number of numbers, and they thought of all infinities (if there
were any) as being equal.
1. Russell 1990a, 52.
2. Ibid.
3. Ibid., 53.
64
It was not long before Russell came in contact again with Cantor's
work, in spite of his first negative reaction. He was asked to review an
article for the British philosophical journal, Mind. This time Russell
inspected Couturat's De l'injini mathematique (1896). Unlike Hannequin,
Couturat had a very positive and sympathetic reaction to the treatment of
Cantor's theory of transfinite numbers. Couturat stated that, Cantor's
exposition of the theory of the infinite was completely free from contradictions. In fact, one of Couturat's main goals was the defence of the
concept of actual infinite. Although Couturat was unable to influence
Russell's negative opinion of Cantor's ideas, he was, nevertheless, completely convinced of the importance of the question of the relation between the categories of number and quantity to the foundations of mathematics. In his review, Russell chose three of the existing theories concerned with the relation between number and quantity. These three theories, which were radically different, were developed by Mill, Hannequin
and Couturat. The first two, as I haVe( already mentioned, rejected the
central concept of Cantor's transfinite numbers: the supposition of the
existence of the whole sequence of natural numbers as a given totality,
that is to say, as an infinite set. On the contrary, Couturat argued that
infinite quantity was a priori without any connection with infinite number. Couturat also supported number and quantity as completely independent categories, contrary to the assertions of the proponents of the philosophical systems underlying Russell's philosophy, such as Hegel and
Kant.
Nevertheless, by late 1897, Russell showed the first signs of an

intellectual shift vis-a-vis neo-Hegelianism. Spadoni has already clearly
shown that Russell did not deviate suddenly from the intellectual position
of his mentors. On the one hand, in 1897, Whitehead asked Russell to
read Hegel's books on logic, and not those of other neo-Hegelians; as a
result, Russell was now convinced that all that Hegel had said concerning
the foundations of mathematics was "ignorant and stupid". 1 I suspect
that Russell reacted in the way that Whitehead expected. Whitehead also
read Hegel once, arriving at the same conclusion. Perhaps, Russell no1. Russell 1956., 24.
65
ticed that Hegel made mistakes similar to those of many philosophers and
mathematicians in the eighteenth century: there was no correct definition
of continuity and, therefore, no way to avoid the paradoxes of infinite
number that Hegel had thought he had solved. Also there were also
several fallacious proofs in the infinitesimal calculus. But although Russell disagreed strongly with Hegel's views on the foundations of mathematics, he did not abandon neo-Hegelianism completely.
Sources of early 1898 [see: Russell 1898c] show that Russell was
still working on the 'dialectic of the sciences' (metaphysics), i.e. the final
synthesis in which all contradictions of all the sciences will have been
overcome. In early 1898, Russell also thought of mathematics as a dialectical science; notwithstanding the fact that he had started to consider that
mathematics could be completely true in and of itself, and therefore, "not
merely a stage in dialectics".1 In March of 1898, Russell was still working on the foundations of physics. He was also reading Dedekind's Was
sind und was sol/en die zahlen? [l888a] and Whitehead's Universal
Algebra [1898a] as background material for a new book on the foundations of arithmetic.
A few months later, in July of 1898, one finds Russell working on
a book on the principles of mathematics, which had been his "chief
ambition ever since age of eleven". 2 The complete title of this obscure
and difficult manuscript is "An Analysis of Mathematical Reasoning:
being an inquiry into the subject-matter, the fundamental conceptions,
and the necessary postulates of mathematics" [Russell 1898e]. The general structure of the book was divided in five parts: an introduction; book
I, on the manifold; book II, on number; book III, on quantity; and, book
IV, on the infinitesimal calculus [see: Table I].
Some signs of neo-Hegelian influence are found in the manuscript
because Russell insisted in keeping the final status of mathematics contradictory; in other words, mathematics was still a defective science. There
1. Russell 1944a, 12.
2. Russell 1929a, 15.
66

Table I. An analysis of mathematical reasoning.
Introduction. This is an outline of the contents of the extant manuscript version of

Russell's" An analysis of mathematical reasoning"-there is a second typed version too segmented. In fact, this description is based on both the Table of Contents contained in the first folio of
the manuscript and the extant folios. This essay was written approximately in July 1898, although Russell's correspondence with G. E. Moore seems to indicate that the manuscript was not
entirely completed by this date. Nevertheless, there are some folios of subsequent sections
contained in later manuscripts which suggest that Russell continued to work on this draft. Today,
only Book I (contained in File 230-03300-Fl) remains complete. Books II, III and IV (kept in
File 230-03300-F2) are incomplete, but some of their folios might be found in the version of the
"Principles of Mathematics" completed in 1899-1900. The topic: "The Unit and the Number
One" (Book II, Chapter III) was listed in the table of contents but was not developed in the
manuscript. (Delentions are represented by square brackets).
An Analysis of Mathematical Reasoning
being an inquiry into
the subject-matter, the fundamental conceptions &
the necessary postulates of mathematics.
Introduction
The Manifold [fhe Logical Calculus]
Book I
I The elements of judgments [On the conception of a manifold]
Chapter
II Subject and predicate [On addition]
III The manifold [Correlation of the two kinds of addition]
IV Addition and synthesis [Interpretation of the logical calculus]
V The logical calculus
Number
Book II
I Existential judgments and number [fhe cardinal integers]
Chapter
II Necessary connections between cardinal integers [Numbers]
III The unit and the number one
III Ratio
IV The relativity of the unit and the number continuum
[Defmition of a particular [relativity of the] unit]
V Ordinal Numbers
Quantity
Book III
I Transtition from number to quantity
Chapter
II Equality and inequality
III Conditions for the predication and quantity
IV On the distinction of sign
V Extensive quantity
The Infmitesimal Calculus
Book IV
I Zero
Chapter
II Infinity
III Approximation
IV On elements in an extensive continuum
V The differential coefficient
67
was a contradiction (the contradiction of relativity, described above)

inherent in mathematics. The only reference to this contradiction in the
whole text of the surviving folios reads as follows: "The contradiction of
the difference between two terms, without a difference in the conceptions
applicable to them". 1 This contradiction, along with the concepts of
'addition' and 'manifold', seemed to define the realm of mathematics.
Other signs of neo-Hegelian influence lay in Russell's design of a
triadic structure of categories. Russell argued that each mathematical
concept presupposes the fundamental ideas of 'identity', 'diversity' and
'unity'.2 Another sign of the remaining influence of the neo-Hegelians on
Russell rests on the fact that Russell divided judgements into two categories: those of intention and those of inclusion, thinking, perhaps, that
somehow one was the antithesis of the other.
But there were other influences on Russell. Contrary to Bowne, I
cannot assert that Russell was completely "unimpressed by [the] philosophical implications of [Whitehead's Universal Algebra]".3 We agree
with Bowne that there were others who were more impressed than Russell by the philosophical implications of Whitehead's book. Couturat, for
example, concluded that the algebra of logic appeared to be the foundation of mathematics or, at least, the most fundamental branch of mathematics, prior to the ideas of 'number', 'order' and 'magnitude'. In a
much later recollection, Russell commented that he was "greatly excited"
with Whitehead's Universal Algebra. 4
The reader should not be guided solely by the printed records. It is
true that Russell did not appear, at least in public, to be greatly influenced by his former instructor's work. Nevertheless, there are at least
two very clear hints of Whitehead's impact on Russell which were reflected in this unsuccessful attempt to write a book on the principles of
mathematics in 1898. The first indication is Russell's discussion of the
1. Russell 1898a, Introduction, Folio 6 [Russell 1990a, 106].
2. See, also: Rodriguez 1988a, 28.

3. Bowne 1966a, 48.
4. Russell 1959a, 39.
68
concept of 'manifold' as one of the basic ideas of mathematics; a second,

and more obvious sign was Russell's effort to base the explanation of the
basic concepts of arithmetic upon those of the logical calculus. This
seems to be the foundation of Russell's logicist program! It would be
natural, then, to ask why did he not develop logicism as early as 1898,
instead of late 1900 or early 1901? There are some plausible explanations. One is that this logical calculus was not 'the true' logical calculus. 1 On the other hand, by early 1898, Russell was still under the influence of neo-Hegelian doctrines. His conception of logic was still 'too
philosophical', and he was also skeptical of the value of algebraic logic.
Finally, he speculated, contrary to the order of presentation of his book,
that arithmetic was more general than the logical calculus. Even by 1900,
when Russell delivered a paper at the First International Congress of
Philosophy, he considered logic a more philosophical vain, arguing that it
treated questions concerned with reasoning in general.
Unfortunately, Russell avoided a detailed analysis of the concept of
'number'. He argued that nothing could be said about its nature. He took
the nature of 'number', when a particular number was going to be applied to a particular manifold, to be something known (i.e., a priori).
But, what about infinite numbers? Russell argued that there was a chance
for infinite numbers to arise as reflected, for example, in the extension of
number itself. But he was still contemplating the idea of infinite number
as related to the number of numbers; probably he was unable to explain
those philosophical antinomies already linked with the misrepresentation
of the idea of an actual infinite set. Russell promised his future readers
that he would deal with the topic later on. Due to the incompleteness of
this manuscript "An Analysis of Mathematical Reasoning", it is impossible to judge whether Russell had an effective means to deal with those
philosophical antinomies related to the number of numbers. According to
a letter to Moore, however, Russell could not find a solution at that
1. We should not forget that the logical calculus of Whitehead's Universal Algebra was still basically
Bollean algebra. Whitehead's goal was "to present a thorough investigation of the various systems of
Symbolic Reasoning allied to ordinary Algebra" [Whitehead 1898a, v); but he was not attempting to
express mathematics from such principles. Whitehead's Universal Algebra belongs to an algebraic
tradition concerned with the use of parts of mathemstics and versions of the calculus of predicates.
69
moment; he simply "skated over [the antinomies ... ] leaving them to be

discussed later". 1
According to a retrospective note added to the manuscript, perhaps
much later, Russell finished this book in July 1898. At once, Russell
requested Moore to read it. Moore, who for some years had "fulfilled
[Russell's] ideal of genius" ,2 had recently overame the influence of the
neo-Hegelians. Apparently, Moore only read the first book (chapter) of
the manuscript and then communicated his criticism to Russell orally.
This review, apparently, was extremely critical, prompting Russell to
finally abandon neo-Hegelianism. He now thought that there was no need
to discuss mathematics merely in terms of a stage in dialectics, among
several disciplines. If the nature of mathematics could not be explained
by the principles of neo-Hegelianism in a reasonable way, it had to be
discarded. Russell's rejection of neo-Hegelianism became so extreme that
now he was tempted to oppose all neo-Hegelian philosophy, not just neoHegelian philosophy of mathematics. Thus he thought that the concepts
of 'number', 'space', 'time', and 'matter' no longer had a self-contradictory character; rather they existed as entities in themselves. Russell's
conception of the nature of numbers shifted completely. Now he
imagined all the numbers sitting in a row in a Platonic heaven [... and] thought that
points of space and instants of time were actually existing entities, and that matter
might very well be composed of actual elements such as physics found convenient. 3
There was yet another element that may have contributed to Russell's final rejection of neo-Hegelianism: his study of Leibniz's philosophy, in particular, his logic. Why did Russell become interested in Leibniz's logic? Was there any genuine interest of Russell in Leibniz's philosophy per se? Russell's answer, like many of his comments on his own
work, is simple and direct. "My book [on Leibniz] is the result of a
purely philosophical course of lectures which I was obligated to give at
Trinity College. ,,4
1. Letter: Russell to Moore, 20 July 1898.
2. Russell 1967a, 64.

3. Russell 1959a, 48.
4. Letter: Russell to Couturat, June 21, 1900, my emphasis.
70
Russell began to prepare his lectures on Leibniz in early 1898}

His study of Leibniz helped him to apply his new views on logic, which
he developed after receiving Moore's criticism. It also convinced Russell
of the great importance of the 'theory of relations'.
The title of Russell's book was A Critical Exposition of the Philosophy of Leibniz, with an appendix of leading passages, in which Russell
attempted to follow an epistematic approach in analyzing the truthfulness
of Leibniz's system. By an epistematic approach, Russell meant that his
goal was not to carry out an historical study of Leibniz's work, but to
reconstruct the logical structure of the doctrine. Russell's main thesis was
to prove that Leibniz's metaphysics was grounded on his logic, but at the
same time, that there were some inconsistencies. Some inconsistencies
resulted from Leibniz's reluctanse to discuss issues that logically arouse
from this thinking, but were considered objectionable in his day. Others
arouse from his first five original premises.
I will bypass commenting on Russell's analysis of Leibniz's philosophy, because it has already been done very throughly elsewhere. 2 However, I shall examine the way in which Leibniz's concepts affected the
development of the 'paradoxes' and Russell's Logicism. In the first place,
Leibniz's position regarding the acceptance or rejection of an actual
infinite is ambiguous. In some places Leibniz spoke positively of the
actual infinite, and in some other places he rejected it on grounds of its
inconsistency. 3 But, according to Russell, one of the basic premises of
Leibniz's philosophy was to assert that there was 'an infinite number of
monads'-the simplest substance. At the same time, Russell's analysis
showed that Leibniz also repudiated the idea of 'an infinite number'. So,
one could not conclude that Russell thought of the concept of the infinite
in a specific manner because of Leibniz. Despite indirect evidence, one
1. According to O'Briant, "the evidence, [... ], indicates that Russell was preparing these lectures
during the latter balf of 1898 and first quarter of 1899" [O'Briant 1979a, 166].
2. See: O'Briant 1979a.

3. "I am so much in favor of the actual infinite, that [... ] 1 hold that nature affects it everywhere"
[Russell 19OOa, 109]. Cf. Ibid. P 110.
71
does not get a detailed picture of Russell's own ideas on logic at the time
because the Leibniz book was written as a didactic work. In particular,
there is nothing in the text indicating whether Russell still rejected the
concept of the actual infinite. A footnote in the text, however, seems to
indicate that Russell allowed, at least, a possible solution to the antinomy
of a given infinite number.
How did Russell react towards Leibniz' s idea of a Characteristica
Universalis-or universal language? According to Leibniz, this characteristica universalis would allow the expression of all knowledge in
terms of very simple characters that were easy to learn, comprehend and
use. Simple concepts would be expressed with these simple characters;
more complex ideas would be expressed using combinations of these
characters. Russell thought that Leibniz was seeking something similar in
nature to symbolic logic, "which [was] definitely a branch of mathematics".l Russell thought that symbolic logic could be useful in mathematics, but that it was not suitable as a method of philosophy. Russell's
conclusion was opposed to the opinion of Peano and his colleagues. Some
years later, once Russell became deeply influenced by Peano and his
followers, Russell would modify his position towards the value of Leibniz's work.
The correspondence between Russell and Moore reveals that by
May 1899 Russell was working again on the principles of arithmetic, and
that he was mired in difficulties. At The Bertrand Russell Archives there
is a manuscript entitled "The Fundamental Ideas & Axioms of Mathematics" [see: Table II.A & II.B]. There are some differences in the general
structure of this manuscript compared with Russell's earlier manuscript
entitled "An Analysis of Mathematical Reasoning". First of all, in this
work there was no logical precursor of the idea of 'number'. The treatment of symbolic logic had been displaced to a brief section of chapter II
dedicated to the application of numbers. More important, Russell displayed, for the first time, the explicit influence of the Weierstrassian
school of mathematics. Now, Russell was convinced that Weierstrass and
his followers had shown that "all pure mathematics should be regarded as
1. Russell 1900a, 170.
72

Table II.A The Fundamental Ideas & Axioms of Mathematics.
Introduction. This outline of Russell's manuscript "The Fundamental Ideas & Axioms of
Mathematics" is taken from its Table of Contents and the surviving folios. This manuscript was
written during the first balf of 1899. Little remains of Russell's second attempt at his 'big book'
on the principles of mathematics. The remaining sections of Part I, Part IV and Part VI are kept
in File 230-030-310. Some other missing folios are now contsined in the final draft of The
Principles of Malhemalics, the one Russell claims to have finished on May 23, 1902.
The Fundamental Ideas and Axioms of Mathematics
Table of Contents
Part I
Chapter
Part II
Chapter
Number
I Pure number
II Application of numbers
III Analysis of addition
IV Number in connection with whole & part, quantity & order
Whole & Part
I Implication of whole & part in application & addtion of numbers
II Divisibility & plurality
III Whole & part as a relation
IV Whole & part in relation with classes
V Whole & part without definite numbers
VI Fractions
VII Antinomy of infinite number
Part III
Chapter
I
II
III
IV
V
Order
Mesning of order
The range of order
Sign
Can cyclic orders be one dimesnional'?
Orden & position
I
II
III
IV
Quantity
Mesning of magnitude, greater & less
The range of quantity
Numbers as expressing quantity
Continuity, zero & infinity
Part IV
Chapter
(continues .. )

Table II.B The Fundamental Ideas & Axioms of Mathematics.
Part V
Chapter
Part VI
Chapter
Part VII
Chapter
Part VIII
Chapter
I
II
III
IV
V
Extensive Quantity
Meaning of extensive quantity
Zero & infinity
Approximation
Points
The differential coeficient and the integral
I
II
III
IV
Space & Time

Points & instants exist
Relations & projective geometry
Dimensions
No antinomies peculiar to space & time
I
II
III
IV
V
Matter & Motion

There are no things
Motion
Motion is not relative
Impossibility of explaining away quantities
Mass as quantity of materiality
I
II
III
IV
Motion & Causality

Laws of motion must be casual
Mass & Newton's laws
Antinomy of causality
Constancy of sums
73
74
dealing exclusively with numbers". 1 Russell also pointed out that this
development, the arithmetization of analysis implied that there was an
even greater separation between the concepts of 'number' and 'quantity',
as did the discovery of new algebras dealing also exclusively with numbers. Perhaps, Russell understood this distinction after reading Whitehead's Universal Algebra or Dedekind's Was sind und wass sollen die
zahlen?
Russell remained concerned with the philosophical antinomies
relating to the concept of infinite number. There are three different ways
of arriving at these philosophical antinomies: through the idea of all
numbers, the idea of the largest of all numbers, and the idea of the last
possible number. The antinomies resulting from these three ideas could
be formulated in such a way that their existence could be both proven
and disproven and, according to Russell, no existing metaphysics avoided
these antinomies. He promised his readers that he would discuss such
antinomies in part II, which is now lost. It is important to keep Russell's
new term 'antinomy' in mind when refering to old problems associated
with the idea of infinite number. Under the previous neo-Hegelian philosophical influence, these propositions were described as 'contradictions',
and the concept 'infinite number', at least when it was discussed in relation to Cantorian transfinite numbers, as 'self-contradictory'. But now,
Russell is using the term 'antinomy', most likely following Kant directly.
Finally, the table of contents of this second manuscript shows that there
was some doubt in Russell's mind about the role of symbolic logic in
mathematics. The analysis of the application and addition of numbers
supposedly explained how the concepts of 'whole' and 'part' originated,
whereas he thought that possibly the role of symbolic logic, as the calculus of 'whole' and 'part', was logically prior to arithmetic.
But Russell abandoned this enterprise almost immediately. By July
1899, he was working on yet another draft on the principles of arithmetic, only two months after his previous unsuccessful attempt. This time
he entitled the manuscript "Principles of Mathematics". The key to ex-
1. Russell 1899b Part II Chapter I Folios 1-2. (In the table of contents, the discussion of the concept of
'Quantity' appears as Part IV, but in the manuscript occupies Part II).
75
plaining why Russell did not simply keep working on the fundamental
ideas contained in the previous manuscript, rather than beginning a new
one, seems to be the influence of Cantor's work. According to Russell's
small notebook What shall I read?, he read Cantor again in July 1899.
Most important, in a letter to Moore, Russell stated that he was rewriting
his 'big book' and that he had found the topics of 'order' and 'series' (sequences) both important and fruitful, even though they had been "almost
entirely neglected by philosophers". 1 These two concepts had recently
acquired a special status in the study of the foundations of mathematics,
especially, in relation to the principles of number. Cantor's and Dedekind's works, among others, had shown that their analysis could bring
light to the characteristics that differentiated the diversity of numbers
(e.g., natural, rational, irrational, real, among others).
The impact of Cantor's work on Russell is easily found throughout
the manuscript, not only because of the explicit references to Cantor's
name, but also because of Russell adopted Cantor's conceptualization
(e.g., definition of denumerability), and his use of some of Cantor's
results. On this occasion, the manuscript seems to have been planned in
seven parts: 'number', 'whole and part', 'quantity' [now missing], 'order', 'continuity and infinity', 'space and time' and, finally, 'matter and
motion' [see: Table III]. From the general structural point of view, there
do not seem to be great changes in the organization of the manuscript
over the previous one. The same topics were discussed in both manuscripts, although there were some changes in their order of presentation.
In particular, the parts on 'Quantity' and 'Order' were interchanged. But
this new sequence of presentation does not imply fundamental changes in
Russell's understanding of the concept of 'number'. It simply stresses,
once more, the profound distinction between the categories of 'Number'
and 'Quantity'.
Once again the first point to be discussed was the concept of Number. Two major topics had to be explained: the relations of numbers inter
se, and the concept of collections or aggregates. Both terms, 'number'
1. Letter: Russell to Moore, July 10, 1899.
76
and 'collection' were fundamental and indefmable, with the latter arising
from a certain way of counting elements. These two concepts were also
independent of each other, with neither of them presupposing the other
one. They both had "equal logical priority".1 The only reason why Russell began by explaining the nature of the term 'collection' instead of
'number' was because, in his view, the former required a smaller effort
of abstraction. The concept of 'counting' could not be regarded now as a
foundation of arithmetic. According to Russell, the number of a collection resulted from counting: this process always presupposed certain
characteristics of the nature of collections and the pure relations of integers. There was a small number of indefinable terms: 'points', 'instants',
'numbers'. Collections were formed from the addition of units, and then
numbers could be derived from these collections. This process of addition
was important because it still reflected the Kantian influence on Russell.
For example, Russell said that the equation '7 + 5 = 12' was "certainly
a synthetic proposition".2 The reason was that the meaning of '12' was
not defined by the equation but it was presupposed by it. It was reasonable to suppose that numbers were undefinable because they could not be
defined by means of additions or by collections. In order to arrange
numbers in a sequence it was necessary to find a relation capable of
carrying out such a task without presupposing the concept of 'addition'.
Russell thought he had found such a relation in terms of 'order'. This
once again reflects the impact of the Weierstrassian school and Weierstrass's program of arithmetization of analysis. This program required the
rigorization of the real number system, and from it, the rigorization of all
its operations, including addition. In mid nineteenth century some mathematicians thought it necessary to construct certain branches of mathematics more rigorously. Some of these reasons had to do with the development of non-Euclidean geometries and abstract algebras. It was natural to
attempt to build mathematics upon the concept of 'number', without
resorting to personal intuition and helpful diagrams. For Russell it became a fundamental requirement to presuppose the concept of 'number'
for defining the addition of numbers.
1. Russell 1899h Part I Chapter I Folio 1.
2. Ibid. Chapter III Folio 19.
77
Table III. Principles of Mathematics.

Introduction.- This is an outline of the Table of Contents of Russell's "Principles of Mathematics" developed from the surviving folios. The manuscript was written ca. July 18, 1899 (according to a
letter to Moore) and ca. June 1900 (according to a dated folio of Part VI). There is a note, at the left
corner of folio I, indicating that Russell revised this section after meeting Peano, in approximately September or October of 1900. Some of the surviving folios of this manuscript had been included in the typed
draft of "An analysis of mathematical reasoning". Part VII, on Matter & Motion, is missing in its entirety.
Today, it is contained in the fmal version of The Principles of mathematics.
Principles 0 f Mathematics
Number
Collections
Pure number & their mutual relations
Arithmetical addition
Multiplication
Rational fractions
Infinite colections
Whole & Part
Meaning of whole & part
Totality
Whole & part without the use of numbers. The logical calculus.
Infmite wholes
Ratio in connection with whole & part
Order
Meaning of order
On the distinction of sign
On asymmetrical relations
Various kinds of series
Ordinal numbers
Various questions concerning series
Continuity & Infmity
On the correlation of series
Irrational numbers
Degrees of continuity
Zero & infmity
Transfmite numbers
Approximation
The differential coefficient
The defmite integral
Infmite divisibility & quantitative infmity
Space & Time
Points & instants
Simple relations between instants
Spatial distance & the straight line
Dimensions
78
There are other signs indicating Cantor's new influence on Russell.

Russell now believed in the existence of 'actual infinite collections',
Cantor's basic concept for developing the arithmetic of transfinite cardinal and ordinal numbers. Now, Russell asserted that natural numbers
were infinite in number. This basic idea, along with the introduction of
the concept of 'order', would allow Russell to develop his ideas on 'sequences'. Nevertheless, Russell was still very critical towards the ideas
he adopted from Cantor. As usual, he did not assimilate Cantor's ideas
blindly. In the section of the book dedicated to the concept of 'number'
there are signs that Russell was beginning to reject some of Cantor's
conclusions. In the first folio of the manuscript, Russell had claimed that
'collections' implied 'numbers'. However, he later restricted the indiscriminate use of this principle. There was always a question, he argued,
whether or not any given collection has a number, because it could contain an infinite number of elements. Having no doubt that there were
infinite collections, now Russell even attempted to "free them from the
contradictions which cling to them." 1 But what about infinite numbers?
Russell could only answer the question in one direction: at this time "no
assignable number is the number of all numbers.,,2 Notwithstanding the
difficulties of Cantor's system, Russell was greatly impressed by Cantor's
'mathematically invaluable analysis of continuity'; he was also deeply
interested in its philosophical implications. He was convinced that Cantor's ideas were fundamental for a sound explanation of the principles of
calculus and the theory of irrationals, but, he argued, "I cannot persuade
myself that [Cantor's] theory solves any of the philosophical difficulties
of infinity, or renders the antinomy of infinite number one whit less
formidable" .3
It can be seen from Russell's manuscript that he was skeptical of
Cantor's second principle of generation, which allowed Cantor to introduce '",'-the first transfinite number. This number was not definable
from Cantor's first 'principle of generation' which created numbers only
1. Russell 1899h Part I Chapter VI Folio 45.
2. Ibid.
3. Ibid. Part V Chapter V Folio 40.
79
by adding units to the ones already known. Russell believed that the
generation of number Ow' violated the usual principles of arithmetic.
According to him, if the number previous to Ow' was a finite one, so was
'w'. If Ow' was the number of all finite numbers, it would have to be the
last finite one. Hence, Russell concluded, in opposition to Cantor, that
'w' was not a "definite [and] unique number." I Russell was also tempted
to drop the question of whether there was a number of numbers, because
he thought a positive answer to this question led to contradictions. Russell had asserted earlier in this manuscript that there were infinite collections (e.g., the collection of numbers), but that such collections did not
have a number. Russell, nevertheless, thought that Cantor's theory had
been useful in shedding light on the difficulties surrounding the infinite,
which was "for philosophy, [the transfinite numbers] principal and invaluable merit."2 In spite of its implications for the mathematical theory
of continuity and infinity, Russell claimed that philosophically "Cantor's
infinite is as ambiguous & as little definite as the older kinds, & cannot
disprove the proposition that every number is finite.,,3
Concluding, there were important changes between this manuscript,
probably finished by June 1900, a couple of weeks before Russell attended the First International Congress of Philosophy in July, and the manuscripts of 1896, 1898 and 1899. In 1896, Russell had even refused to
discuss the possible mathematical uses of Cantor's theory of transfinite
numbers. By 1900, on the contrary, although he was still unclear about
the philosophical foundation of such a theory, he now applied Cantor's
ideas to explain the complex nature of certain basic mathematical concepts, such as those of 'order', 'sequences', 'continuity' and 'infinity'.
1. Russell 1899h Part V Chapter V Folio 48.

2. Ibid., Folio 53.
3. Russell 1899h Part I Chapter IX Folio 86.
I thought [the writing of The Principles of

Mathematicsl was nearly finished, but in the month of May I
had an intellectual set-back [...J. Atfirst I supposed that I
should be able to overcome the contradiction quite easily,
and that probably there was some trivial error in the reasoning.
Gradually, however, it became clear that this was not the case.
Bertrand Russell.
Chapter IV.
Russell's discovery
of the
'paradoxes'
4.1 Introduction.- The main goal of this fourth chapter is to
present a plausible historical reconstruction of the actual way in which
Russell discovered the 'contradiction' of the greatest cardinal number,
known today as the "Cantor paradox"; how Russell discovered his own
inconsistency of the class of all classes which are not members of themselves; and how he presented the elements that would give rise to the
'contradiction' of the greatest ordinal number, today called the "Burali-Forti paradox".
In order to carry out such a task, it is of fundamental importance to
analyze how Russell wrote the 'final draft' of his book on the foundations
of mathematics, after meeting Peano at the First International Congress
of Philosophy (Paris, July 1900). In this book, which Russell had been
attempting to write since 1898, he presented his logicist thesis, claiming
that pure mathematics was a branch of formal or symbolic logic. At this
time, Russell was attempting to present a new philosophy of mathematics
by showing that its foundations were consistent, that is to say, that they
were free from contradictions.
82
The discovery of the 'paradoxes'
I will try to show that only a single draft of The Principles of

Mathematics was written after Russell met Peano; this is not in agreement with some historians and philosophers, who-following Russell's
confusing recollections-claim that Russell wrote several manuscripts of
this book between 1900 and 1902. In the second section, I will describe
how Russell wrote his book in different periods. During the first stage,
between November and December 1900, Russell developed Parts III, IV,
V and VI. In the second stage, May and June 1901, Russell wrote Parts
I and II. In the third stage, Russell rewrote Part I (originally written in
May 1901), and Part VII (originally written in approximately June 1900).
Later on, in a fourth stage, Russell added two appendices in November
1902 and the preface in December 1902. I believe that the folios written
during these four different stages were assembled into one single manuscript.
Unfortunately, Russell's own emphasis of the influence of Peano
on his thinking has hidden the tremendous influence of Cantor. I will
present evidence that Russell was heavily influenced by Cantor's ideas
while writing The Principles of Mathematics, especially during the first
two periods. During the first period, Russell encountered certain problems with Cantor's work, which he labeled 'mistakes' or 'fallacies';
during the second stage, Russell concluded that these 'mistakes' were not
merely simple errors, but 'contradictions' related to the most basic concepts of symbolic logic and the theory of sets.
In the third section of this fourth chapter, I will show how Russell
discovered his own inconsistency from the simultaneously gradual discovery of the argument of the greatest cardinal number.
Finally, when Russell attempted to finish his book, he incorporated
into its text those passages discussing the "Cantor paradox", and the
elements that prompted the discovery of the "Burali':Forti paradox".
Throughout this section, I will argue that Russell was motivated to
create the 'paradoxes' in part by his realistic attitude towards the existence of a 'universal class', and by his complete acceptance of Cantor's
theory of sets.
83
4.2 The First International Congress of Philosophy,

Paris, 1900.- Russell spent the first half of 1900 working on his book
on Leibniz. He probably sent the manuscript to the publisher in April
1900, and Russell's personal correspondence shows him asking Moore to
revise and correct the page proofs sent from Cambridge University Press.
The last unpublished manuscript of The Principles of Mathematics-entitled simply "Principles of Mathematics"-indicates that Russell
worked on it a few days before he went to the congress. Russell arrived
at Paris with a firm belief in Platonic Ideas as a result of Moore's influence. He also rejected subject-predicate logic as a result of Bradley's
influence. At the congress, Russell read an essay in which he argued
strongly in favor of the actual existence of 'points' and 'instants'. This
newly advocated logic did not have yet any relation to the symbolic logic
developed by Peano and others:
It was a new metaphysical logic, [ ... ], and was based upon philosophical views advanced by G. E. Moore. These views were, in Russell's opinion, logical because they
assigned a new philosophical status to relations. I
For Russell, the most important event of the conference ocurred

when he met Peano. In his autobiography, Russell claims that he knew
Peano by name and that he had seen his work before the conference, but
he had not taken the trouble to master his notation. Some biographers
have followed Russell's account uncritically; some have even pointed out
specific times and places where Russell met Peano and encountered his
work-on one occasion as early as 1897, at the First International Congress of Mathematicians. It was thought that Peano's work stimulated
Russell's original interest in the philosophy of mathematics. 2
On the other hand, other evidence suggests that Russell had not
seen Peano's work before the congress. There is nothing in his What
shall I read? to indicate that he read Peano, or any members of his
school, prior to the congress. There are no references whatsoever in any
of Russell's previously published books and articles, or even in his unpublished material written at the time.
1. Bowne 1966a, 55.
2. Kennedy 1963a, 262.
84
There is only one place in which Peano's name is associated with

Russell before the congress. This occurs in a letter from Russell to Couturat, dated October 9, 1899. Russell agreed with Couturat in that Peano's distinction between of' (symbol of membership) and' C' (symbol of
containment) was unimportant. But this assertion does not necessarily
mean that Russell had already read Peano in order to substantiate his
position. Couturat had recently published [Couturat 1899a] a long and not
very positive review of Peano's work for the French philosophical and
mathematical community, printed in the pages of the Revue de Metaphysique et de Morale. In this essay, Couturat criticized Peano's method,
asserting that Schroder's was simpler; he also rejected Peano's distinction
between the symbol of membership and the symbol of containment by
stressing that he believed it to be unimportant.
Russell's acceptance of Couturat's judgement without reading
Peano is not unique. On other occasions, Russell had also accepted the
arguments of others, Couturat's in particular, without having made a first
hand examination of the ideas in question. For example, in the letter of
October 9, 1899 to Couturat, Russell agreed with Burali-Forti that the
trichotomy law was false for ordinal numbers and he even suspected that
it was possibly false with respect to cardinal numbers. In the same letter,
Russell asked Couturat to send him the Burali-Forti articles because, at
the time, he had not had the opportunity to read them. In fact, there is a
letter from Russell to Couturat which provides evidence that Russell did
ignore Peano's work, at least, until October 1899. Only then, Russell
informed Couturat that he had read his essay-review on Peano, who was
unknown to Russell before then. 1 So, it appears that prior to the 1900
Congress of Philosophy, Russell had only heard Peano's name through
his correspondence with Couturat, and knew his work indirectly through
Couturat's essay. Possibly, after reading Couturat's essay-review-if in
fact Russell read it, Russell thought that he would not find anything
worth studying in Peano. I should mention that Schroder wrote another
critical essay-review on Peano's work, printed in The Monist. Russell
1. It seems that, before 1900, the only member of the Italian school from which Russell had knowledge
was Mario Pieri (1860-1913). At that time, Russell owned a reprint of one of Pieri's papers [GrattanGuinness 1986a, 107 note 10; and Rodriguez 1988b, 285].
85
appeared to have ignored Schroder even after meeting Peano in 1900, or

perhaps disagreed with the main ideas of the German logician, because
Russell later intended to publish an essay-review of his own.
It is also important to stress that it was not Peano's specific contribution at the congress that mostly impressed Russell. Peano had read an
essay on how definitions should be done in mathematics. Of the three
different procedures outlined by Peano, Russell accepted only one, even
after he had been deeply influenced by the Italian mathematician. In
retrospect, Russell explicitly pointed out that he was impressed greatly by
Peano's successful debating ability. The clarity and perfection of Peano' s
ratiocination were impressive. Russell concluded that Peano's superiority
over all of his colleagues was probably due to his mathematical logic. 1
Having convinced himself that this was so, Russell approached Peano and
asked him for all his previous publications. Now, Russell believed that
Peano's notation
afforded an instrument of logical analysis such as [he] had been seekmg for years,
and that by studying him [he] was acquiring a new and powerful technique for the
work that [he] had long wanted to do. 2
In short, it was at the First International Congress of Philosophy

where Russell realized the importance of mathematical logic. 3 There is
no doubt that, as he said, this congress produced a turning point in his
intellectual life. Returning to England in mid August 1900, he was now
persuaded that Pea no and his colleagues were the best researchers in
1. It seems that what initially impressed Russell from Peano was his discussion on the need for a
notation for the defmite article [the]. Such discussion probably took place between SchrOder and Peano,
once the latter delivered his contribution during the congress [Grattan-Guinness 1975d, 110].
2. Russell 1967a, 144. For other descriptions of what Russellieamt from Peano see: Grattan-Guinness
1975d, 110-112 and Rodriguez 1988b.
3. Perhaps, Russell became aware of the scholars who are considered to be the pioneers of mathematical and algebraic logic through reading Whitehead's Universal Algebra. Nevertheless, if Russell did in
fact read them, their work had little impact on him, before meeting Peano in 1900. Russell gave his
own account: "Most of [Boole's, Peirce's and Schroder's] works were already familiar to me [before
meeting Peano], but I had not found that they threw any light on the grammar of arithmetic" [Russell
1959a, 65]. It is difficult to estimate what Russell meant by the word 'familiar'. Did he study them?
Or, was he only acquainted with the titles? A few years later, in 1902, Russell informed his first wife
Alys that he was planning to study all the great logicians "to begin an exposition of all previous works
of any importance" [Letter: Russell to Alys, June 12, 1902].
86
mathematical philosophy. Russell devoured all possible publications of

the Italian school, spending most of August mastering Peano's notation.
By now his previous conception of Peano and his contributions had
changed completely, almost a year after his correspondence with Couturat. Russell was so impressed by Peano's techniques that he even promised Stout, editor of Mind and his former professor at Cambridge University, a long essay review on Peano and his followers (or perhaps he
wanted to modify Couturat's and Schroder's possible negative opinion).
Although Russell kept notes for this essay review [see: Russell 1901fj he
never published them. Russell commented to various people on different
occasions that he was planning to write the review. At almost the same
time, Russell published another article on the principles of mathematics
showing that all recent advances were due to Zeno's work [see: Russell
1901e]. In this popular essay, Russell was very critical of philosophers,
who obviously disliked the article. Perhaps, Russell was aware of Stout's
discontent, and he decided not to publish his essay review in Stout's
journal as Russell originally had planned. It is also possible that, after
writing the essay, Russell realized that it was inappropriate for Mind.
From the mathematical point of view, there were several conventions, concepts and techniques that Russell took from Peano. The first
and most important was the symbolic language which allowed him to
express complicated mathematical arguments using a small number of
unequivocal signs. Secondly, he now believed that the distinction between
the relation of membership '' and of containment 'c' was of fundamental importance, contrary to what he had said in his letter to Couturat of
October 1899. Russell argued that this distinction allowed one to discern
the proposition 'Socrates is a man' (an element of a set) from the proposition 'all men are mortal' (a set contained in another set). Thirdly, it
was possible to distinguish between a set with a single element, and the
single element by itself (i.e., the difference between '{x}' and 'x'). Fourthly, there were other useful innovations in Peano's work such as the
study of geometrical properties without the use of figures.
In September and October 1900, while he was still reading Peano's
works, he decided to extend his study to the logic of relations. Russell's
previous works on the logic of relations resulted from his study of Leib-
87
niz and his abandonment of neo-Hegelian doctrines. He had even written

an essay entitled "The Classification of Relations" [Russell 1899a]. But
now he had the appropriate technique to express his ideas. The goal of
the new work was to simplify the logic of relations by the use of Peano's
notation. One of the most important factors relating this study was Russell's new definition of an 'infinite class' as one which is similar (i.e.,
there is a one-to-one correspondence) to itself when a single element has
been substracted.
A 'finite class' was simply defined as a class which was not infinite. Russell justified his definition of infinite class by explaining that
although it was possible to define the finite numbers by mathematical
induction and then take the definition of infinite class as a primitive
proposition, he had been unable to deduce any properties from this second definition of infinite class. He refused to define an infinite class as
one that contains a part (or subclass) which can be put into a one-to-one
correspondence with itself. Apparently, he could not show that the part
obtained by substracting a single individual from the totality was isomorphic with the totality itself! Finally, and most important, Russell did not
define an infinite class by the property that an infinite class remained
coordinatable with itself whenever a term was added. He considered that
this definition would exclude the 'class of all individuals' (the universal
class), because it was impossible to add anything to this class. I will
argue that it was this metaphysically realistic attitude towards the existence of a universal class containing all possible individuals that enabled
and motivated Russell to consider the formulation of the set-theoretic
'paradoxes'. It is important to keep in mind that Russell must have believed at this time that it was possible to assume the consistency of the
vast field of mathematics, and he did not expect to find any contradictions within it. Once he had finished this article in October 1900, he sat
down to write, he hoped, a new and definitive version of his book on the
principles of mathematics.
4.3 Russell's writing of the 'first' final draft of The

Principles of Mathematics, November - December 1900.According to Russell's late recollections and his personal correspondence,
he sat down to write The Principles of Mathematics in October 1900 and
88
finished it by December 31 of the same year. If this is so, the question

arises of why he did not immediately send the book to the publisher?
After having worked on the principles of arithmetic for at least three
years, and after several unsuccessful attempts to write a book on this
same topic, it is reasonable to suppose that Russell would be very anxious to publish his thoughts. Perhaps, he may have had a cautious attitude
after having gone through so many revisions. Or, alternatively, it could
be argued that the delay was due to Russell's putative discovery of "Cantor's paradox" in January 1901. Nevertheless, throughout the analysis
presented in this chapter, I will attempt to show that none of these explanations is satisfactory in itself!
In fact, contrary to what Russell expressly said,l I propose that he
did not finish the manuscript by December 31, 1900. He asserted that
Parts III, IV, V and VI were published almost without change, as he had
written them that autumn, while Parts I, II and VII were rewritten. But I
believe that the book (including all later changes) was not finished until
January 1903 and that there are no traces of the original versions of Part
I and II that Russell allegedly wrote in 1900. 2 Moreover, if one wanted
to be accurate, Part VII was not rewritten, but revised from a previous
manuscript of The Principles of Mathematics.
In order to understand the process by which Russell completed the
book, it is useful to consider the writing techniques that he used to complete the earlier drafts of The Principles of Mathematics (and not his
mature writing methodology that he claims to have developed after 1902). Whenever he had previously written something on any relevant
topic, Russell inserted into the new manuscript anythiIlg from the previous drafts that he considered still valid and correct. An examination of
the table of contents of Russell's earlier manuscripts on the foundations
of arithmetic indicates that he inserted sections and, in some cases, long
chapters from his earlier versions.
1. "I finished this first draft of The Principles of Mathematics on the last day of the nineteenth century-i.e., 31th December 1900" [Russell 1959a, 56].
2. Cf, Blackwell 1971b, 7-8; 1984a, 275-277.
89
So, when Russell came back from the congress and read Peano's
publications, he did not simply sit down to write yet another book on the
principles of mathematics. He first revised his latest effort at such a
book, the one which he was still considering a few days before going to
Paris in July 1900. There is, in fact, a large amount of external evidence
of Russell revising the draft of 1899-1900 [Russell 1899h] of The Principles of Mathematics before he started his new approach. For example,
there are several notes added to this version referring to Peano' s work
and some others to the work of Bernard Bolzano (1781-1848). 1 There
were also two large notes commenting on some problems with the approach and how they might be rectified in the new version.
I have mentioned before that there are no traces of Part I (,The
Indefinables of Mathematics') and Part II ('Number') in the manuscript
version allegedly written in 1900. The suggested explanation is, I would
say, straightforward: Russell simply did not write Parts I and II at that
time. Al least two different alternative explanations can be derived from
Russell's own description of the events surrounding the preparation of
The Principles oj Mathematics. In a letter to Jourdain dated April 15,
1910, Russell stated that "during the rest of the year [possibly late October, November and December] I wrote parts III-VI of my [The] Principles [oj Mathematics] (Part VII is largely earlier, Parts I and II wholly
later, May 1902); ... ".2 In another letter to Jourdain dated September
11, 1917, Russell informed him that he had written Parts III to VI of The
Principles in the autumn of 1900. 3 There is yet another source of evidence that supports my conjecture. According to the personal diary that
Alys Russell kept at that time, her husband only finished two thirds of his
book by the end of 1900. 4
Why, then, one would ask, did Russell inform Helen Flexner
(1871-1956) that the book was finished on December 31, 19OO? It was,
perhaps, a 'small Russellian exaggeration'. Here the reader faces an
1. Russell seems to have read Bolzano, for the first time, in August 1900 [see: Russell 1983a, 363].
2. Letter: Russell to Jourdain, April 15, 1910; quoted in Grattan-Guinness 1977a, 133.
3. Letter: Russell to Jourdain, September 11, 1917; quoted in Grattan-Guinness 1977a, 143-144.
4. This diary is kept at The Bertrand Russell Archives.
90
historiographical and methodological question. On one hand, one of

Russell's best known biographers has claimed that he concluded that,
given any quote from Russell, there is another that contradicts the first.
This implies that Russell's quotes are not always reliable. In such cases,
when I find contradictory quotations of Russell, I have selected, obviously, those that support my new interpretation of the events surrounding the
preparation of The Principles of Mathematics. In particular, when I have
found these contradictory statements in Russell's personal correspondence, I have attempted to weigh the credibility of these statements by
examining the individuals involved. In this case, there is no doubt in my
mind that Russell was more careful when writing to Jourdain (who was a
colleague as well as a historian of mathematics, who was publishing
essays on the recent developments of theory of sets and mathematical
logic) than when he was writing to Helen Flexner (whom he was, perhaps, simply trying to impress). Obviously, other historians might argue
that the dating of the misives is of fundamental importance and that Ms.
Flexner's letter should be considered first.
A second letter to Helen Flexner may also be questioned on the
same grounds as the first. In the second letter, Russell told her that he
now was working on the second volume of The Principles of Mathematics, and that his work was nearly finished. 1 But there is no material in
The Bertrand Russell Archives supporting his assertion. 2
One could claim that the manuscript was not finished by the date
cited because at least, Part VII (,Matter and Motion') was not added to
the 'final manuscript' until May 23, 1902 [see quote page 120]. In following this new interpretation, I suggest that there were no separate
drafts of the manuscript of The Principles (e.g. 'first', 'penultimate' and
'final'). I propose instead that there was only one draft composed in the
following order: (i) Parts III, IV, V and VI were (added and) written
J
1. Letter: Russell to Helen Flexner, August 2, 1902.

2. Nevertheless, there is no resson to discredit Russell's achievements. Even if he did not finish the
whole manuscript of '!he Principles ofMalhemolics by December 31,1900. He did finish Parts ill, IV,
V and VI, a great accomplisment in itself. These sections of the book contained almost five hundred
folios distributed in thirty four chapters. Russell argued, accurately in this esse, that he could write an
average of ten folios a day.
91
during the fall of 1900; (ii) Parts I and II were originally composed in
May and June 1901, but Part I was rewritten in May 1902; (iii) also in
May 1902, Part VII was taken from an earlier attempt to describe the
foundations of mathematics and added to the current manuscript; (iv)
between May and November 1902, Russell wrote appendices A and B, as
well as carrying out more changes and additions; (v) the Preface was
delivered to the publisher in December 1902; (vi) finally, in January
1903, Russell added some new folios to the manuscript in which he
summarized some of his ideas. I believe that, if Russell originally contemplated a book divided in seven parts, then he wrote the book over
three global periods of time: October - December 1900, May - June
1901, and, May 1902 (- January 1903)-excluding the galley revisions
from the period of composition. Throughout this time until May 23, 1902
the manuscript was incomplete. This is the reason why most folios of The
Principles of Mathematics were not duplicated (or triplicated!) during its
composition, with the clear exception of those contained in Part I which
were effectively rewritten. If some folios were renumbered in the 'final'
manuscript, it is because they were written during Russell's earlier attempts to write a book on the principles of mathematics [Russell 1898e,
Russell 1899b, & Russell 1899h], before he met Peano at the First International Congress of Philosophy in Paris in July 1900. Thus even if Parts
I and II were originally written in 1900, the manuscript remained incomplete by the end of that year. I think, however, that Russell did not write
Parts I and II before December 31, 1900. It is very curious that the most
significant parts of the original manuscripts of The Principles of Mathematics are still in existence, but nothing remains of Parts I and II. It
could be argued that, in fact, all the manuscripts are incomplete, even the
'final' one. Possibly in some sections almost the whole text is missing,
not only one or two Parts. But, I do not think that this was the case.
There is an internal element contained in the manuscript of the
final written version of The Principles of Mathematics that might support
this new historical hypothesis. In defining the irrationals, Russell asserted: "Arithmetical theories of irrational could not be treated in parts I or
II since they depend essencially upon the notion of order".' In the print1. Russell 1900e Part V Chapter III Folio 40, my emphasis.
92
ed version, Russell specified Part II. This discrepancy seems to indicate

that, by that time, November 1900, he was not aware of the particular
section in which he had discussed the theories of irrational numbers.
Another argument suggesting that Parts I and II were written after
1900 is of an external technical character. I mentioned before that by
1897 Russell had already concluded that one of the fundamental problems
in the philosophy of mathematics was that of the relation between the
categories of 'number' and 'quantity'. However, by 1900, Russell was
convinced that the complete separation of 'number' and 'quantity' had
been proven. 'Quantity' had lost its mathematical importance because all
of its theorems could be generalized from the properties of 'number' and
those of the concept of 'order'. 'Quantity' was not definable in terms of
logical constants, and it was therefore an idea that did not belong to pure
mathematics. He discussed the category of 'quantity' because it was
traditional to do so when discussing the foundations of mathematics, but
stated that it was not important. Possibly, Russell thought that he had
summarized his current understanding of the topic of quantity in his
previous manuscript on the foundations of mathematics (finished approximately in June 1900). It is important to notice that Part III devoted to the
discussion of the notion 'quantity' diappeared from that attempt and it is
possible that Russell transfered it to the new manuscript, since this part
only required superficial changes. In the new manuscript, only a few
changes were needed. As he said in a note added to the same manuscript,
his entire conceptualization of the logical calculus was incorrect and Part
I needed to be completely rewritten. Perhaps, while writing The Principles of Mathematics, he had similar thoughts about his ideas on 'Number'
(Part 11), and decided to leave these two Parts to the end. First of all, he
would work on the material that only needed superficial revision. I agree
with the conjecture that it was in November (or late October) 1900 when
Russell began to write what he hoped would be the final version of The
Principles of Mathematics [see: Table IV].
Most importantly, Russell was now convinced that all of the philosophical antinomies could be solved or reduced to a simple difficulty
related to the concept of infinite number, which could be solved by an
adequate philosophical analysis of the term 'any' (the existential quanti-
93
Table IV. The Principles of Mathematics, Parts III-VI (Oct-Nov 1900).

Introduction.- The following Table of Contents summarizes the material that I believe Russell
wrote in the last two months of 1900 (a folio of Part V is dated 'Nov 24, 1900'). The reader should recall
that I also claim that Part VII was not composed at that time, but taken from an earlier manuscript (with
some revisions and additions). The fll"St folio of Part III is dated 'Nov, 1900', but this seems to be a note
added to indicate when Russell revise this section. All the corresponding folios are contained in the fmal
draft of The Principles of Mathematics.
Quantity
The meaning of magnitude
The range of magnitude
Numbers as expressing magnitudes: measurement
Zero
Infmity, the infinitesimal and continuity
Order
The genesis of series
The meaning of order
Asymmetrical relations
Difference of sense and difference of sign
On the difference between open and closed series
Progressions and ordinal numbers
Dedekind's theory of numbers
Distance
Infmity & Continuity
The correlation of series
Real numbers
Limits and irrational numbers
Cantor's first defmition of continuity
Ordinal continuity
Transfmlte cardinals
Transfinite ordinals
The infmitesimal calculus
The infmitesimal & the improper infmite
Philosophical arguments concerning the infmite
The philosophy of the continuum
The philosophy of the infmite
Space
Dimensions and complex numbers
Projective geometry
Descriptive geometry
Metrical geometry
Relation of Metrical to Projective and Descriptive geometry
Dimensions of various spaces
The continuity of space
Logical arguments against points
Kant's theory of space
94
fier). He was probably referring to those antinomies related to the concept of 'infinite number' that he had stated, in previous drafts on the
foundations of mathematics to be insoluble. Earlier, he had criticized
Cantor's Mengenlehre as lacking a correct philosophical foundation to
deal with the antinomies. Now, Russell argued, the origin of the 'intolerable contradictions' was located in the supposition that well defined
infinite sequences (or ordered series) have a last term. Two years or so
before, he had been 'indignant' towards Cantor because of Cantor's
supposition of the existence of 'w', when there was no last term in the
sequence of all natural numbers. Russell now distinguished between
infinite and finite numbers by arguing that infinite numbers did not obey
the principle of mathematical induction. This was true, since
Finally, although Russell emphasized the complete separation of 'number'

and 'quantity', stating that the latter did not belong to the realm of pure
mathematics, he did not express, at any point in this section of the manuscript, the strong logicist thesis of his book: the fact that all mathematical
concepts and theorems could be defined in terms of logical concepts and
could be deduced from them. At this early stage of the writing of the
book may not yet have perceived such an important role for logic. As a
result of further reading of Peano and Cantor and of his own efforts in
writing Parts IV and V he became convinced of the complete dependence
of mathematical concepts on logic.
Part IV ('Order') was also written in November 1900. He justified
his lengthy explanation of matters relating to the concepts of 'order'
because he used his correct philosophical analysis of 'order' to solve all
the inconsistencies associated with the idea of 'infinity' that he had encountered up to that date.
Part V of The Principles of Mathematics (on 'Infinity and Continuity') was also written in November 1900. In it, Russell explained the
applicability and the significance of Cantor's ideas for the foundations of
mathematics. Now, Russell argued, the fundamental problems of the
philosophy of mathematics focussed on the concepts of 'infinity' and
95
'continuity'. Here one finds that Russell finally accepted Cantor's 'w'.
When considering the philosophical discussion of transfinite cardinal
numbers, it was of prime importance to understand that a one-to-one
relationship was not at all relevant to the concept of 'number'; that is to
say, there was no need to count the elements of an infinite set to find out
if it was 'denumerable' or not (i.e., equivalent, symbolized by '-', to
the sequence of all finite cardinal numbers). On the other hand, the concept of 'similarity' was also independent of the concept of number (as I
mentioned before, similarity occurs when two ordered-sets can be put
into a one-to-one correspondence while maintaing the ordering relations,
symbolized by ':::::' which allows one to compare the cardinality of transfinite classes [see: pages 9-10]). Logically speaking, a proper substitution
of the concept of 'any' instead of that of 'all' would avoid the possibility
of problems relating to infinite totalities.
As I said before, Russell was very critical of the ideas he took
from others. In his discussion of transfinite cardinal numbers he therefore
doubted that it was possible to show that they were comparable. His
discussion of transfinite ordinal numbers showed a completely new attitude towards Cantor's 'w' different from that expressed in his [1898a]
manuscript and his essay of 1896 [Russell 1896d]. Russell explained that
some of its basic properties could be expressed by the fact that
l+w=ww+1.
One of the main goals of Russell's fifth part was to show that the notion
of 'continuity' was free from contradictions; it was of fundamental importance now to prove that both mathematics and, especially, the new
philosophy of the infinite, were consistent.
Russell asserted, on numerous occasions, that Cantor's theory of
continuity was free from contradictions. He was using this term believing
that mathematics was a consistent science. In future, contradictions would
prove to be a serious challenge. He examined Zeno's famous paradoxes
in detail: the dichotomy, Achilles, the arrow and the measure. Russell
claimed that Zeno' s arguments could not lead to the conclusion that there
were inherent contradictions in Cantor's concept of continuity. Cantor's
96
concept had remarkable features: It was associated with the doctrine of

limits, and it affirmed that infinitesimal segments did not exist. Russell's
adoption of Cantor's ideas, as well as his metaphysical realism, had
progressed so far by this time that he not only believed in infinite classes,
but actually thought that he could prove their real ontological existence.
According to Russell, infinite classes could include the number of terms
and number of ideas.
Nevertheless, there were Cantorian results that Russell had difficulty in accepting, in particular one which Russell originally thought to be a
mistake. Cantor had shown that there was no greatest cardinal number.
He had proved that
2n
> n, for any n.
This meant that given any transfinite (or finite) set, it was always possible to show the existence of a greater one. In short, it was impossible to
consider an 'universal set', containing all sets, or all classes. But Russell's metaphysical realism, which originated as a consequence of his
rejection of neo-Hegelian doctrines, made him believe that there was such
a class: The class of classes. So, he stated that,
there is a certain difficulty in regard to the number of numbers, or the number of
individuals, or of classes. Numbers, individuals, & classes, each form a perfectly
dermite class, & it will be remembered that we found a general proof, from the
reflexiveness of similarity, that every class must have a number. Now the number of
individuals must be the absolute maximum of numbers, since every other class is a
proper part of this one. Hence, it would seem, the numbers have a maximum. 1
In order to show that there was no contradiction between this idea
and Cantor's, Russell examined the earlier difficulties and their solution.
First of all, he explained, by showing that the class of classes was similar
to a proper part of the class of individuals and vice versa, then that the
number of classes was the same as the number of individuals. Secondly,
he argued that "the number of numbers, which always formerly led to
contradictions, can now be freed from all contradictions". 2 There was no
1. Russell 1900e Part
v Chapter XII Folios 188-189.
2. Ibid., Folio 189-190.
97
difficulty among the finite numbers. If one thought of 'n' as the number
of numbers, it was also the greatest number. There was no difficulty
among transfinite numbers either. Because, "the number of numbers up
to and including a o (No in Cantor's notation), is still a o."! This remained
the case until one had run through as many cardinals as there are ordinals
of the second class, and one could go through the whole hierarchy so that
the number of numbers was always less than the greatest number. Russell
was completely convinced that there was a greatest cardinal number: that
of the class of all individuals. But, as he said,
Cantor has given two proofs [ ... ] that there is no greatest number. If these proofs be
valid, there would seem to be still a contradiction. But perhaps we shall find that
[Cantor's] proofs only apply to numbers of classes not containing all individuals [ ... ].
It is essential, however, to examine this point with care, before we can pronounce
infimty to be free from contradlctions. 2
The first of the proofs did not present great difficulties for Russell.
He simply rejected its basic premises. There was no way, he argued, to
support the hypothesis that there was a one-to-one correspondence between the transfinite cardinal and ordinal numbers. Russell also rejected
the notion that ordinals had no maximum. I have already mentioned that
Russell did not accept Cantor's proof of trichotomy law for the transfinite
cardinal numbers, and based on this assumption, he could not think of the
transfinite cardinals,
as being ordered, and forming a sequence (or ordered class). Russell

continued
the proof in the second of the passages above referred to is quite different, & is far
more definite. But here again, I think, we shall find out, when the impossibility of
repetition of any individual is realized, the proof fails to be universal. 3
In the article in which Cantor presents his proof of this theorem, he

makes three major points. He made a simple demonstration that there are
1. Russell 1900e Part V Chapter XII Folio 190.
2. Ibid. Folio 189.
3. Ibid . Folio 193.
98
higher powers than No using his now famous diagonal procedure. Secondly, he demonstrated that this diagonal method can be applied to any power or cardinal number. Finally, he proved the existence of powers higher
than that of 'the continuum'. 1 Russell's rebuttal of the second part of
Cantor's proof of the second point ran in the following way:
the argument by which it is to be shown that the number of classes of classes exceeds
the number of classes may be disproved in the following manner. We have u =
class, so that x is a u means x is a class. When! is not a class of classes, let k,.
be the class of classes whose only member is!. When! is a class of classes, let k,.
be ! itself. Then we define a class u', in accordance with the above procedure
[Cantor's diagonal method), as containing every! which is not a member of its k,.,
and no ! which is a member of its k,.. Thus when! is not a class of classes, ! is not
a y'; when! is a class, or class of classes, or class of classes of classes, or etc., ! is
not a y'; but when! is any other class of classes, ! is a u'. Then Cantor infers that
u' is not identical with k,. for any value of !. But u' is a class of classes, & is
therefore identical with k,... Hence Cantor's method has not given a new term, & has
therefore failed to give the requisite proof that there are numbers greater than that of
classes [ .. ). Thus what Cantor has proved is, that any power other than that of all
classes can be exceeded, but there is no contradiction in the fact that this power
cannot be exceeded. 2
After completing this demonstration, Russell thought he had disproven all

possible occurrences of contradictions associated with the infinite, and
concluded the section by asserting that "no definite contradictions can be
proved concerning [the infinite and continuity].,,3
Part VI, that dealt with the concept of 'space', was written in
December 1900. Here, Russell was able to show a close logical connection between the concepts of arithmetic and those of geometry. In particular, Cantor's findings indicated that the continuity of the straight line
was exactly the same as that of the real numbers, or 'the continuum'. In
an additional folio, Russell stressed his logicist thesis, which asserted that
all mathematical concepts could be derived from the indefinable precepts
1. See: Cantor 1932a, 278-281 & Dauben 1979a, 165-168.

2. Russe1l1900e Part V Chapter XII Folios 197-198.
3. Ibid., Folio 199. [This folio, in particular, was dated 'Nov 24, 1900').
99
-=--- _
. d
f -
~ ~~~.&. ~ ..........
Q.L
ct.....
t ....Jl.-,J'). .. ...J-... ....,.j;t .......
~ ") , .r...n,:-.., 1<. ~f\.
,..
I ......
., . . . .
~ ~ .:.....u..(:.;.~ . ~~-. -~d:....t........... r,( ~a.ltQ.

~ ...... Jf , ~ " ~ ' H-..... ...-. ........ ~a...
'-~ ~ ~ /-a...~ :... k....-. If.1Q..
if- ~( ~,
..-'oJu> C.
--..l ~...'l.- ~ - ; .
, atcS1": ~ gcJl..;,.h~ ~ ~
.lJ. :' .
:1..:......
' -r---
...Q.,,,
.,)I
. Vla. ~ Lb..-
Q.... ~ C .,' c..chrv.. l.

t!..,,- ~ ,~ ~
..... ........ ~ ~d;;..., ~o., .OD ... ....-. ~ , ~f

""
~
c... c:t...~ rt /0(,

,;
~ 14:.. ~ C ,
~ ;........ bf
I<.
.:...
0/-
0 . ,
'"
CL~ ~ .
<{ . . -
....~ r .;,
cb- ~ ... ~ f
" I '4-:"
,:,,~ , -d.. ' 'i ~
....
9,.-.#'11-
?J._ \... ....-.... t
~ -~.:..rJ:.. L~ . --.~'- ~ =--~~~"<' ~ fI.... ~ ~
...... .--, ~ . t!l n..~ a..J'~ ~~ tL. .'
r '..
f'
"....
h(l o~", ~ ~
.....
......
,~ .
f-:' .... .
...:., ........ .... n. / ct...1.........--,,~
~~ l.o ~ --~ .J'\... __ . D ...n.......:.u.~,w<-, oI' !' - ~ ..,..

~', ............. ~~
tSl
1C 1\.
u'\.&. ,
...
.. .:..
CULL
~ l ~~
~ !(~~
'1 ~'
nIb,;
t..<.' t- r"" ~ ... ~
(.I'...a.r '" , n- .
Plate 7. Part VI Chapter III Folio 54a.
" r')...,.;.,.,q.. ~ ::.-u
~'.
,.
l ~'- -;...
.....
100
of logic. 1 There are other statements in later sections of this Part that
reinforce the logicist thesis, but the folios of the original manuscript are
lost and there is no way to determine whether or not these comments
were later additions.
4.4 The 'paradoxes'.- Even before Russell had finished this

supposedly final manuscript, he had informed Couturat of the 'mistakes'
that he found in Cantor. On December 8, 1900, Russell explained to
Couturat: "I have discovered a mistake in Cantor, who maintains that
there is no largest cardinal number. But the number of classes is the
largest number [... ]".2 At this stage, Russell did not claim to have found
a contradiction in Cantor's ideas. He was hoping to find a consistent
explanation of the principles of mathematics, and Cantor's theory played
a fundamental role in Russell's task. At the time, he thought he would be
able to show that Cantor was mistaken, but had not yet found any errors
in Cantor's second proof.
In the first days of January 1901, Russell wrote an essay for The
International Monthly [Russell 1901e] , where he publicly proposed his
logicist thesis. In order to determine exactly when Russell first publicly
stated the logicist thesis it is worth noting a letter from Russell to Jour-
dain on April 15, 1910: Jourdain approached Russell on questions concerning the origin and development of his ideas on mathematical logic.
Russell states:
As regards your questions of autobiography, the first thing I wrote showing the
influence of Peano was the article [Russell 1901b] in R[ivistaj d[ij M[atemolicaj,
[Volume] VII. I don't remember whether that was published before or after the
article [Russell 1901e] in The International Monthly: if after, the [7hej International
Monthly was my first Peanoesque publication.]
1. It is impossible, as far as I know, to determine exactly when Russell added this folio. It might have
been immediately after finishing this sixth part in 1900: or maybe later, after he had composed Parts I
and II and was revising the manuscript for publication. It is worth mentioning that most references
explicitly supporting the logicist thesis seem to have been added after the first attempt to write the book
was completed.
2. Letter: Russell to Couturat, December 8, 1900: quoted in Moore & Garciadiego 1981a, 325.
3. Letter: Russell to Iourdain, April IS, 1910; quoted in Grattan-Guinness 1977a, 132.
101
Nevertheless, the careful reader should distinguish between the article

containing his logicist thesis and another showing Peano's influence. The
reader should also keep in mind that while there might be some difficulty
concerning the dates of publication of the article, there is no uncertainty
concerning the dates of composition. The Rivista di Matematica article
was written in early October 1900, before Russell began to write The
Principles of Mathematics and before he had read Cantor's seminal articles of 1895 and 1897. Consequently, he wrote the Rivista article before
he had developed his logicist thesis. The International Monthly essay,
however, was written in early January 1901, once four important sections
of The Principles were finished.
There are important philosophical, mathematical and historical
issues contained in this popular article from the The International Monthly entitled "Recent developments in the philosophy of mathematics". One
finds a brief description of the work of the Italian mathematicians, especially Peano. Here, Russell was very critical of certain philosophers
(perhaps having in mind his former Cambridge professors) because of
their ignorance of the latest innovations in mathematical philosophy.
Russell also discussed how the new philosophy of the 'infinite' and 'continuity' solved the fundamental questions of the philosophy of mathematics, including Zeno's four arguments involving the concepts of 'continuity', 'infinity' and 'motion'. But, most importantly for the issues raised
in this book, Russell also mentioned the existence of a mistake in Cantor's reasoning-although fallacy was the term he used:
There is a greatest of all infinite numbers, which is the number of things altogether,
of every sort and kind. It is obvious that there cannot be a greater number than this,
because, if everything has been taken, there is nothing left to add. Cantor has a proof
that there is no greatest number, and if this proof were valid, the contradictions of
infinity would reappear in a sublimated form. But in this one point, the master has
been guilty of a very subtle fallacy, which I hope to explain in some future work. I
Russell did not use the word contradiction to describe Cantor's mistake
or fallacy. At this time Russell still believed that mathematics was a con-
1. Russell 1901e, 95.
102
sistent science; but if Cantor's fallacy were to stand, then Zeno's contradictions would still be valid. But Russell did not discuss the issue further.
In the meantime, Russell's correspondence with Couturat had
continued throughout January 1901. Russell kept insisting on the validity
of the concept of 'class of classes', while being pressed by Couturat who
disputed the utility of this concept on the grounds that it could be associated with some sort of contradictions. Through this line of thinking,
Russell eventually perceived Cantor's theorem not as an error but as the
seed of a real contradiction.
It is uncertain which problems in philosophy of mathematics occupied Russell during February, March and April of 1901. But there is a
folio, probably written in April 1901, containing a table of contents of a
first version of Part I of The Principles of Mathematics, on 'The IndefinabIes of Mathematics' [see: Table V]. This table of contents has some
similarity to another one that Russell prepared approximately a month
later. In addition to Part I, this second table of contents outlines Part II,
which dealt with 'Number' [see: Table VI]. It should be noted that there
is no mention of the contradictions in either table. 1
Assuming that Part I was not written in the fall of 1900, the first
version of Part I was likely composed in May 1901 and thoroughly rewritten a year later. Although Russell partitioned the first draft when he
composed the second one, Blackwell has been able to reconstruct it [see:
Table VII]. 2 There are several important issues contained in this first
draft of 1901, which was originally divided into eight Parts or chapters.
First of all, it includes Russell's definition of 'pure mathematics' as "the
class of all propositions of the form "a implies b", where a & bare
propositions each containing at least one variable, & containing no constants except logical constants such as can be defined in terms of logical
constants."3 This is a definition of pure mathematics derived from his
1. There may have been some discussion on the 'contradictions'. However, the fact that the 'contradictions' are not listed in either table suggests that Russell may not have considered them important at the
time.
2. See: Blackwell 1971b and Blackwell 1984a.
3. Russell 1901i Part I Chapter I Folio 1.
103
Table V. Part I. The indefinables (April 1901).
Introduction.- Russell drafted this Table of Contents of Part I

of The Principles of Mathematics in April 1901. This Table of Contents is now kept at The Bertrand Russell Archives in File 230-033-
50.
Part I. The indefinables of mathematics
Chapter
"
"
"
"
"
"
"
"
"
"
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
Definition of pure mathematics

Terms & concepts
Propositions & assertions
Implication
Propositions of given form
The variable[: conjunction & disjunction]
Conjunction, disjunction & intension
Classes: extension & intension
Relations
Definition
Peano's symbolic logic
104
Table VI. Parts I & II (May & June, 1901).
Introduction.- This is an outline of the Table of Contents of

the surviving folios of the first two parts of The Principles of Mathematics. Both parts were originally written in May and June 1901.
Part II was published (in May 1903) as it was originally formulated.
Part I had to be rewritten once Russell became completely convinced
of the importance of his 'contradiction'.
The Principles of Mathematics
Chapter
"
"
"
"
"
"
"
Chapter
"
"
"
"
"
"
"
II
III
IV
V
VI
VII
VIII
Part I. The variable

The definition of pure mathematics
Terms & concepts
Classes & relations
Conjunction & disjunction
The variable
Implicaction
Summary of Part I
I
II
III
IV
V
VI
VII
VIII
Part II. Number

Definition of cardinal numbers
Addition & mutiplication
Finite & infinite
Theory of finite numbers
Addition of terms & addition of classes
Whole & part
Infinite wholes
Ratios & fractions
105
logicist thesis. But the most remarkable aspect of this draft of Part I, for
the present discussion, is that it does mention a contradiction within the
logic of classes. According to Russell,
We saw that some predicates can be predicated of themselves. Consider now those of
which this is not the case. These are the referents (& also the relata) in what seems
like a complex relation, namely the combination of non-predicability with identity.
But there is no predicate which attaches to all of them & to no other terms. For this
predicate will either be predicable or not predicable of itself. If it is predicable of
itself, it is one of those referents by relation to which it was defined, & therefore, in
virtue of their definition, it is predicable of itself. Conversely, if it is not predicable
of itself, then again it is one of the said referents, of all of which (by hypotheses) it
is predicable, & therefore again it is predicable of itself. This is a contradiction,
which shows that all the referents considered have no exclusive common predicate, &
therefore, if defining predicates are essential to classes, do not form a class. [ ... ].0
It follows from the above, that not every defmable collection of terms forms
a class defmed by a common predicate. I
At this point, Russell thought that this contradiction meant that

there was a need for some kind of limitation when defining classes. Cantor had already arrived at the same conclusion a few years earlier when
he distinguished between consistent and inconsistent multiplicities.
There are two remarkable points closely associated with Russell's
passage quoted above [see, also: Plates 8 & 9]: First, and most importantly, this is the earliest statement of Russell's contradiction of the class
of all classes which are not members of themselves. Secondly, it is possible that both Russell and Cantor had begun thinking about these contradictions as a result of the theorem proven by Cantor in 1890 and 1892.
Russell's conclusion, derived from the above contradiction, asserts that
Peano holds [ ... J that every proposition containing only one variable is reducible to
the form 'x is an a'. [ ... J. But we shall see (Chapter IX) that at least one such proposition is not reducible to this form. 2
More than six months after his first discussion of the potential
errors in Cantor, Russell was still unable to demonstrate the nature of the
fallacy in Cantor's work. There are important similarities concerning
1. Russell 1901i Part I Chapter IX Folio 22-23. My emphasis.

2. Ibid. Folio 56.
106

Table VII. A The Principles of Mathematics, May 1901.
Introduction.- In May 1901, Russell developed a tentative Table of Contents of his

entire manuscript (kept in File 230.03350). The headings of Parts III and IV are missing and the
title of Part VII is mentioned but the topic was not developed. Perhaps this is an indication that
Russelll ignored, at this time, the chapters composing Part VII and that he did not write it (under
the new influence of Peano) in December of 1900. Note that nothing has been written about the
contradictions.
Part I
Chapter
Part II
Chapter
Part III
Chapter
I
II
III
IV
V
VI
VII
VIII
The Variable
The definition of pure mathematics
Terms & concepts
Classes & relations
Conjunction & disjunction
The variable
Implication
Summary of Part I
I
II
III
IV
V
VI
VII
VIII
Number
Definition of cardinal numbers
Addition & multiplication
Finite & infinite
Whole & part
Infinite wholes
Ratios & fractions
I
II
III
IV
V
(Quantity)
Meaning of magnitude
Range of quantity
Measurement
Zero
Infinity, infinitesimal & continuity
(continues ... )

Table VII.B The Principles of Mathematics (May 1901).
Part IV
Chapter
(Order)
I Genesis of series
II Meaning of order
III Asymmetrical relations
IV Difference of sense & of sign
V Open & close series
VI Progressions & ordinal numbers
VII Dedekind's theory of numbers
VIII Distance
Part V
Chapter
Infinity & Continuity

I Correlation of series
II Real numbers
III Limits & irrational numbers
IV Cantor's first definition of continuity
V Ordinal continuity
VI Transfinite cardinals
VII Transfinite ordinals
VIII The infinitesimal calculus
IX The infinitesimals & the improper infinite
X Philosophical arguments concerning the infinitesimal
XI Philosophy of the continuum
XII Philosophy of the infinite
Part VI
Chapter
Space
I Dimensions & complex numbers
II Projective geometry
III Descriptive geometry
IV Metrical geometry
V Relations of metrical to projective and descriptive geometry
VI Definitions of various spaces
VII The continuity of space
VIII Absolute & relative position
IX Kant's theory of space
Part VII
Matter & Motion
107
108
Russell's earlier denial of the existence of 'w', and his current rejection
of Cantor's theorem on the cardinality of the power set. In 1896, Russell
had denied the existence of 'w', arguing that it was impossible to determine 'w' when there was no last element in the infinite sequence of finite
numbers. Having accepted 'w', Russell argued that there was a limit to
the sequence of transfinite cardinal numbers. There was a greatest of all
classes: The class of classes. His metaphysical realism did not allow him
to accept the unlimited existence of the transfinite cardinal numbers,
because, as he said, once one has considered the class of all classes there
is nothing left to add. The rejection of Cantor's theorem seems to be a
reiteration, with different philosophical justifications, of the reason Russell had used to substantiate his denial of the existence of 'w'.
As a result of Russell's reasoning, two contradictions were emerging almost simultaneously. 1 First, Russell had examined Cantor's theorem in order to avoid the contradiction known today as the 'Cantor paradox' or 'the paradox of the greatest cardinal number' [see page 32]. On
the other hand, he was also considering another problematic argument
which arose as a consequence of his considerations of Cantor's theorem.
I already pointed out that Russell wrote, in his draft of Part I, written in
1901, a statement contradicting Peano' s principle that every proposition
containing only one variable was reducible to the form 'x is an a'; that is
to say, that every proposition of this kind defined a class. But Russell
had shown that this was not the case when one considered the 'predicates
which were not predicates of themselves'. His argument runs as follows:
there is a class defined by all classes which are not members of themselves. The question was whether this class was a member of itself or
not. If it were to be then it would mean that it was not a member of
itself. If it were not, then it would mean that it was one of those classes
which were not members of themselves and therefore it was a member of
itself. So from each statement its own negation followed. Russell concluded that this contradiction meant that there was a limitation to Peano' s
1. In fact, Grattan-Guinness has asserted that it is possible to discuss four different types of 'paradoxes'
when following Russell's arguments, but he was unable to visualize all of them [Grattan-Guinness
1981a, 17].
109
principle. But there are no signs in the draft, or other unpublished material written at the time, to determine whether Russell changed his overall
philosophy, in those days, as a result of his discovery. There are no signs
that he either deduced this concept ('the predicates which are not predicated of themselves') or applied it to Cantor's second proof. I am suggesting here that perhaps Russell thought of this class when examining
Cantor's theorem. It was not necessary to apply the concept to any other
argument, because it was a contradiction in itself, contradicting Peano's
principle.
The earliest surviving manuscript of Part II was originally written
in June 1901. This section of The Principles of Mathematics clearly
shows the impact of Cantor's influence on Russell. Throughout the text
of Part II there are several footnotes referring to Cantor's work, especially to his "Beitdige zur Begriindung der transfiniten Mengenlehre" parts I
and II, in which Cantor finally expressed the principles of his theory on
transfinite cardinal and ordinal numbers in a clear and concise way. 1
Russell had read these two essays in November 1900,2 and their impact
on him was already evident in Part V ('Infinity and Continuity'). There
were several aspects of Cantor's treatment of numbers that fascinated
Russell which were incorporated into Part II. For example, Cantor gave
a definition of infinite number that was completely independent of the
definition of finite number. Russell intended to make this distinction in
his first contribution to Peano's journal. The operations of addition and
multiplication for cardinal numbers had been defined without specific
reference to the finiteness or infiniteness of number. The procedure of
comparing the 'equivalence' (and therefore the cardinality) of different or
equal classes through one-to-one correspondences had eliminated the
process of counting (from the Empiricist approach, for example). It had
further been shown that this process of counting was irrelevant also for
the definition of 'whole and part'. Further, those instances of one-to-one
1. See: Dauben 1979a, 168.
2. Russell 1983a, 364. [CI. Grattan-Guinness [1980a, 76] who asserts that "Russell did not read
Cantor's 1895-97 paper until 'a good deal later' than 1898, and perhaps not by 1901, when early in the
year he discovered Cantor's cardinal paradox for himself"].
110
correspondences were defined independently of the concept of 'number',

and so a new and consistent philosophy of (infinite) number was formulated.
Russell was so taken with Cantor's formulation of the theory of
number that he wrote an essay for Peano's journal in which he would
apply his theory of relations to the main body of Cantor's results. In this
article, Russell explicitly mentioned that Cantor was the source of most
of his definitions and theorems. In a later recollection, Russell explained
to Jourdain that he was only testing his "logic of relations by its applicability to Cantor". 1 It is essential to note that the French manuscript of
this article was written in July or August 1901 [Russell 1901g], in addition to an English version which had been written previously [Russell
1901h].2 It may be that Russell wrote the English version during June
1901, when historians think that Russell discovered his contradiction of
the class of all classes which are not members of themselves; this date is
also the one provided by Russell. I would argue, however, that Russell
had already derived his contradiction, as suggested by his draft of Part I,
which was written a month earlier. He probably became convinced of his
discovery as a result of writing this article for Peano's journal. The
reader should keep in mind the following arguments. First of all, if that
specific folio of Part I is not the first description of Russell's contradiction then, Grattan-Guinness has correctly claimed that the pages where
Russell originally outlined the discovery of his 'paradox' do not seem to
have survived-unless these are the folios reproduced in pages 111-112,
making it impossible to determine exactly how and when Russell made
his discovery. Coffa argued that it was not a sudden discovery but "slowly emerged over a period of time,,3 and did not have an immediate impact. If these observations are correct, then it is possible that Russell
realized the validity of the discovery while writing the Rivista di Mathematica article in June 1901. Now, he distinguished his contradiction from
the arguments associated with Cantor's theorem, although perhaps not as
clearly as might be desired. It is important to keep in mind that Russell
1. Letter: Russell to Jourdain, April IS, 1910; quoted in Grattan-Guinness 1977a, 134.
2. Blackwell no date, item # C0204.
3. Coffa 1979a, 37.
111
~ \4
..v...u..
J.-...c...~
~ J.,'l( ~ (k:'l...riA...
~'t~
.,..
'~.
;....
fi~
-:.
~ (""'~
~A").. ":""", ~ ..... ~ .- : h ....<>-tL. ........~
.~"h, UJ\.'a....:....a.~
.,s.. ')
~..;.......
/1-....,
~1..:>.....u
"""-- " - tv-d... .

~o.~
....
~& oIl~
Iv
e...... <l
I ..
'I
j."
1.0 ........ '}-
~-"
....
t:f..,..Jl '.., ~ ~n~ ,~~~,~ ")~

~
~ ~~;:..:....
..u..-.>
~ ~ ,~ ~ 0. (~
'" --" , ,t. .....l..,QQ

,
"'0 ~. :. a.I!..
!" ,v.
t-
....,
' ;: -, S ~
;:::] .......... '"--' , t!.&JlIlc..-All tt:..
__
'-"r'""1
Q
.d
'
.....
...
.....-1
?-o .. r """;'"
<;> ...........
~ t.r-,~i\,. ~
e..............
1t..;.~......,t.:\'1~ ... ~.
" ...,c.. ~
.....,...;:>
l-'>- 'k-- t ...w...~9
t:t...-t..m (."4
*'i _
e::::4:
.h.lW. 81
..
(1
&1 _ .. . ...... ,c:.JCt:
.....:.
~.........."
"...r;,,~ at-.. tt- _--:- ..... .
_u
4- ~~ ~
;J,.A.
dt. .. ~
. ~"'7 '
....
-r--' rca
"
; . '}-.~
. '>LU-
a:
il_
... '--
~.,.... ~
..J'JR..
a..
or
~ ~
OJ'
..
rl;.
..
r:
.;~~
(~ ,
I... t. f......, , .......1 .L.
f -
,' .
--
. i ..
I~
. -....L..t
c:L4A.
(\.
... Q.J.~
-...Q -
~Af,..~ ",Ii ,.J.A....~
..t l.11-A'L.., J...
'
....... 1..... '4
f-c -I.. .......

J,..
, J , ....I'1.J'...A
..J\-.,,t 'J... -." ~::'-J , q -t...
~ oJ _ _ ;."
),."
4-"> 1
.-~~ ~
~~
&Q..v, . ~
..... l--..r/..:_.. cL.
?c?=
...Jt ~P OM
.... .... _ ..... ..,_ ..J-'/t ..... ,~ ........... 4<t.,.l-"-:

1. "'-, ~ too- r ("'" ~1-"'" .' ~ 904 b
c..-.\o.
tt...
u:"..
"" . \ Tf
tL. .....~ (I
vL '''''. "~
in
i=tL:"",j
;';1:
0.....
4-iI..--
etL!WJ
.....
' .. ,
'<!..~J.. ........iJ...
,9.0 Q..aj)-.....
d... ~ ~,~ruJfl~~~ ~ ~
~
In
a...
~ f.&&~~lo~Ge~
(~U>.. ~
c}}-.. a12.R.. ..;., a
...Q.:..~~" k~ ~
-'9
..
to.-
ft
f?.. . .
I... ..... .n::h,
Lri",
'.IV\
h...o..
Plate 8. Part I Chapter IX Folio 22 (May 1901).
112
.......- Cl <-Ca., ') .
. ,.. ,,,
;. / ' - ~ ~"4
1,....-:4CD-fl-.
~~t ~
'c..--...-. ~
(A .A- ~
r .
H e..
e........
'f
......-( "1-' /L....
_.'-u
au...
a..,,~
'J ~ ~ ~u:k~ c.~~ e. . . . . . . -7- cL........, f -
~f'--~=-~("'~"uf)r
, ra.... ~~~....
~ JtL. '-.Ak~) ~ 't ;1). ~ ; 9-- ~ fk
a.
;, -'\- Fd....<. ~
~ J4 ~rc
)~
~~ d~ t-. ..t ... ,
t-S).
~~I'i-u-.~
""-~ '.~
.. ".
' I
(k
1-.'). "i~
(#t'1..<..C~ ~ ;l/:..c.Qf> ,. /I JJ- ~ ~ ( l ~ 7 t;J... ?~t(l411.;..J>

tud..:tPl'''''''.9~ ~ ~ \...Q~~~ -~ j..J .(......,\4.0
~~.:>It. :,~.0 ~ . . ."l..A..
. ~ o:.H.&. In ~ L (..d.o.uJI.. ')
--~' .,
c:.I.
)1
r- ri... ~v-t..
~ ~
0. r!a.:V,)
..Q. ....,.. __7- ~ ckJ~~ ~ 'l-~
~~ C,
0..
eu-.-.
~ ~ to.'1.c (2...,;
0.,,1 ...........A G.< ~ ;,... '~, " v~ ~ ~ 1.u J...." ................

.d).
0.
~J)....;......~).. 0-... ~ ~ t..,~:
~~ ~ ~e....o<1 t.. ~ c..-~ i--. "'""'"1 ~ (~
Il.
~ ~ 11~"'.~ ~r:.... ~~
t.o"4
Ci
(a.-w .
-,
""4. ~ t.r~
~ .~1......-.. ant..v ..:t.:.J .~:t... ~ ,~ ~U . .. _ _
. ).
r
........ a...:....
~" ' c*-:l
Or ,Q... a.t } L
Qco .. ,... .ti ti..
..5\...'IL... ..,~, ~
C"
't.--r ..,
C.-.
a ( ,.,
:>
it
1,
..
.5 ...
4'
I
,.4.,.. 'J~. 9~ ~
44 "",~I.. .
Plate 9. Part I Chapter IX Folio 23 (may 1901).
o......-r;...':].;....
113
gradually became aware of his inconsistency; he did not become concious

of it all of a sudden. For a while, he confused the origins of his inconsistency with Cantor's. Russell never claimed to have discovered Cantor's
inconsistency, although he did so unconsciously. As mentioned earlier,
Russell described his contradiction, for the first time, in a manuscript
composed in May 1901. Also, he publicly revealed Cantor's 'contradiction' for the first time in May 1903 in his The Principles of Mathematics.
There was yet another new element in this story indicating Russell
had a problem with the theory of ordinal numbers. On February I, 1901
he had written to Couturat thanking him for sending an article in which
Burali-Forti had proven that the trichotomy law did not hold among
transfinite ordinal types. On the other hand, Cantor had recently shown
(1897) that this law did hold among transfinite ordinal numbers. Russell
was now confronted with evidence of yet another potential contradiction.
In a footnote added at the end of the paper written in June 1901, Russell
briefly mentioned that Burali-Forti had denied what Cantor had positively
proved. Russell thought there was a possible flaw in Burali-Forti's proof,
namely the assertion that the set of all ordered types was an order type
itself.
Thus, by June 1901, Russell was involved intellectually with three
potential contradictions. First of all, Cantor's 1892 theorem contradicted
Russell's concept of a class of all classes or universal set. Secondly,
Russell's own contradiction, that was based on the class of all classes
which are not members of themselves. This contradiction forced him to
deny Peano's basic principle on the formation of classes, described by
Russell in his May 1901 draft of Part I. Thirdly, there were also two
theorems on the trichotomy law, one posited by Burali-Forti and the
other by Cantor, leading to, apparently, opposite conclusions. Prior to
June 1901, there does not seem to be any evidence that Russell took
simultaneously all of these contradictions into consideration while thinking about the foundations of mathematics. He seems to have failed to
grasp their importance at the time. He himself claims that,
at first, [he] thought that there must be some trivial error in [his] reasoning [concerning the contradiction of the class of all classes which are not members of themselves].
114

[He] inspected each step under a logical microscope, but [he] could not discover
anything wrong. l
Moreover, he does not seem to have communicated his results to anyone. 2

Needless to say, he probably felt very uncomfortable with these
'contradictions'. He was very close to finishing his book, yet he was
finding new problems and no answers. In his autobiography, he says that
the setback he suffered when he discovered the contradictions in May
and/or June 1901 was as severe as the one he suffered when he witnessed
Mrs. Whitehead's heart attack in early 1901, but on this occasion it was
of an intellectual and not an emotional character. Russell described the
discovery of his own contradiction in the following terms.
Cantor had a proof that there is no greatest number, and it seemed to me that the
number of all things in the world ought to be the greatest possible. Accordingly, I
examined his proof with some minuteness, and endeavoured to apply it to the class of
all the things there are. This led me to consider those classes which are not members
of themselves, and to ask whether the class of such classes is or is not a member of
itself. I found that either answer implies its contradictory. 3
The reader should examine this quote carefully. It is clear that the first
four lines are concerned with the inconsistency of the greatest cardinal
number. The rest of the paragraph is concerned with a second inconsistency: the one Russell claimed was his own. The contents of Russell's
manuscripts and correspondence that have survived seem to support Russell's recollections if they are taken literally. The remaining folios of
Russell's The Principles of Mathematics of November 1900 indicate, as
the quote given above suggests, that Russell's examination of Cantor's
proof (done originally to avoid the inconsistency of the greatest cardinal
number) led him to develop the formulation of the concept and then to
formulate the 'contradiction' of the class of all classes which are not
1. Russell 1959a, 76.
2. There are some undated letters to Stout, his former philosophy professor, kept at The Bertrand
Russell Archives. A solution for Russell's contradiction was discussed in these letters. Unfortunately,
we lack internal or external evidence indicating whether this correspondence took place in, or immediately after, June 1901.
3. Russell 1967a, 147.
115
members of themselves. I agree with Coffa that the process of discovery

was not an immediate one. The sources show Russell talking of the existence of mistakes (as he said to Couturat) in Cantor, since December,
1900, at the earliest. On the other hand, Russell may have become convinced that there was nothing wrong with Cantor's proof when Russell
wrote his paper "On well-ordered series", which I believed probably
occurred in June, 1901.
It may be that Russell originally thought, as he said, of the class,
but not the contradiction, of the class of all classes which are not members of themselves, as a consequence of his detailed examination of Cantor's proof. Unfortunately, as Grattan-Guinness claims, there are no surviving sources, explicitly supporting this conjecture. The sources available do not seem to indicate, at least directly from the text, any relation
between Cantor's proof and Russell's original formulation of his contradiction contained in the first draft of Part I of The Principles of Mathematics written in May 1901. However, I will argue (as others have done
before, including Crossley 1973a, Grattan-Guinness 1978a & Coffa 1979a) that Russell became interested in such classes (those which are not
members of themselves) from his discovery of possible 'fallacies' or
'mistakes' in Cantor. In this regard, there is another relevant quote from
Russell, written in some of his later recollections.
I was led to this contradiction [of the class of all classes which are not
members of themselves] by considering Cantor's proof that there is no greatest
cardinal number. I thought, in my innocence, that the number of all the things there
are in the world must be the greatest possible number, and I applied [Cantor's] proof
to this number to see what would happen. This process led me to the consideration of
a very peculiar class. Thinking along the lines which had hitherto seemed adequate, it
seemed to me that a class sometimes is, and sometimes is not, a member of itself.
[... ]. The application of Cantor's argument led me to consider the classes that are not
members of themselves; and these, it seemed, must form a class. I asked myself
whether this class is a member of itself or not. [... ]. [E]ach alternative leads to its
opposite and there is a contradiction. I
Russell was simply led by Cantor's proof to consider the class of all
classes which are not members of themselves. The contradiction originated solely from this class itself on questioning whether or not such a class
1. Russell 1959a, 75-76.
116
was a member of itself. In each case, the answer implied its opposite,
resulting in a contradiction. Russell first concluded that this class denied
Peano's basic principle stating "that every proposition containing only
one variable is reducible to the form 'x is an a"'.1 On further analysis,
perhaps unconsciously, Russell realized that he could derive a contradiction without relying on Peano's work. The formulation of the class in
itself was contradictory. Russell's first surviving account of his discovery
was addressed to Frege, about a year after it took place (June 1902).
Russell explained to him that he applied his special relation (between all
objects and all classes) to Cantor's proof, finding that the class of all
classes which are not members of themselves "was left over, even though
all classes had already been enumerated."2 But I shall warn the reader
that this argument is not a contradiction in itself, but only an indirect
proof. If all the logical steps followed by Russell were correct, then it is
impossible to establish a one-to-one correspondence between all objects
and all classes. It is possible that Russell originally conceived of such a
class as a result of applying his special relation to Cantor's proof; but at
this stage, the class was not considered to be contradictory in itself, as he
would claim later on.
Russell considered his 'own' contradiction to be the simplest and
most primitive. The formulation of the contradiction involved very simple
technical elements. In order to generate this contradiction one needed
only the concept of class and the relation of pertenence (). By contrast,
the contradiction of 'the greatest cardinal number' involves, at least, the
concept of set and power set, and the 'Burali-Forti paradox' involves part
of the theory of ordinal numbers.
By June 1901, therefore, Russell had a new challenge: The newly
discovered set-theoretic paradoxes had to be solved before The Principles
of Mathematics could be presented to the publishers. At the beginning he
thought this a trivial matter but gradually he would realize what an enormous task it actually was.
1. Russelll90li Part I Chapter IX Folio 56.
2. Letter: Russell to Frege, Iune 16, 1902, printed in Frege 1980a, 134; and van Heijenoort 1967b,
125.
117
4.5 A last effort to finish The Principles of Mathematics, April 1902 - May 1903.- After finishing Part II of The Princi-
ples of Mathematics and his article for Peano's journal, and teaching his
lecture course on the philosophy of mathematics at Trinity College (Cambridge University), it is reasonable to suspect that Russell spent part of
the second half of 1901, trying to find a solution for the contradictions,
especially his own. Unfortunately, almost at the same time, he suffered
another terrible emotional setback. As he described in his autobiography,
while bicycling one afternoon, in early 1902, he suddenly realized that he
did not love his wife Alys anymore. The misery and the agony suffered
by both of them is reflected in a diary kept by Russell at the time [see:
Russell 1902g]. The ensuing depression and unhappiness that followed
had a negative impact on Russell's intellectual activities, in particular in
the writing of The Principles of Mathematics. He felt that he was not
going to be able to finish the book properly, and that he would simply try
to patch together something suitable for publication.
In April 1902, he sketched another table of contents of Part I [see:
Table VIII]. Russell needed to reexamine Part I carefully: The contradiction of the class of all classes which are not members of themselves had
not been solved and the principles of classical logic had to be scrutinized.
It was probably around this time when he wrote to Peano asking his opinion on the matter of his contradiction. But the Italian master did not reply.1
Russell's general methodology of writing at this time is apparent in
the revision of Part I: after he determined the necessary changes, Russell
inserted most of the folios of the May 1901 draft in the new draft of May
1902-only those folios requiring revision were changed. The global
thesis of the book remained the same: To prove that pure mathematics
follows from mathematical logic. This time, there was, at least, one major change evident from the Table of Contents of the new draft, a special
chapter entirely devoted to the treatment of Russell's contradiction.
1. See: Letter Russell to Frege June 16, 1902; van Heijenoort 1967b, 125 and Frege 1980a, 131.
118

Table Vill. Part I. The variable (April 1902).
Introduction.- Russell drafted this outline of Part I of The

Principles of Mathematics in April 1902 (now kept in File 230.0303350). Nevertheless, this section of the manuscript would be extensively modified again, largely due to Russell's discussions with
Frege.

Book I. The variable
Chapter
"
"
"
"
"
"
"
"
"
I Definition of pure mathematics

II Symbolic logic
III Implication & formal implication
IV Proper names, adjectives & names
V Denoting
VI Classes
VII [The variable] Assertions
VIII The variable
IX Relations
X The contradiction
119
Russell described his contradiction in terms of the predicates which

are not predicates of themselves. The same analysis, even the same folios, were taken from the previous draft. He mentioned he had discovered
his contradiction when he was trying to
reconcile Cantor's proof that there can be no greatest cardinal number with the very
obvious fact that the class of all terms (which we have seen to be essential to all
formal propositions) has necessarily the greatest possible number of terms. I
But, again, Russell confused the derivation of Cantor's 'paradox' with

that of his own inconsistency. It can be inferred, reading between lines,
that Russell used his analysis of Cantor's second proof to deduce first the
class of all classes which are not members of themselves and then his
'contradiction'. In this section of the book Russell stated his contradiction
in a number of different ways to provide the reader a survey of all possible implications. Nevertheless, the strongest and simplest form of the
inconsistency was the one in terms of classes. He also discussed some of
the possible solutions without going into too much detail. These solutions
were rejected, because they would have refuted many of the ideas contained in his new book. The most reasonable solution would have been to
reject the concept 'class of all classes', but, according to Russell, the
concept of 'all' was fundamental to mathematics.
Still, the manuscript remained incomplete. According to the table
of contents that Russell had drafted in May 1901 (when he had attempted
to finish the book), the work would include a final part-Part VII-on the
concepts of 'matter and motion', or the principles of dynamics. At the
beginning, he thought it necessary to rewrite the previous version completely, which was contained in the last draft [Russell 1899h]. Upon
reading again the previous version, however, he decided that it required
only a few corrections and probably of a very superficial nature. At the
time, while going through a temporal separation from his wife, he was
corresponding with her almost every day. On May 24, 1902, he informed
her that
1. Russell 1902fPart I Chapter X Folio 197.
120

[you] will be surprised & amused, after all my talk of 2 months, to hear that I
finished [The Principles] yesterday. I found that a pile of old Ms, which I had
expected to have to re-write, required only a few additions & corrections, so I
arrived at a sudden termination. I
This 'pile of old Ms' (as he called the earlier version) belonged to
the previous draft that he had attempted to finish a few days before the
start of the First International Congress of Philosophy in mid-19oo. By
that time (ca., June 1900), Russell had already concluded that the main
questions relating to the basic problems of the foundations of dynamics
did not belong to physics or dynamics, but to pure mathematics. If so, by
May 1902, according to Russell's new thesis, these concepts could be
derived from logical constants and definitions. Russell explained that he
had conceived the analysis of rational dynamics as a branch of pure mathematics not related to observation of the actual world. Most of the revisions consisted of brief explanations of how concepts traditionally thought
of as belonging to dynamics, could be understood in terms of logical
constants. For example, the idea of 'occupying a point or an instant' was
explained in terms of 'relations'; the concepts of 'time' and 'space' were
replaced by one-dimensional and n-dimensional sequences, respectively.
The notion of 'motion' was expressed in terms of 'continuous sequences'.
According to Russell, all statements of dynamics were capable of being
translated into the language of pure mathematics and, therefore, into that
of logic also. Thus, he concluded that
the a priori truths involved in Dynamics are only those of logic: as a system of deductive reasoning. Dynamics requires nothing further, while as a science of what
exists, it requires experiment and observation. 2
So, by the end of May 1902, Russell had finally finished his book
on the principles of mathematics as planned. It contained seven parts and
supported his thesis that pure mathematics could be derived from the
principles of symbolic logic. Russell's 'contradiction' remained unsolved
and Russell, as far as I know, had not heard from Peano expressing his
thoughts on it. Now, Russell decided to write to Frege, the great German
logician, philosopher and mathematician. Russell had become aware of
1. Letter: Russell to Alys Russell, May 24, 1902.
2. Russell 1902f Part VII Chapter V Folio 59.
121
Frege through a review of Frege's work written by Peano. Russell had

recently begun to read some of Frege's works, and he had found that
they shared much in common. 1 Of particular interest, one of Frege's
major objectives was to show that arithmetic could be derived from logical constants and definitions. Perhaps Russell thought that the most appropiate person to judge the value of his inconsistency (and may be the
most capable of finding a solution for it) was Frege. On June 16, 1902,
Russell sent him a letter containing the following description of his contradiction:
I find myself in complete agreement with you in all essentials, particularly when you
reject any psychological element [Moment] in logic and when you place a high value
upon an ideography [Begriffsschrift] for the foundations of mathematics and of formal
logic, which, incidently, can hardly be distinguished [ ... ]. There is just one point
where I have encountered a difficulty. You state (p 17) that a function, too, can act
as the indeterminate element. This I formerly believed, but now this view seems
doubtful to me because of the following contradiction. Let w be the predicate: to be
a predicate that cannot be predicated of itself. Can w be predicated of itself? From
each answer its opposite follows. 2
The answer came almost immediately, showing Frege's disturbance at

what Russell had revealed to him.
Your discovery of the contradiction caused me the greatest surprise and, I
would almost say, consternation, since it has shaken the basis on which I intended to
build arithmetic [ ... ]. It is all the more serious since, with the loss of my Rule Y, not
only the foundations of my arithmetic, but also the sole possible foundations of
arithmetic, seem to vanish [ ... ]. [Y]our discovery is very remarkable and will
perhaps result in a great advance in logic, unwelcome as it may seem at first
glance. 3
Russell mentioned to his wife that he had received 'a most candid letter'
from Frege. This letter probably finally convinced Russell that his discovery was certainly not trivial and, likely, very important. And, as a
consequence, he then tried to find a solution.
1. The reader should know that around this time (ca June 1902), Russell was reading the main logicians (including Frege) for his own book on logic [see: page 85 note 3].
2. Letter: Russell to Frege, June 16, 1902; printed in van Heijenoort 1967b, 124-125 & Frege 1980a,
130-131.
3. Letter: Frege to Russell, June 22, 1902; printed in van Heijenoort 1967b, 127-128 & Frege 1980a,
132.
122
By the end of May 1902, the manuscript had been sent to the
publisher and Russell had already received some specimen-pages. He
decided to expand the chapter on 'the Contradiction' as a result of his
correspondence with Frege. Some pages of this chapter were heavily
revised in proof. Russell also decided to alter some paragraphs of earlier
parts of the book and add two appendices: One on Frege, explaining his
work, and another one on a possible solution of the 'contradictions'.l In
the first appendix Russell mentioned that Frege's second volume of Grundgesetze der Arithmetik (1903) was already in the bookstores, and that it
included a discussion of 'the Contradiction'. In the second appendix,
Russell presented a tentative solution, that he realized needed "to be
transformed into some subtler shape before it can answer all difficulties" .2 This solution was devised in terms of a theory of types, ordered
in a strict hierarchy, that did not allow one to consider classes of classes
as being of the same type as the simple classes. Nevertheless, Russell's
theory of types did not eliminate all of the mathematical contradictions;
although it provided a solution for Russell's 'contradiction', it did not
solve Cantor's 'paradox'. The last sentence of The Principles of Mathematics reads,
What the complete solution of the difficulty [the contradiction of the greatest cardinal
number] may be, I have not succeeded in discovering; but as it affects the very
foundations of reasoning, I earnestly commend the study of it to the attention of all
students of logic. 3
The preface of the book was written on December 2, 1902, and

received by the printing house about a week later. In it, Russell stated
that the book had two main objectives: To prove the logicist thesis, and
to explain the fundamental concepts of mathematics that were accepted as
indefinable. He also pointed out that the contradictions implied that something was wrong somewhere, although he had been incapable of finding
out exactly what it was. He also mentioned that a second volume was in
1. Both appendices were received by Cambridge University Press on November IS, 1902. See: Russell
1902f Appendix A Folio no number.
2. Russell 1902f Appendix B Folio 1.
3. Ibid. Folio 9.
123
preparation, with Whitehead's collaboration, which would be aimed exclusively at mathematicians. l The preface also contains some important
notes for those interested in the history of ideas. First, Russell stated
accurately that he had been almost completely ignorant of Frege's work
at the time that The Principles of Mathematics was about to be printed.
He also described briefly the historical development of his ideas, beginning with the study of problems in dynamics, leading to a reexamination
of the principles of geometry which forced him to analyze the philosophy
of 'the continuum' and the 'infinite'. This in turn motivated him to study
the principles of symbolic logic. In a note, which is not present in the
manuscript, he excused himself for publishing a book which included
unsolved difficulties. Finally, in January 1903, he added some folios to
the manuscript in which he summarized the full contents of the book.
Thus, it seems that by January 27, 1903 The Principles of Mathematics was finally complete [see: Table IX]. But allow me to consider
again the intellectual background of Cantor's contradiction (on the greatest cardinal number), and Burali-Forti's contradiction (on the greatest
ordinal number). I believe that Russell wrote the English version of his
paper "On well-ordered series" in June 1901. It was published later on in
French in Peano's journal. About this time, Russell became increasingly
convinced that Cantor's theorem was not false, and that it implied a
mathematical contradiction; furthermore, there were theorems, one by
Cantor and another by Burali-Forti, which proved, apparently, opposite
results. Even though this article was written around June, 1901, its material was not included in the version of The Principles of Mathematics
being written at the time. In that particular month, Russell was writing
Part II of his book where he discussed the notion of 'number'. This
second part, however, does not contain a single reference to the contradictions. A question then arises: when did Russell incorporate the references to the other contradictions into the text? It is possible that the other
contradictions were added to the text on May 26, 1902, when he made
1. As soon as Whitehead had read some of the page proofs of The Principles of Mathematics, he and
Russell began their collaboration. Beginning in September, 1902, Russell often mentions that they were
working together on a project. Probably, Russell was very pleased that his former professor, who he
admired greatly, was now his colleague [see: page xv note 1).
124
Table IX.A 'Final' draft of The Principles of Mathematics (January 1903).

Introduction.- The following Table of Contents outlines the contents of the 'final' draft
of The Principles of Mathematics, finally published in May 1903. The manuscript was largely
completed by May 1902 but there were other folios added in later dates and some of the changes
were, presumably, introduced in galley. The manuscript is also incomplete-there are some
missing folios from Parts IV and VI. The second folio of Part VII Chapter I indicated that it was
probably written in 1899. Another folio (Part VII Chapter III Folio 30) is dated June 1900 on the
left upper comer, indicating that Russell originally wrote this section few days before leaving
England for The First International Congress of Philosophy held in Paris, where he met Peano
for the first time. Folio 76 (Part VII Chapter VII) is dated May 23, 1902 (on the right lower
comer) indicating the date that Russelll made a revision. In most cases, in which a folio contains
a double numeration, it indicates that that particular folio originally belong to a previous manuscript. Finally, the file number is 230.03350 and it is divided into twenty folders.
Preface
Part I
Chapter
Part II
Chapter
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
The Indefinables of Mathematics

Definition of pure mathematics
Symbolic logic
Implication & formal implication
Proper names, adjectives and verbs
Denoting
Classes
Propositional functions
The variable
Relations
The contradiction
Number
Definition of cardinal number
Addition & mutiplication
Finite & infinite
Whole & part
Infinite wholes
Ratios & fractions
(continues ... )
125

Table IX.B 'Final' draft of The Principles of Mathematics (January 1903).
Part III
Chapter
Part IV
Chapter
Part V
Chapter
Part VI
Chapter
XIX
XX
XXI
XXII
XXIII
XXIV
XXV
XXVI
XXVII
XXVIII
XXIX
XXX
XXXI
XXXII
XXXIII
XXXIV
XXXV
XXXVI
XXXVII
XXXVIII
XXXIX
XL
XLI
XLII
XLIII
XLIV
XLV
XLVI
XLVII
XLVIII
XLIX
L
LI
LII
Quantity
The meaning of quantity
The range of quantity
Numbers as expressing magnitude: measurement
Zero
Infinity, the infinitesimal & continuity
Order
The genesis of order
The meaning of order
Asymmetrical relations
Difference of sense & difference of sign
On the difference between open & closed series
Progressions & ordinal numbers
Dedekind's theory of numbers
Distance
Infinity & Continuity
The correlation of series
Real numbers
Limits & irrational numbers
Cantor's first defmition of continuity
Ordinal continuity
Transfinite cardinals
Transfinite ordinals
The infinitesimal calculus
The infinitesimal & the improper infinite
Philosophical arguments concerning the infInitesimals
The philosophy of the continuum
The philosophy of the infinite
Space
Dimensions & complex numbers
Projective [space] geometry
Descriptive geometry
Metrical geometry
Relation of metrical to projective and descriptive geometry
Definitions of various spaces
The continuity of space
Logical arguments against points
Kant's theory of space
(continues ... )
126

Table IX.C 'Final' draft of The Principles of Mathematics (January 1903).
Part VII
Chapter LIII
"
LIV
"
LV
"
LVI
"
LVII
"
LVIII
"
LIX
Matter & Motion

Matter
Motion
Causality
Definition of a dynamical world
Newton's laws of motion
Absolute & relative motion
Hertz's dynamics
Appendix A. The Logical & Arithmetical Doctrines of Frege

Appendix B. The Doctrine of Types
127
'final corrections' to the 'final' manuscript of The Principles of Mathematics.

However, there is yet another alternative explanation. I have suggested that Russell became fully aware of the (real) importance of his
discovery a month later when he received an answer to his letter to Frege
on June 22, 1902. This would imply that the additions concerning the
other 'contradictions' were added at a still later date, once Russell was
completely convinced of their importance for his own philosophical system. In conclusion, Russell may have included his discussion on the
'contradictions' of Cantor and Burali-Forti any time between June 1902
and January 1903, but it seems almost impossible to establish the exact
date that Russell incorporated them into the text.
There are contradictory statements in The Principles of Mathematics concerning the 'paradoxes'. For example, in the section dedicated to
'the Contradiction', Russell commented that his own contradiction was
the only one contained in the whole book. But I have already quoted a
passage from the book in which another contradiction was discussed. The
first passage was written in May 1902, and the second one was sent to
publishers in November 1902. Therefore, Russell probably finally concluded that Cantor's theorem implied an intractable contradiction, sometime between May and November 1902.
One of the most remarkable changes between the printed version of
Part V of May 1903 and the original draft of the autumm of 1900 is that
Russell originally appeared convinced that Cantor's theorem was false. In
the earlier version, he even attempted to find possible logical flaws in
Cantor's argument. But, when The Principles of Mathematics was published, Russell presented a completely different approach to Cantor's
theorem.
Cantor has given a proof that there can be no greatest cardinal number, and when this
proof is examined, it is found to state that, if u be a class, the number of classes
128

contained in u is greater than the number of terms of u, or (what is equivalent), if ot
be any number, 2" is greater than ot [ ]. The difficulty arises whenever we try to
deal with the class of all entities absolutely, or with any equally numerous class.!
This was the first time that the 'paradox' of the greatest cardinal
number was stated in print, and credit must be given to Russell's for its
original publication. But there are other passages contained in the book
which are not found in the manuscript concerning the origin of the 'paradoxes'. The most interesting one, which appears as follows contains the
term contradiction.
There is a difficulty as regards the type of the whole series of ordinal numbers. It is easy to prove that every segment of this series is well-ordered, and it is
natural to suppose that the whole series is also well-ordered. If so, its type would
have to be greatest of all ordinal numbers, for the ordinals less than a given ordinal
form, in order of magnitude, a series whose type is the given ordinal. But there
cannot be a greatest ordinal number, because every ordinal is increased by the
addition of 1. From this contradiction, M. Burali-Forti who discovered it, infers that
of two different ordinals, as of two different cardinals, it is not necessary that one
should be greater and the other less. In this, however, he consciously contradicts a
theorem of Cantor's which affirms the opposite. I have examined this theorem with
all possible care, and have failed to find any flaw in the proof. But there is another
premiss in M. Burali-Forti's argument, which appears to me more capable of denial,
and that is, that the series of all ordinal numbers is well-ordered. This does not
follow from the fact that all its segments are well-ordered, and must, I think, be
rejected, since, so far as I know, it is incapable of proof. In this way, it would seem,
the contradiction in question can be avoided. 2
The folios of the manuscript are complete in this section and do not
contain this quotation. In fact, the manuscript does not contain sections
299, 300 and 301 of the printed version, suggesting that they were probably added in a proof stage. On the other hand, Russell uses the word
'contradiction' three times in the passage quoted above, although he referred to different 'contradictions'. First, he commented on Burali-Forti's
conclusion of his reductio ad absurdum procedure which demonstrated
that the trichotomy law did not hold among transfinite ordinal types. In
the second instance, Russell stressed the fact that Burali-Forti's and Cantor's theorems contradicted one another. Hence, in fact, he was talking
of two contradictions. One of them was derived by Burali-Forti himself
1. Russell 1903a, 362.
2. Russell 1903a, 323.
129
which allowed him to deny the trichotomy law for ordinal numbers. The
other showed that the two theorems led apparently to opposite conclusions. Finally, Russell perceived a possible solution to Burali-Forti's
contradiction, by denying Burali-Forti's premise that the entire series of
ordinal numbers was a well-ordered series. But soon, other mathematicians attempted to show the fallacy of Russell's solution. Russell was
satisfied with his rebuttal of Burali-Forti for the moment. However, this
section of The Principles of Mathematics formed the basis for the eventual development of the Burali-Forti paradox. 1 Soon, other mathematicians
attempted to show Russell's fallacy.
The Principles of Mathematics was finally published in May 1903,
presenting Russell's 'contradiction', Cantor's 'paradox', and the elements
that would create the Burali-Forti argument. Although the mathematical
community had already become acquainted with the 'contradiction' of the
class of all classes which are not members of themselves through Frege's
appendix to the second volume of the Grundgesetze der Arithmetik, it is
very difficult to determine how widely and effectively Frege's version
was known. It appears that Frege's book was read by some of the most
important mathematicians of his day (e.g., Cantor and Peano), but it had
very limited influence and impact on his colleagues. It is then reasonable
to suppose that very few scholars knew of Russell's 'contradiction'
through Frege's previous formulation. Moreover, Russell's book was
published only a short time after Frege's book. Possibly, mathematical
communities did not have a chance to become familiar with Frege's
Grundgesetze, even those within Germany. On the other hand, Russell
had discussed his contradiction with several people before the publication
of The Principles of Mathematics. The people included, among others:
Frege, Peano, Whitehead, Jourdain, Couturat and Gilbert Murray (18661957), a close friend-expert on Classic and British literature.
It is not surprising to find that Russell was disappointed with the

book, even before it was finally published. Letters to colleagues and
friends show his dissatisfaction, probably because he was unable to cope
with the inconsistencies and provide a solution to them. He had been
1. Moore & Garciadiego 1981a, 330-331.
130
attempting to find a consistent explanation of the principles of mathematics for some years. When the book was almost complete, however, he
encountered certain contradictions which seemed even more difficult to
explain and solve than the earlier paradoxes (orland philosophical antinomies) that had been explained through Cantor's work. Russell expressed
his feelings as follows: "My book is nearly out, when it appears, I shall
feel as one does when one leaves the dentist. I do not think the book
good, but it will be done, and that is enough. "1
1. Letter: Russell to Helen Flexner, February 7, 1903. See, also, Letter: Russell to Alys Russell, May
16,1902.
The [semantic] contradictions [ ...] are not purely

logical, and call1lot be stated in logical terms alone
[ ...]. So they may due not to faulty logic or mathematics
but to faulty ideas concerning thought and language.
Frank Ramsey.
Chapter V.
The 'Semantic Paradoxes'
5.1 Introduction. - The traditional approach to studying the
origin of the so-called non-Logical 'paradoxes' was to treat them as
immediate consequences of the logical ones discovered by Russell. 1
Since most historians associate the discovery of these new 'paradoxes'
with scholars such as George G. Berry (1867-?), Zermelo, Jules Konig
(1849-1913) and Alfred Cardew Dixon (1865-1936), among others, it
would be reasonable to suppose that Russell's role was simply that of a
guide. However, subsequent research has shown that his role in the
formulation of the new 'paradoxes' was much more complicated, and that
he was, in fact, directly involved with the origin of some non-Logical
'paradoxes' .
Throughout the subsequent discussion, the following points will be
assumed: First, that mathematicians and philosophers simultaneously
became familiar with the diffusion and metamorphosis of the Burali-Forti
'paradox' and the development of other inconsistencies; second, that
Zermelo's proof of the Well-Ordering Theorem (1904) provoked an
immediate debate involving French, German, British, Italian and possibly
other mathematicians; third, that mathematicians and philosophers felt
confused about the origins and implications of the 'paradoxes'. As a
result of this confusion, some mathematical propositions were treated as
'paradoxes', even when they were not originally considered to be so.
1. "Since the discovery of the [ ... ] contradictions within Cantor's theory of sets, additional paradoxes
have been produced in abundance" [Eves & Newson 1965a, 298; Eves 1983a, 475.]
132
The 'semantic paradoxes'
This confusion has persisted until the present: today some propositions
are still treated as 'paradoxes', contrary to the intention of their creators.
In this chapter, only the origin of those non-Logical 'paradoxes' arising
directly from Cantor's theory of sets will be discussed. Others (e.g., the
paradox of Epimenides), were directed by their creators at other questions, even though they have sometimes been discussed in terms of Cantor's ideas.
My discussion will focus primarily on events ocurring between
1903 and 1908. In 1903 Russell published The Principles of Mathematics.
In this book, as we have studied, he described the 'paradox' of the greatest cardinal number (Cantor's 'paradox'), his own contradiction of the
class of all classes which are not members of themselves, and antecedents
of the 'paradox' of the greatest ordinal number (Burali-Forti's 'paradox').
In 1908, on the other hand, Russell presented what he thought could be
his final attempt to solve the inconsistencies. 1 By this time, Russell
thought that he had already proved that all contradictions could be generated from a single general 'paradox' from which it was possible to formulate numerous specific 'paradoxes'.
In his [l908c] article, in which Russell presented his final attempt
to solve the 'paradoxes' (the ramified theory of types), he outlined certain 'paradoxes' that he intended to solve. These included: the "Epimenides paradox"; a similar to Russell's contradiction in terms of relations, as well as the inconsistencies proposed by Berry, Konig, Jules
Richard (1864-1921) and Burali-Forti. Russell did not analyze the 'paradox' of Kurt Grelling (1886-1941) and Leonard Nelson (1882-1927)
[l908a] (formulated in terms of two classes of 'autological' and 'heterological' adjectives) doubtless because he was unaware of this 'paradox' at
the time he published his paper on the theory of types. Nor is there any
discussion of this 'paradox' in Principia Mathematica, in which Russell
1. Some mathematicians and philosophers stated that Russell's solution seemed somewhat artificial and
not (fully) based on logic [see, for example, Weyl 1946a, 272]. According to earlier studies, Zermelo
made an attempt to solve the 'paradoxes' in 1908 [van Heijenoort 1967b, 199]. Moore, however,
argued that Zermelo published his axiomatization of set theory in order to rebut criticism of his Axiom
of Choice [see: page 139 note 2] and his proof of the Well-Ordering Theorem, rather than to attempt to
solve the 'paradoxes' [Moore 1978a, 307].
~'
So
~'
~'
~
S'
!"')
S'
'"
::l
0'
~.
::l
(')
CI>
'i::I
or
fA'"
INTaOOl'CT10~
(,,,
..
,
vt
'j'.
1.1.
... t ..... 1I1M",,"vt.... 11>111
Ih,.",. N-'I' -:L\tJ

tHUn .. n
U.O .. hflllu :' lr.NJ.,.. r .J,I .. 'uaJl1.II'.... ..-l ...... Jj
I."., ...... \..t
U) liurah F"rti"_ CUlltfR,h("ll"U HUn' lJ(~ .. 1..",,1 .h f,II'I\~fOt II ("rUl J

t'hfl",,~n .hal "'\"r\ ... ,II.nnJ.n . . 1M'rl,'''' hllo; n'n rtrdilidl 1IIIInh,,r. th;It I h. .... r;'.... t
"":bnal,, up t .. Rml wrlu,hng ioLIl)" )(1\..,' lO"hll.11 \l.n."t"tf'4 till' J.:Ht1I 11,,1111.11 1~~
I.n,., lind (fill t.rt.1m \~'FJ' nALural ....... Ullll'lIull!i) lh,IL till..' ",,"'rlt-'" "I "II vnlill"l ...
(In u"lt'r II! lIIa~"Hud\.', I. "rdlunh-I't'ti It lullv\\jt/ tltal 1111' "'l n' 'Ii .. ,. all
Ttu .,"
R ..nit -S 111;1)" tK'. H hi\... thot' nbtlcm r tv S" ,!!' ''I'U\ I,lIl hi It ,J.It~ h ..l
hA\'l' Uu- rvl"liull It tv ::;," Hi/on' J.,.'lYIII~ thl' \l4llh' t too 1... 11, U .1IIt! .....
"T hAlt Lh rl'l.llIun T ti. 7' " IIIi ""11II\',,,I.:n1 101" 1" .I... ~ IlCIL II,,,, rll., r"l..u""
(!l) ] .... 't T lA Iht nbt .. ,11 ",hleh ... nl, ... ""I ... 1"'1",,,11 1\1,,, .-i,l""II'" }( allil '"
",hl'nt'H'r H ,It"' .... nut Ii.nt Ih. J"1lal,ull U t .. .s. Th.lI, "kiln, r ,01111t.""
.. tD '" nuL r, II'
or
~pilllt:na"I'""I
(1) Tht .. Id''l '-"ntr.uh('tllll1 .,r 1111' 1..11111 1fI.I"t... I."" ,,. 1111 ' f./I",.".,,,,,,
th, ("rd;U1 .....lIfi th.lt nil ('rd.m .. ".'1'\' )/;\1 ... II1I1I .ill OIL!.. t ... 1.,
flifntJ. mDlI{' by t 'r,I;)n... \\, r,' ('lrt:u.,I,' Ii.", W,1,'f Ilu ... '.1 h. , Th. .. uI11,1 .... ' '"UII
th,-.cllnlmlh"u, ....... atfntlhd h, Ih"III,1I1 ~hll ... " ... " J .1111 hlllJ.: If hl ' l"
IJlng. he 1" ~1)(akinJ.: I", lnllh .. H;,I \ H'j' , . . . . . . . . . .
(21
y. bt, ,hI' d.\.~ ..I' nil lh ..~ l"I. ......... '1'Io "hu 'lI .H, 11111 "I"lul,. nc
lh(lm~ht1t. Tht'n. \If hnl"\'~'r d:I"~ ,I 1I1;l.' 1" ,... J" j ... ,I w ' ,..... " .. \.111111 III ,J I ..
nl,l. an T." ~.('n(", . t:i,ill~ til f Ihl' \;illh"
" r" I" .1.,. I" t-'t"n.ll. 111 I.,
In cx-pl"m Iht'" nU'r,'I~' ;.... ,h. noiluh HJ .... !H 11J.t:llllllall ' u... "I Ifl'~' ull'n ..
Tht' ~llutl'lO IUU ... t tl\' "oCHlJ:ht iu .."till "\1('" "-('rutln,\ .. I IlItIIl ...... III,.1 r"~I4' ll
ldea1l as ha~ bl't'n Rttf'IIII'it",lm Ih. f..r.It:"I1l).:" 'lrl ...... '... ,
\",111. rift' nmtmrltr/'u,u,

We nn' no.... In ~l p~llI"fI 101 l'oh .. " hI!'" th" th .... ,r~ .1' 1,~ I"'" "tll'1t,. ! hi'
!IOlutlOU .,( th(' ("ltntnuJICUt.n"l '" hi('h Im.H h,.... ,t 111 .... 1 h'III.11 i,'n1 I":.:,i,'. "~Ir till"
purpt~. a'e- shn.11 ~WU h~ nn ."IIU11~i.'r'ftII"1I llJ ' -1/11" .,t til' 111',1'\' IUlI .. lrI.11I1 ,IIIfI
illu~lmtl\t:' u( tht~, c..nlr"lhdl.m",. ;\11,1 "lMlIlht'1i .. h.. " hIP" Ih \ ".11 t-IUI ... I~
Yiclou"'-eII'C'I~ raJLI{,ll"". ",ml r"t' .hlrd"r, ~dl ;"101(1..,1 hy I h,"' II",,, ~ 1I11~" I.'M. II
vl'Ili bt" n~tiCt.. llh.1l (h. ..... , ptnld,');,.... 1.. 111.1 nt.h' l'vh'''I\.I,~ 1'1 Ihl " 1,1~ ;L.... t
numb,. 41H'i ~'U:\lIlI' l ..\("t ...... lingl~, lin ..... "11111111 (';ll1 I.. . ~I. I. '1".111 ' '" hl..-II ~. L...
at. i. by no meUIIJ. improbable thRt U Jllihnulll ~. filumt to bo .1 ... lu.jh!. frilin

lOme nther more rundll.tueuLaI and mon' ("-14I"'lll U"IUIU, II ..,. 1'II~ ..ibl." Ih'lL II..
u. th!' vi(,lftulI-CIrc:It" princlpl,', ru; cmbt.uil'd In the ".Ilfn, I,i,f;"'(h~ fir l,.IIIt ...
i.t more dr-.lIc lhltn It. nt-t. . 1 bt-. ,MI lhnt b) It h~ clr.l.'\uC' 11M' .f"., "I'"t'ftloolll
ror the &I1Ofn mIght bt a\olrl..,.1 Such ChMIJ.:\."", h""I.'\"t"t, \\ltlll.J nut ruuil 'r
&DfltJlog
whlt'h hAd h."tn t\S.~T~1 un th ... Ir.;"' .. lIf Ih. fJrIIl4'iptnoio c'I'I.IIII,,,d
above they ",nltl,' 1I1<'n'ly pn)vHit e..... i~r p"""f.. "I lhl ' "IlI.!lII'Hrt'n,,," 1'h.,!'i'
_ould let'm. thl.'I'\(fof\>, IQ t,.. lUll I ht" ~hlillt ........ t 1o: .... lllllfl '~.r II<al"w~ tilit ,hi'
11M or the a .. IOlU l,r n'tlut"ibllity In'l~ 11';,,1 U~ 111141 ,-m'l
60
J::O-C\H.R.\TIO:\
or CO:'l>TLlIiICTlo",s,
tillll, nllll1l .. r
"I "')-'11:11110".
lid, ....
tho
rlallu",
'.1\""
lit ..... 1.10" Hlh'gt1"& O\u"',
1"11'11'11... 1
I.ul
HUIIII .. r .(
.... ltll,
\0,.
hilln ,h. ulh ,1"1IIi.11 " '\"rll" k~ ", h.l\! ,kt;II,,1 X lII.r hmh'
"rd ... nll.lth. ,d .. r' Y .11;.,:1.t ... 1,, .1 ItltIIIt... t .. t f: Thll .. .\" bo.th
,,, 1111 I" ,".1 ,I ""Iul . , 01 ",'
I h ,,11 III. ,,1 ...\'., .""llr,I.JIIC '''II'' ,,,IIll"h .In' 111.-.).h- MI."("th'"" r",.tll .In
1IIIIdililit ntllilito t I tll,r.' I .. a ''''111111,,11 ,h.lr:lt""'''u.,. "l~hh '\to III.I~ 11\~nh.
,~ ",..11"'1,'1,,.." ur 1\,,1.'1 " lIl ...... . Th. fIIII.H~ IO( EI"III"III,I--" mil .... 11 ... 111. 1.
)(",...11 III II .... ~II ....... I. 1I"lld.I"l"'~ I'rll\. \I 1''11 th.'~ lin ' .... I 1II'IIII ... r- ..1.h.'I11M" .... 1111"',.... f H' , Ihl'" Inll"" .. I.. " "1'1,1~- III !IC' , "lud "flUlIIII~ tur Ih.
h .. tll
.. I I.' "'IIV'
li"':IIt. Iii \' I.'/j ... 11111 it II" . '11
llll'II.\' 1"'lhttol'"I'r"IIi.~U tht"'"lHlm'bt r.,

,,1..11.\., JiIlJl. \.dll' II III'~ h.I~' ,h" i.11I h~ur.' 111 X I'" .htf,r-Ill
III" ,.lIt 110:111" 'II II" , ohllr .. 1 Ih. d. -Itn"ll ... /.. IIII ...... 'I,/t: t". illftl til' n flO"' .\'
dUI:11 \"llId. 1".1 ,.. ul'-ldll.:"''''I;

~; I ICldul,r .. I~H .,.I" ,j I... ;I~I;I I"lh.u. III tiLl' 1",,"'1 u .. ldin-Ibll ' l.nlln.,1. II
I'" "... t"II,,~ .. C II ... ,.I, r ,III Ir.'t'III1 . I ... h". ",.11 I.. , ,ldin,.I~) UI,.U\';' "t ., tunt.
l'UI'Ib. r "I '" 'Its; , 1, , f; I.. ,h. d., ... t "'1111, 11'",'III1"I'\, 1'11," J:: h..... N. "'rIu"
h"lw, ' .1 .. 1111 '1111 ... ,,, r,u ,.. ",.1. I. ,1.,,, Ih. 1 .. , :!II,I, ;lfll,
1... 1 .\' 1.. ,I nllllllN(
.Idill ... l II" f.,I1 ....... II III" ,,111 ti;':IIf' lit d .. ~th ,1"'1111.,1 I~ /', l.-t ,hI,' Nlth
.1l1.,". . 1 lilli'
" .. 11 ...\11.11,1. ..... ,,,.1 "Un""!.: th'"'4lh, ... kUI"1 .,. .1\
Hq1(.... oolh~
I."-.. t UII"":"r "'" U;HII. ~.t'.Jt. iu h \'o"r thnn 1I1Il"k"n ...\Jhhl,"t-" fIIn ....1"IU.t, l\
ddlult ~ lilt.'!.,:,, . III '.II'!. II dHI+lr-... 11Ij;; Hln "thl.' 1l'.L.... 1 mt.g.r n.. t
Jt:\lII .... II.I., III I"\\"'r t h ,It lilli' t ... 11 ",~H"hl.~' I.... 1t~1f ..... un, t"IIII"'I"'ling .. r
l"i..:hl''11 !'o.dtll,r .... , II.t hI' 1., .... 1 In"'"\.:,r 1... \ 'I-Itlh.,bl,. HI (t"~"r Ih.1Il nllV',
""'11 ,.\11.1.' .... " , Ih I.. ,.... 111.,1111 "..:jlt,.II ... ~II.,LI.- ... "huh I-.;t ('ulI\m.Jwtl O.
It,. \IlI'ofl;': t 1:1,. ... hllll. ,,"'hll,~I ... ,.IIn. .1II , ... 11,nll'''1. .. hill' ulh,r.... t"an !tut .
I"" tlw ,otla' 1IIIIU .... '1 1 1.......... 11.1., .1. IIIIH.",, I~ N ...... "hll. 'h, mlllll",r lir lnUl".
hlllll' l.fll.lIll .. """""11 .. No, 11'11., Ih",,, IIm .. t b.. Ullldin.,l.h ,,"llIIltl,,.. "wi
.UII'h": Ih, .... , fh"n 11111.-..1 I.. . ,fl'.,... , B'H tlll"'I" d .. hlll ... 1' I.... oo .hl' k.1;.,,1 imlt hll.lblt
or
61
.,..iIlMI .. h;\.~ ,In 'nlih;ll nUlllh_ 'r. n ..."

But. III th..H cn."I4." (hi' -'flt"" CJf 11111
' '1'1 ... :\1 .. IIld ll-ll1llo: n ha..., 'hI' .. "lln.,1 ~1'1lI1)('r U -+ ' . ~hlf'h IUnt b.. greatt!r
tlull n iI"f~'" U ,.. !I l th .. 41 .... 11,..,1 IlHlIID.t .,f nil unllluh
1.-" Th~ 11111111 .. r ~.t" ...,YII.\hl ... In l " ... ~:nJ.::li .. h IUlHl .... !If l1ult( tnlflgr"
,.",1 ... 141 IIlq..l.... ,1."1 Ih, ' Inl' ~. f'OI gr-'''-' hfJ:',r. 1''1111 rull ... , "'f'JDItu,,lh Inrr'f"~'
IlIlkhlllli-ly :>oolllf'."' un I.) II hnll.' H1111\'.,.'r III" n.'fIIl"" ('art b.. uwii' .,.fh " glY"n
III
W
W
......
'"
><
('1)
CI>
s-
R<>
~
'"
134
and Whitehead again proposed the ramified theory of types. Since Russell
apparently was completely unaware of the existence of the Grelling and
Nelson inconsistency, no attempt will be made here to analyze its origins.
On the contrary, the Konig-Zermelo 'paradox', which was not discussed
in Russell's [l908c], was included in this book because it was mentioned
in the debate between Russell and Poincare.
5.2 Berry's argument.- The first non-Logical 'paradox' was

discovered by Berry in about 1904. He described his 'contradiction' to
Russell in a letter [see: Appendix pages 167-168], but Russell did not
publish a modified version of it until 1906. 1 Little is known about Berry.2 He was working part time at the Bodleian Library (at Oxford) without holding an academic position when he met Russell. According to
Russell, Berry was the only person there who knew anything about mathematical logic.
In his letter to Russell, Berry claimed to be disappointed with The
Principles of Mathematics and Russell's articles published in Peano's
journal Rivista di Matematica, because Russell did not prove that the
ordinal numbers could be well-ordered. This point is quite important to
understand the background of the 'paradoxes', especially the Burali-Forti
'paradox', because the absence of a proof allowed Russell, originally, to
avoid the Burali-Forti 'paradox'. Berry maintained that Cantor had virtually proved the existence of the well-ordering of the ordinal numbers by
showing that ordinals of the second class are well-ordered. 3 Berry insist1. Russell 1906e, 645.
2. Blackwell and Grattan-Guinness have carried out independent investigations on Berry's life and
works. Unfortunately, the results are very limited. Berry is also known for his English translations of
Gomperz 1913a and Langlois & Seignobos 1898a.
3. Cantor defmed the second number class as "the totality [Gesamtheit] {a} of all order types a of
well-ordered sets of cardinality No" [Dauben 1979a, 202]. Berry claimed that Cantor had proven that
the second number was well-ordered [sic] but in the first part of the "Beitriige zur Begriindung der
transfiniten Mengenlehre", Cantor simply indicated that "we shall show that the transfinite cardinal
numbers can be arranged according to their magnitude, and, in this order, [they] form, like the finite
numbers, a 'well-ordered aggregate' in an extended sense of the words" [Cantor 1915a, 109]. Russell
did recognize the validity of Berry's proof of the well-ordering of the ordinals a year later [Russell
1906a, 36 second footnote].
135
ed that "it is very easy to prove that [... the] series [of all finite ordinal
numbers] is well-ordered which has none but well-ordered segments". 1
Berry also wrote, "you did not involve yourself in a contradiction which
I expected to find you had.,,2 This implies that by the time he had read
Russell's book (as well as his articles in Peano's Rivista di Matematica),
Berry had already investigated the possible consequences of the well-ordering of the ordinal numbers and he had previously discovered a contradiction. In other words, Berry's discovery seems to have been independent of Russell's writings. When did Berry prepare the argument he sent
Russell? How did he arrive at these results? Was he aware of the 'paradoxes' prior to reading The Principles of Mathematics, implying that he
might have discovered the 'paradoxes' independently of Russell?
It has been impossible to discover the reasons for Berry's interest
in these problems. He did not publish in mathematics. Perhaps, he was
interested in the 'paradoxes' for their own sake. As Russell said in another context paraphrasing Leibniz: Berry seems to be one of those who
"wished to be known only by [his] paradoxes".3 This statement seems to
fit with Berry's personality, whose friendship with Russell focussed on
the 'paradoxes'. Berry is also well known for the way in which he introduced himself to Russell. There is support in Russell's original sources
and in some recent studies that Berry presented himself to Russell by
handing him a note saying: 'The statement on the other side of this paper
is true'. On the other side of the paper was written: 'The statement on
the other side of this paper is [false].'4 Consequently, Berry has been
credited also with the concept of the 'visiting card paradox'.5
Berry's interest in 'paradoxes' relating to set theory is not surprising. Some scholars had become intrigued with this subject some years
before Russell published The Principles of Mathematics. Zermelo, for
1. Letter: Berry to Russell, December 21, 1904. The Bertrand Russell Archives [see: page 167].
2. Ibid.
3. Russell 1899g, 120 & Russell 1900a, 72.
4. Grattan-Guinness 1977a, 50.
5. Church, adding to the already long list of misconceptions and errors, has credited Jourdain with this
paradox [see: Church 1942a, 224].
136
example, had already discovered, independently of Russell, the 'paradox'

of the set of all sets that are not members of themselves.! Hilbert, in a
letter to Frege, acknowledged receiving a copy of his Grundgesetze der
Aritmetik. adding: your 'contradiction'
at the end of the book (p. 253) was known to us here [I believed Dr. Zermelo
discovered it three or four years ago after 1 had communicated my examples to him);
1 found other even more convincing contradictions [ ... ); they led me to the conviction
that traditional logic is inadequate [ ... J.2
As it has been pointed out earlier, Hilbert probably encountered the 'contradictions' from his correspondence with Cantor in 1897. Moreover, I
have already indicated that Russell discovered, at least, two of the most
important 'paradoxes' (Cantor's and his own) because he accepted Cantorian set theory, not because he rejected it. The 'paradoxes' became well
known only after Russell's mathematical and logical formulations. In
fact, some of the theorems of the theory of transfinite cardinal and ordinal numbers seemed to completely contradict common sense that Cantor's
expression 'I see it, but I can't believe it' could had been applied to all of
them.
Now, the original formulation of Berry's contradiction was the
following:
Some ordinals, e.g., w, wl, wWare definable in a finite number of words. Let us
suppose that there is any ordinal which is not so definable. The ordinals less than this
particular one are a well-ordered series. Hence if among them there are any which
are not finitely definable, there is one of these less than all the others. The least
member of the class is then the least ordinal which is not definable in a finite number
of words. But this is absurd, for 1 have just defined it in thirteen words. Going back
a little 1 infer that one not-finitely definable ordinal cannot exceed another such;
therefore there is at most one such; and this again is absurd. 3
As I mentioned before, Russell published a modified version of

Berry's 'paradox' in 1906. Previously, there is no indication that Berry
1. Zermelo 1908a, 191 note 9. [See, also, Thomas & Rang 1981a).
2. Letter: Hilbert to Frege, November 7, 1903; printed in Frege 1980a, S1.
3. Letter: Berry to Russell, December 21, 1904. The Bertrand Russell Archives [see: pages 167-168).
137
had communicated his 'contradiction' to anyone else. In other words, the

inconsistency seems to have been unknown to the mathematical community at large up to 1906, nor did it playa role as a precursor to other similar arguments developed before this date.
In conclusion, Berry's proposition seems to have originated independently of Russell; that is to say, following Frank Ramsey's (19031930) dichotomy, apparently the first example of a non-Logical 'paradox'
was not a direct consequence of the 'contradictions' already discovered
by Russell. Berry's correspondence with Russell shows that, even several
years after he discovered the 'contradictions', he was enchanted with the
topic. On the other hand, as far as we know, Berry's work does not
appear to have influenced the thinking of other researchers.
5.3 The Konig-Zermelo proposition.- By 1883, Cantor had

affirmed but not proven that the principle that every 'well-defined' set
could be reduced to a 'well-ordered' set. He asserted that this principle
was a "fundamental logical law of great consequence, being noteworthy
by its universal validity". 1 By 1895, approximately, however, Cantor
realized he needed to prove this principle. The theorem became well
known in 1900 when Hilbert posed it-along with the Continuum Hypothesis, asserting that 2" = N1-as the first of his famous twenty-three
problems [see: page 34 note 2].
Some attempts to prove the Well-Ordering Theorem were published
but none of them was valid. 2 It was not until August 10, 1904, at the
Third International Congress of Mathematicians at Heidelberg, that Konig
'proved' that the theorem was false. 3 He argued that not all transfinite
sets could be well-ordered. He supposedly demonstrated that the set of all
real numbers could not be well-ordered, and, consequently, the Continuum Hypothesis itself must be false. Zermelo, however, found an error in
Konig's proof the day after its presentation at the congress. Konig had
1. Cantor 1932a, 169.
2. See, among others, Hardy 1904a, Jourdain 1904a and 1904b.

3. Konig 1905a, 144-147.
138
used one of Felix Bernstein's (1878-1956) ideas on the exponentiation of

alephs. The proposition in question is
N: = NO!. 2\ for
every ordinal ct.
Later on, Bernstein claimed that this result was applicable only in the
case of finite ct, but Konig did not specify such a restriction.
On September 24, 1904, Zermelo sent a letter to Hilbert, then
editor of Mathematische Annalen. Part of Zermelo's letter appeared in
this journal two months later. 1 But in the portion of the letter published
by Hilbert, there was no reference to the mistake in Konig's proof discovered by Zermelo a day after Konig's presentation. Those mathematicians who attended the congress but were not close associates of Zermelo
and Hilbert, may have mistakenly assumed there was an inconsistency
associated with the work of Konig and Zermelo. On the one hand, Konig
had supposedly proven that the set of all real numbers was not well-ordered; on the other, Zermelo had just proven that any set could be well-ordered. 2
Soon, the French mathematician Jacques Hadamard published a
brief report of the Heidelberg Congress. 3 Although he disqualified himself from giving a detailed analysis of the papers presented at the conference, he noted one that was of particular interest: Konig's article devoted
to the question of whether the continuum could be arranged as a well-ordered set. Hadamard went on to say that this question, formulated at
the 1900 congress, had apparently been answered in the negative in the
1904 congress. Nevertheless, Hadamard stated that it should not be forgotten that Zermelo had arrived at exactly the opposite conclusion. 4 A
1. See Zermelo 1904&, 139-141
2. Copi has argued [1971a, 10) that both articles were already contradictory at the time of the congress. 'Ibis cannot be, however, because Zermelo's proof of the Well-Ordering Theorem was not
developed until after the congress had finished.
3. [Hadamard) 1904a, 961-962; see, also, Borel 1912a, 157.
4. [Hadamard) 1904a, 961 note 4.
139
few months later, Hadamard discussed the question again,l emphasizing

that the two conclusions were contradictory. He noted that Konig's contribution was still unpublished; consequently, it was impossible to analyze
it, even though mathematicians were already debating Zermelo's Axiom
of Choice. 2
Several interrelated points arise from Hadamard's notes. Firstly, he
did not resolve the apparent 'contradiction' between Konig's and Zermelo's contributions. Instead, he mentioned this was not the first time that
(apparently) contradictory arguments had arisen in set theory. He then
discussed what he considered was an earlier inconsistency. He called
attention to some contradictory findings of Cantor and Burali-Forti. Although Cantor had postulated the existence of a series of transfinite ordinal numbers, Burali-Forti had shown that the existence of certain transfinite numbers, with the properties discovered by Cantor, resulted in
contradiction. Hadamard was not surprised or disappointed by the appearance of these 'contradictions'-not because he was opposed to Cantor's
theory, but, because inconsistencies seemed, historically, a natural occurrence in mathematics. After all, as Hadamard noted, mathematicians had
encountered 'paradoxes' when they first introduced incommensurable,
negative and imaginary numbers. 3 And mathematics had always grown
stronger and richer from these developments.
Ironically, both Zermelo and Konig already knew their results were
not contradictory because Zermelo had found a mistake in Konig's proof,
rendering it invalid. In fact, when Konig's contribution to the congress
1. [Hadamard] 1905a, 241.

2. Hadamard probably had in mind the dispute among some French mathematicians carried out through
personal correspondence. Later on, this polemic was published in the Bulletin de La Societe Mathematique de France [see: Borel 1912a, 150-160; & Moore 1982a, 311-320]. Perhaps, on the other hand,
Hadamard made reference to the polemic generated after Cantor's lecture at the meeting of 1903,
polemic that Schoenflies commented some years later [see: page 37]. The Axiom o/Choice asserts that
"even for an infinite totality of sets there are always mappings that associate with every set one of its
elements, or, expressed formally, that the product of an infinite totality of sets, each containing at least
one element, itseld differs from zero" [Zermelo 1904a, 141]. Cantor previously used a weak version of
this principle in his proof that Ito was the smallest of all transfinite cardinal numbers [see: Cantor
1915a, 104-106].
3. [Hadamard] 1905a, 241.
140
was published in 1905, he had already modified his position, claiming

only that 'if' Bernstein's result were true in general, 'then' the set of all
real numbers could not be well-ordered. Hadamard's recollections concerning Burali-Forti and Cantor left open the possibility that a real 'contradiction' may be discovered within the theory of sets even though none
was present with respect to the work of Konig and Zermelo.
There is, and has been, confusion concerning the status of the
non-logical 'paradoxes', especially the so-called "Konig-Zermelo paradox". Poincare also discussed the Zermelo-Konig antinomy (his term) as
one of the already numerous inconsistencies arising from set theory. 1
Poincare was hunting for 'paradoxes' because he was trying to discredit
both Cantor's theory of sets and Russell's logicism. Unfortunately, Poincare only mentioned the Zermelo-Konig antinomy, but did not expand
upon it or discuss it in detail. It is therefore impossible to know why he
thought there was a 'contradiction', or to what extent he thought of these
as 'contradictions'-or antinomies as he preferred to call them.
There is, yet, a third and different interpretation of the Zermelo-Konig 'paradox'. This account states that
Konig drew the conclusion that the set of real numbers cannot be well-ordered,
contradicting a result obtained by Zermelo shortly before [ ... J. After the discovery of
the Richard paradox, in which the well-ordering problem plays no role, Konig's
conclusion can hardly be accepted. Rather, taking for granted Zermelo's result that
the set of real numbers can be well-ordered, we infer the existence of a first element
in the complement of the set of finitely definable real numbers, and we thus have a
paradox, sometimes called the Zermelo-Konig paradox, bearing on the notion of
definability. 2
5.4 Richard's contradiction. - Richard seems to have developed the inconsistency associated with his name from Hadamard's notes.
Apparently, he overlooked the dispute over Zermelo's proof of the Well-Ordering Theorem, nor was he aware of the concrete formulation of the
1. Poincare 1906a, 333 and 1912a, 530.
2. van Heijenoort 1967a, 47. My emphasis.
141
Burali-Forti 'paradox' which was also described by Bernstein and Schoenfties in the Mathematische Annalen. Their description was based on
Jourdain's comments and a paraphrasing of Russell's version. l
In a letter to Louis Olivier (1854-1910), editor at the time of the
Revue Generale des Sciences Pures et Appliquees, Richard described
another 'contradiction' (his term).2 He commented that he had read Hadamard's note in which certain 'contradictions' found in set theory were
discussed. Richard did not specify the 'contradictions' to which he was
referring, but probably they were the Zermelo-Konig and Cantor-Burali-Forti inconsistencies, the two mentioned by Hadamard. Richard did not
seem to know the detailed formulation of the Burali-Forti 'paradox',
further suggesting that he came upon these arguments through Hadamard's note, in which this 'paradox' was only cited, not discussed. It is
important to keep in mind that at this time (ca., 1904) Zermelo, Konig,
Cantor and Burali-Forti did not claim to have discovered these 'paradoxes' that bear their names.
In his letter, Richard mentioned that it was not necessary to go so
far as the theory of ordinal numbers to find 'contradictions':
I am going to define a certain set of numbers, which I shall call set E,
through the following considerations.
Let us write all permutations of the twenty-six letters of the French alphabet
taken two at a time, putting these permutations in alphabetical order; then, after
them, all permutations taken three at a time, in alphabetical order; then, after them,
all permutations taken four at a time, and so forth. These permutations may contain
the same letter repeated several times; they are permutations with repetitions.
For any integer p, any permutation of the twenty-six letters taken p at a time
will be in the table; and, since everything that can be written with finitely many
words is a permutation of letters, everything that can be written will be in the table
formed as we have just indicated.
The definition of a number being made up of words, and these words of
letters, some of these permutations will be definitions of numbers. Let us cross out
from our permutations all those that are not definitions of numbers.
Let u/ be the first number defmed by a permutation, U 2 the second, u3 the
1. Moore & Garciadiego 1981a, 334. I asserted, in the introduction of this chapter, that an important
premise for this account was the supposition that the metamorphosis of the Burali-Forti 'paradox'
paralleled the origin and development of other paradoxes'.
2. Richard 1905a, 54.
142

third, and so on.
We thus have, written in a definite order, all numbers tluJt are defined by
finitely many words.
Therefore, the numbers that can be defined by finitely many words form a
denumerable infmite set.
Now, here comes the contradiction. We can form a number not belonging to
this set. "Let P be the digit in the nth decimal place of the nth number of the set E;
let us form a number having 0 for its integral part and, its nth decimal place, p + 1
if p is not 8 or 9, and 1 otherwise." This number N does not belong to the set E. If
it were the nth number of the set E, the digit in its nth place would be the same as
the one in the nth decimal place of that number which is not the case.
I denote by G the collection of letters between quotation marles.
The number N is defined by the words of the collection G, that is, by finitely
many words; whence it should belong to the set E. But we have seen that it does not.
Such is the contradiction.'
There are some remarkable points of this argument: first, by arguing that it is not necessary to go as far as the theory of ordinal numbers,
Richard is implicitly affirming that his contradiction may be found in the
mere roots of the concept of set; second, Richard explicitly argued that
he had discovered a contradiction, not a paradox or an antinomy. He
went even farther and suggested a solution. The collection G had meaning only if the set E was defined in totality; this could not be done except
with infinitely many words. Poincare agreed with Richard's solution and
called him 'sagacious'.2 Perhaps, Poincare thought that Richard was proposing that mathematical 'contradictions' were implicit in the concept of
any infinite set. Therefore, the concept of infinite set had no value in
mathematics. It should not be forgotten that Poincare was strongly opposed to: both Cantor's introduction of the actual infinite in mathematics;
and, Russell's logicism. 3 Perhaps, Poincare thought that Richard would
be equally critical of Cantor and Russell. Later on, Richard explicitly
stated that the actual infinite was the source of his 'contradiction', perhaps in response to certain criticisms, especially those of Peano, who
argued that "Richard[,s example did] not belong to mathematics, but to
linguistics" .4
1. Richard 1905a, 541.

2. Poincare 1906b, 305.
3. See: Poincare 1905a, 1906a, 1906b and 1906c. See, also, Gray 1991a.
4. Peano 1973a, 218.
143
5.5 Polemics within the London Mathematical Society

over the Well-Ordering Theorem.- Zermelo's letter to Hilbert of
1904 provoked an immediate and strong reaction from the mathematical

community. This criticism which consisted of several publications opposed to certain features of Zermelo' s proof. Perhaps the most interesting
aspect of these criticisms lies in the fact that different aspects of the proof
came under attack from different authors.
In addition to controversies in the German and French mathematical communities, there was also a dispute among British mathematicians.
Ernst William Hobson (1856-1933) initiated an exchange of ideas in the
Proceedings of the London Mathematical Society when he published an
article critical of G. H. Hardy (1877-1948) and Jourdain. Hobson criticized their attempts to prove that every transfinite cardinal number was
necessarily an 'aleph'. As far as can be determined, Hobson was the first
mathematician to suggest that it was necessary to carry out "a further
scrutiny of the foundations of [... set] theory", 1 because of the inconsistencies involved in conceiving the sets of all ordinal and cardinal numbers.
Hobson argued that Jourdain's explanations of the 'paradoxes' were
not satisfactory, because they were based on an unclear distinction between consistent and inconsistent sets, a distinction that Jourdain claimed
to have introduced independently of Cantor. Hobson insisted that Cantor
and Jourdain failed to provide a criterion to decide whether a set was
consistent or inconsistent. 2 Hobson proposed a new definition of the concept 'set' in terms of objects satisfying a prescribed 'norm'. This meant
that a set was defined by conditions sufficient to determine whether an
object belonged to an aggregate or not. Unfortunately, this kind of definition cannot assure that a transfinite set could be ordered. For this to happen it would be necessary to include a principle implicitly or explicitly in
1. Hobson 1905a, 171.
2. See: Jourdain 1904b, 67 first footnote. Cantor's similar distinction between consistent and inconsistent multiplicities, contained in his personal correspondence with Dedekind, had already been rejected
by Hilbert because of the lack of "a precise criterion for this distinction" [van Heijenoort 1967b, 131].
144
the norm (Hobson's term) defining the aggregate by which the aggregate
could be ordered.
There were replies to Hobson's article from Hardy [1906a], Dixon
[1906a], Russell [1906a], among others. Hardy announced that Russell
was preparing a lengthy reply on general aspects of Hobson's criticisms,
including Hobson's concern that no non-Denumerable aggregate could be
well-ordered. Hardy, therefore, restricted his comments to criticisms
Hobson had made of his 1904a article, even though Hardy also reviewed
the Multiplicative Axiom and Zermelo's Axiom of Choice.! Hardy accepted, provisionally, the Multiplicative Axiom, but made no conclusion about
Zermelo's Well-Ordering Theorem.
Russell's reply to Hobson was read before the London Mathematical Society on December 14, 1905. Russell insisted that Hobson had
confused difficulties relating to inconsistent aggregates with other problems concerning Zermelo's axiom. Russell identified his own concept of
'propositional function' with Hobson's concept of the 'norm' of a set,
and discussed the 'contradictions' described by Hobson. Most importantly, he showed that these 'contradictions' were all particular examples of
a more general 'contradiction'. Russell also claimed to have shown that
the 'contradictions' were not based primarily on arithmetic, but on logic.
Therefore, in order to resolve them, the underlying logical assumptions
would have to be modified. In addition, Russell proposed three different
logical theories to solve the 'contradictions', all of which had their roots
in The Principles of Mathematics.
The 'Zig-Zag Theory' would eliminate propositional functions that
were too complex, but would permit others, such as the one asserting
that the complement of a class is a class. It would allow a largest cardinal
number but not a largest ordinal number. The axioms Russell introduced,
1. The Multiplicative Axiom states that: "given any class of mutually exclusive classes, of which none
is null, there is at least one class which has exactly one term in common with each of the given
classes" [Russell 1919a, 122J. This Multiplicative Axiom, as Russell called it, is needed to prove that
the infinite product of either finite or transfinite numbers is zero when one of the factors is zero. The
Multiplicative Axiom corresponds to a weak form of the Axiom of Choice.
145
however, were quite complicated and "lacked intrinsic plausibility". 1

Russell's second proposal limited the size of classes. A similar solution,
as has already been mentioned, was proposed by both Jourdain and Cantor. Within this theory, classes like 'the class of all entities' would be
prohibited, as well as classes such as 'the class of all ordinal numbers'.
However, it was necessary to include all of Cantor's ordinals. Thus,
there remained a difficulty: how to state a limitation of class size precisely? To cope with this problem, Russell proposed a third solution, 'the
no-class theory'. In order to avoid the 'contradictions' entirely, this theory banished classes and relations altogether. In this theory a propositional
function would also be eliminated (by substitutions) simply as an abbreviation of an argument for one or several of its values. Russell anticipated
a number of possible objections to the theory: that classes seem too obvious to be denied; that one would have to refute a great part of Cantor's
theory of transfinite numbers; that working with the theory was very
complicated. The no-class theory was clearly the most radical of the
three. Although Russell at first doubted that it provided the final solution
to the 'contradictions', two months later he added a note to the article
saying that he now had no doubts that the no-class theory afforded the
complete solution. 2 Nevertheless, Russell published yet another 'final'
solution to the contradictions in 1908. This later solution was based on
the theory of types and proved to be definitive.
Russell devoted the rest of his article to Zermelo' s Axiom of
Choice and his own Multiplicative Axiom. Russell exemplified the use of
Zermelo's axiom by his now famous example of the man who owns No
pairs of boots. 3 He also correctly stated that his own axiom could be
2. Russell 1906a, 53, note added February 5, 1906.

3. The example runs in the following way: there is a millionaire who owns Mo pairs of boots and Mo
pairs of socks. The question is: How many boots and socks had he? "One would naturally suppose that
he had twice as many boots and twice as many socks as he had pairs of each, and that therefore he had
Mo of each, since that number is not increased by doubling" [Russell 1919a, 126]. It would be easy to
imagine how many boots he had in total because Mo + Mo = Mo' This is easy to prove for the case of
boots because one can distinguish between a right and left boot of each pair. Therefore, one would
have no trouble selecting a boot from each pair. Unfortunately, manufacturers of socks do not make
any difference between right and left socks. Thus there is no criterion for picking one sock from each
pair [see: Letter Russell to Hardy, June 22, 1905 (quoted in page 191); Russell 1906a, 47-48 &
146
deduced from Zermelo's, but Russell not only doubted the validity of the
general case-Zermelo's Axiom of Choice-but also of his own.
5.6 Konig's and Dixon's arguments. - In the meantime,
Konig kept attempting to disprove Cantor's Continuum Hypothesis. In

order to do so, Konig was trying to find a proof that the continuum could
not be well-ordered. His second attempt, following the Heidelberg 1904
Congress, was made in a lecture originally delivered before the Hungarian Academy of Sciences on June 30, 1905. 1 This occurred at almost the
same time as Richard's letter appeared in the Revue Generale des Sciences Pures et AppUquees. Konig's aim, however, was quite different from
Richard's. Richard very clearly presented a mathematical 'contradiction'.
He admitted that he was motivated to write his note because of the publication of the other 'contradictions'. Konig, on the other hand, was not
concerned with 'paradoxes'. In the first place he argued that the indiscriminate use of the word 'set' resulted in the apparent 'paradoxes'. In
fact, he proposed solutions to the 'paradoxes' in two different places in
the article without going into a detailed analysis, suggesting that he did
not consider them to be of major importance. In his first solution, he
commented that the second number class should not be considered to be
a completed set, defined as a totality of well distinguished elements that
are entirely distinct conceptually. This solution leaves out the Burali-Forti
'paradox', because w, representing the totality of all finite ordinal types
as a completed set, was effectively excluded. His second solution, the
more general one, involved a distinction between the concepts of 'set'
(completed sets) and 'classes' (sets in the 'act of becoming').
The main goal, it should be stressed, of Konig's article was not to
solve the inconsistencies, but to show that the continuum could not be
well-ordered. Konig's proof was formulated in terms of 'finitely defined'
elements of the continuum and the complements of sets. First of all he
gave a definition of the 'finitely defined' element:
Russell 1919a, 125-126].

1. See: van Heijenoort 1967b, 145-149.
147
[ ... ] an element of the continuum will be said to be 'finitely defined' if, by means of
a language capable of giving a definite form to our scientific thinking, we can in a
finite span of time specify a procedure (law) that conceptually distinguishes that
element of the continuum from any other one. I
Konig's initial hypothesis in his proof was to suppose that the continuum was a well-ordered set, which he took to be a false assumption.
Considering the continuum to be a well-ordered set, the elements that
could not be finitely defined, formed a subset M of the supposedly well-ordered set. The cardinality of the set of all the finitely defined elements
of the continuum is No and because the continuum is not denumerable
(that is to say, it is impossible to find a one-to-one correspondence between this set and the set of all natural numbers and, therefore its cardinality must be greater than No), there exist elements of the continuum that
cannot be finitely defined. The subset M has to be well-ordered (because
the subsets of a well-ordered set are well-ordered sets) and, therefore, it
contains a first element. But if the continuum is a well-ordered set, then
there is a one-to-one correspondence between the continuum and a set of
specific ordinals. As a result of the one-to-one correspondence, there
would have to be a first ordinal that was not finitely defined. But this is
impossible, because "there exists a definite (well-ordered) set of finitely
defined ordinals that follow each other in an unbroken sequence beginning with the first".2 Konig then obtained a 'contradiction' to the assumption that the ordinal in question could not be finitely defined by
observing that this ordinal was defined by the phrase "the first ordinal
that exceeds all of these in magnitude".3 The 'contradiction' resulted
from the supposition that the continuum was well-orderable. Once again,
it is important to keep in mind that Konig, as well as Burali-Forti, Cantor
and Dixon (who is discussed below), did not claim to have discovered an
inconsistency. He was only trying, unsucessfully, to prove that the continuum could not be well-ordered. Konig's subsequent interests evolved
in a different direction, but that is part of another story. 4
1. Konig 1905c, 146.
2. Ibid., 147-148.
3. Ibid., 148.
4. In later papers Konig argued that he had in fact discovered a 'paradox', reflecting increasingly the
influence of Poincare. Finally, being one of the few who modified his original ideas, he came to accept
Zermelo's proof of the Well-Ordering Theorem [see: Konig 1907a, 1914a, and Moore 1982a, 119--
148
Some historians believe that both Richard and Konig discovered a

'paradox'. Remarkably, both "were written during the same weeks, perhaps even the same days, and independently of each other."1 In addition,
Copi claims that Dixon published another 'paradox' in an article read
before the London Mathematical Society on December 14, 1905. However, according to Dixon himself, "[ ... ] the idea on which this article [1906a] is founded was mooted by me in a letter to Dr. Hobson in June of
this year [1905] ...... 2 If Dixon had discovered another inconsistency by
June 1905, then these coincidental discoveries were even more remarkable, as they involved not only two independent cases, but three. 3 It is
true that Richard, Konig, and Dixon each devised their arguments independently of the others in June 1905. But Konig, as explained earlier,
considered his argument as part of his rebuttal of the well-ordering of the
continuum. Similarly, Dixon did not consider his argument to be a 'paradox', but rather an attempt to justify the proposition that no transfinite
aggregate could be well-ordered by a finite set of rules.
As already mentioned, Dixon's ideas were originally stimulated by
the same article of Hobson's that provoked the debate among British
mathematicians. Dixon thought that it was possible to find the cardinal
number of the set of all objects that could be defined in finite terms.
What did he mean by 'define'? An object was defined, he held, when the
properties necessary to distinguish it from all other objects of mental
activity were stated. Since the number of elements available for use in
any definition is finite, say p (including the symbols of all known alphabets, signs of punctuation, etc), the cardinal number (or power) of the set
of objects described by such finite specifications could be neither greater
nor less than No. It was not greater because a non-Denumerable set could
121].
1. van Heijenoort 1967b, 142

2. Dixon 19068, 20 footnote.
3. In fact, it could be argued that these were 8 total of four discoverers. One should not forget that
Berry's argument, which he sent Russell in December 1904, is very similar and also involves the least
indefinable number defined in 8 finite number of words. These 'paradoxes' (Berry's and Richard's)
were so similar, that some mathematicians believed that Berry and Russell had merely simplified
Richard's argument, to create yet another paradox [Bourbaki 1960a, 47]. This is incorrect. One of
many misconceptions appearing during the century concerning the set-theoretic 'paradoxes'.
149
not be formed with only a denumerable number of elements; it could not

be less because the sequence of cardinal numbers alone already provided
a set of objects finitely defined with cardinality Na. Because the set of all
real numbers had a cardinality greater than No, there had to be elements
of the real number system which could not be given a finite definition of
specification. However, because of Zermelo's theorem that any transfinite
set could be well-ordered, each number could be finitely defined. If this
were not the case, then
there must be a first [element] which cannot [be finitely defined], and it may be
described in finite terms as 'the first in the series determined by the rules R which
cannot be described in finite terms'. This is absurd. I
This argument was immediately applied to Cantor's theory of

ordinal numbers and the second number class. Now the absurdity laid in
the phrase "the least number of the second class that cannot be specified
in finite terms.,,2 Instead of concluding that he had found a contradiction
or an inconsistency, Dixon inferred that it was impossible to say that a
transfinite aggregate could be well-ordered by a 'finite set of rules'. This
did not contradict Zermelo's theorem, because Zermelo had simply
shown that it was possible to provide a well-ordering of a transfinite set,
but without showing how to produce a specific arrangement of the set, as
Hilbert had originally hoped. 3
Russell, in his now classic formulation of 'the theory of types'
[l908c], discussed Dixon's and Konig's arguments as if they involved
'contradictions'. To my knowledge, this is the first time that these arguments were treated as 'contradictions', implying that Russell not only
created some of the most famous logical 'paradoxes', but also some of
the semantic ones as well. Russell transformed Konig's and Dixon's
argument into a simple and direct contradiction:
(5) Among transfinite ordinals some can be defmed, while others cannot; for the total
number of possible definitions is Ito, while the number of transfinite ordinals exceeds
Ito. Hence there must be indefinable ordinals, and among these there must be a least.
1. Dixon 1906a, 19.
2. Ibid.
3. Hilbert 1900b, 447.
150

But this is defined as "the least indefinable ordinal', which is a contradiction!
In a footnote Russell listed Konig [1905c] and Dixon [1906a] as sources,

but he did not mention whether he thought that either of them was thinking in terms of 'contradictions' or not. Russell's omission represents one
additional contribution to the considerable confusion and misunderstanding associated with the development of the set-theoretic 'paradoxes'!
1.
Russell 1908c, 223.
Conclusions
In May 1901, for the first time, Russell presented an argument

related to his 'contradiction' of the class of all classes which are not
members of themselves. Later on, he began to regard some mathematical
inconsistencies in Cantor's work as contradictions (Russell's term) rather
than a fallacies or mistakes. Also at this time, he became aware of the
possible existence of areas of conflict between certain aspects of the work
of Burali-Forti and Cantor.
Insight into the reasons why Russell discovered these 'paradoxes'
can be found through the study of the various drafts of The Principles of
Mathematics. These manuscripts, written over six years (approximately),
reveal a gradual transformation in Russell's thinking about the 'paradoxes'. The key factor in explaining this evolution in thinking was Russell's slow adoption of Cantor's theory of transfinite numbers, which he
had first strongly rejected. I proposed, in order to account for this change
in Russell's thinking, that Russell's adoption of neo-Hegelian principles
between approximately 1893 to 1898 allowed him to accept the view that
mathematics was generally an inconsistent body of knowledge. This
viewpoint was expressed in the first manuscript of The Principles of
Mathematics, entitled "An Analysis of Mathematical Reasoning", in
which the foundations of the concept of number was discussed. In other
words, I claimed that when Russell wrote the first version of his 'big'
book, he was convinced that mathematics contained quite properly inextricable 'contradictions'. Later on, I also argued that, by the time he went
to the First International Congress of Philosophy held in Paris in 1900,
he had already modified his first conceptualization, as a result of his slow
152
Conclusions
adoption of Cantor's ideas. No longer accepted 'contradictions' as inevitable; nevertheless, Russell came to view the existence of 'philosophical
antinomies', that he did not expect to solve at this point, associated with
the concept of infinite number.
Based on the analysis of Russell's various attempts to write a book
on the foundations of mathematics, I concluded that Russell wrote only
one final manuscript of The Principles of Mathematics between November 1900, after he met Peano, and January, 1903. Contrary to Russell's
own later recollections, I discussed how his manuscript was completed in
three different stages: the first from November 1900 to December 1900
when Russell wrote parts III, IV, V and VI. The second step was taken
during May and June 1901 when Russell originally wrote parts I and II.
Finally, Russell made a third and final effort to finish the book between
May 1902 and January 1903. At this time, Russell rewrote Part I, incorporated Part VII from a previous manuscript, and added two new appendices. This last effort to write the book was necessitated by Russell's
inability to resolve, at that time, the difficulties caused by the discovery
of the 'paradoxes'.
The study of the philosophical and mathematical background of
Russell's The Principles of Mathematics allowed me to assert that he was
interested in the study of the philosophy of mathematics since an early
age, and that his concern was subsequently reinforced at Cambridge
University. I concluded that the 'paradoxes', or contradictions in Russell's terms, resulted from his eagerness to scrutinize the primitive notions of mathematics, in particular, those related to the principles of the
concept of number. His interest did not originate from the discovery of
the nascent 'paradoxes' of Burali-Forti and Cantor, which, in fact, were
not recognized as inconsistencies at the time.
The inaccuracy of the standard interpretation was closely examined, especially with respect to the role played by Burali-Forti and Cantor. As a consequence, it became necessary to question the legitimacy of
alleged causes and consequences role stated by the standard interpretation
in the development of a crisis in the foundations of mathematics at the
beginning of the twentieth century. It is understood, implicitly and even
153
explicitly, that one of the elements of this crisis was the discovery of
even more inconsistencies; today called semantic (or linguistic or epistemological) 'paradoxes'. But I was able to show, in general, that these
non-Mathematical inconsistencies did not originate directly from the
'logical' ones. The study of primary sources has clarified the role played
by some mathematicians associated with these semantic 'paradoxes'.
Some of these arguments were initially formulated as 'positive' results of
the theory of transfinite numbers and sets-although, they were actually
proving assertions that contradicted Cantor's own original expectations.
In short, the general conclusions drawn from my analysis of the way in
which the semantic 'paradoxes' were discovered, once again, contradicts
the standard interpretation.
This standard interpretation has been used to discourage mathematicians from becoming concerned with the analysis and study of the fundamental notions and concepts of mathematics. But there remain several
unanswered questions concerning the origin, development and acceptance
of these inconsistencies. Two examples, among others, are given here:
what was the role of these 'paradoxes' in the origin of three modern
philosophical schools of thought in mathematics Logicism, Formalism and
Intuitionism? When, how and why Cantor formally discovered the 'contradictions '?
Finally, paraphrasing Russell's last sentence of his The Principles
of Mathematics: I earnestly commend to all students of history and philosophy of mathematics the study of the historiographical, historical and
philosophical consequences of this new historical interpretation of Russell's role in the discovery of the set-theoretic 'paradoxes'.
Undoubtedly the an of letter-writing has been

lost ill the twentieth celltury through the development
of the telephone and, to a lesser extent, the greater
ease of travel. In my youth it was imperative to
master the an of letter-writing if one was to make
one's way ill the world. Leiters had a formality which
today must appear ponderous and even absurd.
Bertrand Russell.
Appendix. Correspondence
6.1 Alys Russell.- Alys Pearsall Smith and Bertrand Russell
were married on December 13, 1894. Shorty before Russell finished The
Principles of Mathematics, he suddenly realized that he no longer loved
his wife. These letters were written during a period of temporary separation. Mrs. Russell was undergoing a rest cure, while Russell was attempting to finish his book.
LETTER 1: From Bertrand Russell to Alys Russell
Friday's Hill.
Haslemere.
May 9, 1902
My Dearest,
Many thanks for thy two letters. I was afraid thee was having a
real relapse, so I was not surprised. I am very much interested that thy
rheumatism is nervous: there is a sort of satisfaction about a physical
symptom, as it seems more definite & possible to combat. I am sure thy
change of plans is wise, & I could tell thee how glad I am thee is giving
up the Summer School, which would have been sure to upset thee again.
I wonder if B'ham even for a day will be allowed in the end. It is great
comfort that thee has grown so reasonable, & amiable to Dr's orders. I
wonder if there really is no one thee would like to go to Switzerland
with. Why has thee abandoned Jane & Helen V.? I feel as if there must
be some one: would a person like Beatrice Creighton suit thee better?
Correspondence
156
Mrs C. comes today: I shall see her at tea, but I dine with the
Davies's: one meal enough for me! I got no news at Dunrozel yesterday
of any sort. I did one of the hardest days of work that I have ever done
in my life: seven hours real work, & two correcting proofs from Peano.
I had the hardest Chapter of my book to write, I was anxious to finish it
within the day, while my mind was full of the subject. I succeeded, tho'
it was 30 pages: but once or twice I found myself forgetting everything
in heaven & earth, as I did during my Tripos. I have been so long without real work, that I have come back to it with a kind of fever: everything else seems unreal & shadowy to me just now, & I work as if I
were possessed. If only I can keep it up, I shall soon get a great deal
done.
Keep up thy spirit, Dearest. I shall allow myself to dream of meeting again on the 29th-the other did not seem real & I kept it out of my
mind.
Thine ever devotedly,
Bertie.

Friday's Hill.
Haslemere.
May 16, 1902
My Dearest,
I am sorry thee is having another attack of depression. Thy letter
of Wednesday is very nice in being so affectionate, but it is depressing in
other ways. I have given up thinking of a fixed date when the separation
will be over & I am not surprised at the Dr's postponing Switzerland.
My plans are not very settled. I mean to go back to Cambridge on Tues-
157
day & stay a week or a fortnight. I work hard there now, & never sit out
chatting in the garden. But I find myself feeling rather lonely when the
Webbs are my only society, & I prefer to be where there are plenty of
people to see.- I expect to have my book quickly finished in another two
months, if only I can keep fit & go on working hard. It will not give me
any feeling of elation, merely a kind of tired relief as at the end of a very
long dusty railway journey. The book will be full of imperfections, &
will raise innumerable questions that I don't know how to answer. There
is a great deal of good thinking in it, but the final product is not a work
of art, as I had hoped it would be. I shall send it to the Press at once, as
the load will not be off my mind until I cannot make any further corrections.- I had two splendid nights at Bournemouth, but last night again
was not so good: between 5 & 6 hours. However I am really rested now,
& feel quite equal to doing my work. The worst time was just after the
separation began, when I had no longer a motive for keeping up.- I
prefer to know, if thee doesn't mind, how thee is from day to day.
I believe now in what Dr. Boyle says about the fat and the thin:
certainly I have a quite extraordinary power of recuperation as soon as I
can get sleep. Yet it was hard to guess that my nerves were stronger than
thine.
I have no definite plans: I mean to go backwards & forwards between here & Cambridge, working in both places; until this separation is
over. - When is thee going to get that 100? If thee will send me one
large cheque when thee does, I will pay the bills. I have just managed to
get on with cash, as I spend none at Grantchester.
Thine ever devotedly.
Bertie.
Correspondence
158
LEITER 3: From Bertrand Russell to Alys Russell

The Mill House,
Grantchester,
Cambridge. 1
Telegrams
Trumptington
May 22, 1902.
My dearest,
No letter from thee, owing no doubt to the move into the country.
I am forging ahead with my book, & shall very soon reach the point
where I shall have only more or less mechanical work to do-if only I
can keep on a little longer. But I do get tired over work. I have learnt to
pull myself together & make efforts which formerly I should have
thought impossible; but I suppose there are limits to the process. I should
like to put as a dedication to my book: "To Molocl this Altar is dedicated
by a Sacred Victim". But I fear the mathematical public might be puzzled.
Yesterday Evelyn was out to lunch and tea. I had lunch with the
children & tea with Alfred: except for a short ride, I worked all day.
Nothing of interest occurred, I have had no thoughts except concerning
matter & motion, which I am now working at.
Take care of thyself & be happy.
Thine devotedly,
Bertie.
1. At this time, Russell was living alternatively at the Whitehead's and Friday's Hill.
159
Telegrams
Trumpington
My dearest,
The Mill House,

Grantchester,
Cambridge.
May 24, 1902.
Thee will be surprised & amused, after all my talk of 2 months, to

learn that I finished my book yesterday. I found that a pile of old Ms,
which I had expected to have to re-write, required only a few additions
& corrections, so I arrived at a sudden termination. I have never known
or ever imagined such a relief as I have been feeling. The weather suddenly turned warm, & I sat by myself in the garden. For the first time
for ever so long, I saw that spring was beautiful: I have had no eyes for
anything of late. Wedd came out, & I enjoyed a long talk with him. He
asked after thee, as he always does, with great solicitude.- Although I
contemplated two months, I feel now as if another day would have ended
me. It is extraordinary how one collapses when one has been going on
nerves: there are few feelings more complete or more delicious. Last
night I slept well, which I had not done for some time. And today, with
thy cheery letter, I feel happier than any time since April 14.- I am glad
thee likes the country, & thee too must be feeling the comfort of having
summer at last.- I have not had time to read Lucy's letter to Evelyn: she
is very full of social engagements, & I have been full of work, so that I
have scarcely seen her of late. I wrote yesterday to R. T. Wright, my
aunt and Lucy. Also at last I began to read Mrs. Carlyle's letters. It was
very kind of thee to send such charming picture p.c's to the children.
I hope my book may be out in October with luck.
Thine most-affly.
Bertrand Russell.
160
Correspondence
6.2 George E. Moore.- Moore was born in London and educated at Cambridge (late 1890s). There he met Russell as a student and,
although Moore was younger, he made a deep impression on Russell. In
fact, Russell always deemed Moore as the one who helped him to overcome the early neo-Hegelian influence.
LETTER 1: From Bertrand Russell to G. E. Moore
Friday's Hill,
Haslemere.
16th Aug. 1900.
Dear Moore,
It is a long time since I have had any proofs from you: if, as I
suspect, you have left Cambridge, shall I send you the Gerhardt?
We got back from abroad last night, after a most successful time.
The Congress was admirable, & there was much first-rate discussion of
mathematical philosophy. I am persuaded that Peano & his school are the
best people of the present time in that line.
Have you ever considered the meaning of any? I find it to be the
fundamental problem for math[ematical] Ph[ilosophy]. E. g. "Any number is less by one than another number". Here any number cannot be a
new concept, distinct from the particular number, for only these fulfil the
above proposition. But can any number be an infinite disjunction? And if
so, what is the ground of the proposition? The problem is the general one
as to what is meant by any member of a defined class. I have tried many
theories without success.
I found that Whitehead has a great reputation: all the foreigners
who knew Mathematics have read & admired his book, & were delighted
to meet him.
161
Yours ever,
Bertrand Russell.
P.S. We should be delighted to have a visit from you any time in September.
LETTER 2: From Bertrand Russell to G. E. Moore
Friday's Hill,
Haslemere.
21th Aug, 1900.
Dear Moore.- I am very sorry we shall not see you in September, but I
shall be up early in Oct. & shall hope to see you then. Very many thanks
for the proofs: I am most grateful to you for the trouble you have taken.
You see by the number of mistakes you found that the trouble was not
wasted.- I suppose your translation of G. II A b 1 (E G b) is correct, but
it expresses a sentiment very contrary to L[eibniz]'s usual advocacy of
natural theology.- The question of any is just the same in the case of
men, only there the conjunction is not infinite. The difficulty is, I think,
very real, & I will send you a paper on it when you come home from
abroad.- I am learning Peano's system, who is splendid-the best thing
that has been done for a very long time.
Yours gratefully & frat.
Bertrand Russell.
162
Correspondence
6.3 David Hilbert.- Hilbert was, perhaps, the most influential

mathematician of the late nineteenth and early twenty centuries. His
famous list of twenty three problems, delivered at the Second International Congress of Mathematicians (1900), has guided original research in
many ways. His Formalist program has also had a tremendous influence
in the mathematical realm; perhaps, in mathematical education, stronger
than any other area in mathematics. As far as we know, there only remain two postcards addressed from Hilbert to Russell. They were originally written in German, but we present our English translation.
LETTER 1: From David Hilbert to Bertrand Russell
Highly honoured Colleague,
I have allowed myself to communicate to you that we have already
been engaged for a long time in the mathematical society in the study of
your theory of knowledge, and that just before the outbreak of the war
we had intended to invite you to G5ttingen-through the Wohlfskehl
foundation, so that you could give us personnally a cycle of lessons on
your solution to the problem of the paradoxes. I hope that the execution
of this plan will not be cancelled by the war, but that it will only be
postponed.
Very respectfully,
z.z.d. 12 April 1916.
Lupang-Paradiso, Hotel Europe.
Hilbert.
[Bestial Hun!]. 1
1. Russell added this comment, many years later.
163
LETTER 2: From David Hilbert to Bertrand Russell

Merlingen (Thuner See)
Saturday, May 24, 1919.
Highly honoured Professor,
Thank you very much for the card that I received 3 years ago via
Hecke-Baser. I trust that you will not regret having then agreed to deliver
lessons on logic and the theory of knowledge in G5ttingen. The situation
in Germany is still very unpleasant, of course. Since then, we have studied your works assiduosly in our mathematical-philosophical circles in
G5ttingen, and expect very much of your personal presence in G5ttingen.
In the hope for better times and for the reconstitution of the international society of scholars, best greetings
respectfully,
Hilbert.
Sender: Prof Hilbert.
The answer will reach me through the address of Dr. Richard Baer in
Zurich, Berost 54.
164
Correspondence
6.4 Cesare Burali-Forti.- Burali-Forti, best known as the

discoverer of the 'paradox' of the greatest ordinal number in 1897,
worked on mathematical-logic under the direction of Peano. Russell
probably met Burali-Forti personally at the First International Congress
of Philosophy held in Paris in July 1900. Since Russell credited Burali-Forti with the discovery of the 'paradox' mentioned above, he likely
regarded Burali-Forti as a very clever and brilliant practitioner of mathematical logic and set theory. As far as we know, the following letter is
the only one remaining from a possible exchange of correspondence. The
letter, addressed from Burali-Forti to Russell, was originally written in
French.
LETTER 1: From Cesare Burali-Forti to Bertrand Russell
Corino. March 22, 1912.
Dear Sir and Colleague.
I feel honoured for your petition about a communication I should
send to the philosophical section of the Congress of Cambridge.
Some time ago I abandoned philosophy studies to devote myself
completely to the modern part of vectorial calculus which in a short
period of time has been extended a great deal more than one could expect.
The previous studies of logic have been very useful to me in order
to establish the vectorial notations. Some news about logic procedure
which has served as a foundation to the modern notations is found in the
notes of the first volume of the new publication Analyse vectorielle generale, which I hope will be printed before the Congress (in collaboration
with Marcolongo).
Do you think that a communication in relation to the formal logical
laws of a general notation system could be of interest to the Congress
assistants board of the philosophical section?
165
If you agree, I can promise you a copy of it, within a short period
of time I would be able to tell you the exact title. It will not have the
significance you would desire and that I would like it to have, but it will
have the merit of demonstrating the simplicity and precision of the notations of your great Hamilton, whom I am admire greatly.
I must admit that it is almost certain that I would not be able to

attend the Congress due to the fact that during August I will still be
lecturing in the Military Academy because of the course acceleration.
I send you all my thanks and my sincere greetings together with
the re-affirmation of my most high esteem.
C. Burali-Forti.
166
Correspondence
6.5 G. G. Berry. - In an holograph note, kept at The Bertrand
Russell Archives, Russell described Berry in the following way,
G. G. Berry, the writer of the following letter, was a man of very considerable
ability in mathematical logic. He was employed in a rather humble capacity in The
Bodleian, his subject being one which The University of Oxford ignored. The first
time he came to see me at Bagley Wood he was bearing, as if it were a visiting
card, a piece of paper on which I perceived the words: "The statement on the other
side of this paper is false". I turned it over & found the words: "The statement on the
other side of this paper is false"[sic). We then proceeded to polite conversation.
LETTER 1: G. G. Berry to Bertrand Russell

12 Hermes Road,
Oxford
Dec 21, 1904
Dear Sir,
I was talking yesterday to my friend & former schoolfellow, Dr.
McTaggart, and he said he thought you would not mind if I ventured to
write to you on certain topics connected with Mengenlehre in which I am
greatly interested but which no one here in Oxford seems to care much
about.
There is, in the first place, a theorem about cardinal numbers
which may be new to you and Mr. Whitehead, and which you might
think of sufficient interest to include, duly clothed in the Peano garb, in
the second volume of your Principles. Shortly stated it is as follows:
If any theorem asserts that from a particular set of linear relations
among a system of variable cardinals another such relation may be inferred, it is sufficient to prove that theorem on the supposition that none
of the variables exceeds <xo ' provided that no variable has, in the supposed set of relations, coefficients which sum up to more than <XO.
167
From this then follows, as immediate consequences, Bernstein's

Equivalence theorem, Zermelos's theorems on additions, and Bernstein's
theorems on the division of cardinals by finite numbers, and on inequalities. The only new theorem I can get out of it is that if there are transfinite cardinals for which 2x > x, these are transfinite cardinals which
may be called prime, because they are not divisible by any finite number.
I should be glad to send you a fuller statement, & the proof of this
theorem, if you desire it.
Secondly, there is the question of the maximum ordinal. It was in
hope of illumination on this point that I originally turned to your book,
but I was disappointed. Your suggestions, both there, and in R d M was
that there is no proof that ordinals are well-ordered. But Cantor proved
it, virtually, in 1895, I think, when he proved that ordinals of the second
class are well-ordered. The limitation is quite otiose in his proof. Apart
from that it is very easy to prove that every series is well-ordered which
has none but well-ordered segments. For a not-well-ordered series must
contain a regression; and this has segments, which likewise are regressions, and in which lie each of them in some segment of the original
series. To my mind the difficulty of the maximum ordinal is of equal
dignity & importance with that of the maximum cardinal; and whatever
sets my mind at ease about the one will, if I am not mistaken, set it at
ease about the other.
I noticed, by the way, that you did not, in your article in R d M,
adopt Cantor's method of proving that ordinals of the second class are
well-ordered. Thus you did not involve yourself in a contradiction which
I expected to find you had. But do you not, on the other hand, assume
the existence of WI without proving it? Suppose I choose to assert that
every well-ordered series is either finite or denumerable, what is there, in
your article, to prove that I am wrong? But if that assertion is true, the
class WI is the null-class.
Here I should like to set down a proof, which I am well aware is
unsound, of this same assertion. Some ordinals, e.g., w, W, wW, are
definable in a finite number of words. Let us suppose that there is any
168
Correspondence
ordinal which is not so definable. The ordinals less than this particular
one are a well-ordered series. Hence if among them there are any which
are not finitely definable, there is one of these less than all the others.
This least member of the class is then the least ordinal which is not definable in a finite number of words. But this is absurd, for I have just
defined it in thirteen words. Going back a little I infer that one not fmitely-defmable-ordinal cannot exceed another such; therefore there is at
most one such; and this again is absurd. Hence every ordinal is fmitely
definable.
Suppose, then, that any well-ordered series has more than a o members. It has also more than a o and these have among them more than a o
types of arrangement or ordinals. Some of these, therefore, are incapable
of finite definition, for a definition is a finite collection of words, such
collections are a denumerable aggregate, and there are not enough to go
round. It follows, from what we have already proven, that no well-ordered series can have more than a o members.
This little piece of sophisty, with others which it has suggested to
me and may to you, seems to have some bearing on the difficulties of the
maximum ordinal, the maximum cardinal, and the Contradiction. But I
have written too much already, and will now only say that if you have
travelled any nearer a solution of these puzzles since writing your Principies, I should be most grateful for any hint you could send me.
I remain
yours faithfully
G. G. Berry.
169
LETTER 2: From G. G. Berry to Bertrand Russell

Hamilton Road,
Summertown, Oxford.
May 7, 1905.
Dear Sir,
I must apologize deeply for not having acknowledged earlier by
letter your kind reply to my first communication. The reason is partly
that I naturally desired to comment on the topics revised; and in respect
to the most important of them, the question of functionality. I still find
myself unable to do so with any confidence. The fact that the contradiction puzzles so many persons of undoubted intellect proves that there
must be some widespread wrongheadedness which affects most public
thought on the subject; and it may be my share in this unhappy endowment which makes it hard for me to admit that the expressions you name
are other than functional. The nature of my difficulty may be perhaps
must emphatically stated by asking the questions: If xRy is not a function
of x, R, and y, what right have you to use the notation? Do you nQt
thereby imply just that care of constant meaning in the expression the
absence of which appears to be your ground for denying its functionality?
Your second letter raises an interesting point which is not entirely
new to me. I perceived long ago that my theorem on 'Limited Validity'
rested on the assumption that x'k always exists, when k is any class of
exclusive classes; but I shelved the consideration of this proposition till
after I had made my peace with this Contradiction and allied difficulties.
Meanwhile, the proposition seemed to me to be tacitly assumed by all
writers on the subject, and I thought it permissible and allowable to build
upon it such temporary and provisional results as are alone to be obtained, so I think, until those fundamental difficulties are solved.
At that time, 'all writers on the subject' did not include you, for I
then knew nothing of your work relating to the subject. Now, however,
170
Correspondence
I must beg you to be included, on the ground of Principles, p 118, l[ine]

l1-"(as is easily done)". The thing you then thought could easily be
done is precisely the thing you now say you cannot do. In this passage I
have just cited you are treating of an hypothesis according to which k and
k' are two similar classes of exclusive classes, and one of the possible
1-1 correspondences between them is such that each term of k is similar
to its correlate in k! You say then that it is easy to prove u 'k sim u 'k';
but to do so you must assume the existence of X'E, where E is a class of
classes of 1 + 1 relations: viz each member of E is the class of 1 + 1 relations connecting some one member of k with its correlate in k'. For,
given that u sim v, we have a class of 1 + 1 relations, and there is no
universally applicable recipe for selecting one single relation from the
class.
Thus, I take it, the existence of x'k is assumed by anyone who
asserts that every class of cardinals has a unique and determinate sum, or
who identifies multiplicative universally with the addition of equal numbers, or exponentializing with this multiplication of equal factors.
Again, the doubtfulness of the existence of x'k begins, I submit,
when k contains a o terms. That is, e.g., you cannot infer u'ksimuk' when
k & k' are progressions of (not-ordered) couples. You can of course,
arrange the couples one by one, but if you arrange this arbitrarily; that is
tum each couple into an ordered couple independently of all the rest, you
can only begin a process which can never be finished, and you cannot
assume that a state of things exists which could only be brought about by
ordering ill the couples. If you can claim the right to argue from the
hypothesis of one intellect capable of performing an infinity of independent acts; or of and infinity of intellects cooperating simultaneously; the
same hypothesis, I should think, would justify the general proposition
3'x'k where k is ~ class of exclusive classes.
If I am right in this, another consequence would be that we cannot
affirm that every class which contains 11 members however great a finite
integer n may be, necessarily contains a class with the cardinal number
a o The hypothesis affirms that however many times we have subtracted
a member from the class, others still remain to be subtracted; but it gives
171
no ground for saying that there is any method of defining a procedure by

which it is settled in advance which term is to be subtracted at every
stage. We have thus to face the possibility of a number {3 which is neither
greater nor less than CXo nor equal to it, viz. Nc'u'k where k is a class of
unordered couples & Nc'k = CXo
The subject seems to me to be connected with the questions raised
by my former argument about finitely definable ordinals, and with your
remarks on infinitely complex wholes, and also with the question of
whether it is possible to prove the existence of a class without giving this
means of isolating one particular member of it, as when one proves the
existence of the irrational by proving the irrationality of Y/2 or some
other determinate irrational.
I cycled back to your house this afternoon & would have called but
that I observed signs of visitors. Perhaps, seeing you have been good
enough to speak of our meeting, you might be willing to take a walk with
me some afternoon in the pleasant district you have come to live in. I
am, however, at liberty on Thursday & Sunday afternoons.
I am glad you & Mr. Whitehead like the theorem, and as I have
said, I only claim for this a provisional and conditional validity.
Yours faithfuly
G. G. Berry.
LETTER 3: G. G. Berry to Bertrand Russell
Hamilton Road,
Summertown.
Sept 6, 1905
Dear Sir,
I still have some literature of yours, namely several of Jourdain's
memoirs; before long I hope to bring them back. The treatment of transfinite ordinals does not seem at all satisfactory to me; a good deal is to
172
Correspondence
my thinking unsound, and the sound part is unnecessarily Cumbrous. Still

I am not sorry to have read them, for they have suggested several new
ideas to me.
One question in particular rouses my interest, and that is the relation between members of ten series w .. W2, , W-Y' and their indices.
If we think out the series of ordinal numbers by omitting every ordinal
which has a predecessor having the same cardinal as itself, we get a
series generally supposed to be ordinally similar to the ordinal one. It
remains so if we further omit the finite number at the beginning, and if
we now put this last series into parallel correlations with the original
series of ordinals, beginning at 0, w-y is the term in the thinned out series
which corresponds to 'Y in the full series. To me, it is a matter of doubt
whether even w, exists; but I am considering the consequences of this
hypothesis that there is an w-y for every ordinal 'Y.
Now Jourdain lays it down as an accepted theorem that there can
be no ordinal 'Y for which w-y = 'Y. I should very much like to know the
ground of that statement, because I think the contradictory of it can easily
be proved, on the hypothesis that for every 'Y there is an w-y.
An ordinal w-y equal to 'Y may be called a self-suffixed ordinal. To
define a self-suffixed ordinal greater than any assigned ordinal 'Y, we
form the series
{3 .. {32' {33' ... ,
{3-y,
defined by,
{31
{3-y
= 'Y+1
= W/3-y-l
The sum to infinity of two progressions of {3's (or their lines, which is
the same thing) is an ordinal 0 such that W,s = o. The least self-suffixed
ordinal will be the incredibly huge structure
(a)",
173
......
which tends off to infinity. Suppose we denote this by 0, and the next
highest self-suffixed ordinal by Oh then O2 , , Ow, .. & so on, the
suffixes of the Os will be all possible ordinals, and by the same process
as before we can get a self-suffixed Ov that is an Ooy equal to 'Y.
Thus if there is a good proof that Woy > 'Y, we can infer that the
series w, Wh w2 , is shorter than the series of ordinals. But I suspect
that the statement rests on a vague conception of "ordinal that can be
reached by Cantor's principles of generation" containing an unnecessary
limitation as to the way those principles can be applied.
I have no fresh light on the contradictions. Have you?
Yours sincerely,
G. G. Berry.
Hamilton Road,
Summertown.
Sep 19, 05
Dear Russell,
Thanks for Hobson and Hardy.
The general criticism of the former on the latter-that, so much
being left to arbitrary selection, the series of type w, cannot be said to be
defined-is one that I agree with, and is in fact the criticism I was expecting to find applicable.
But the other objection, according to which the second term of a
Hardy sequence must get infinitely big, is, I think, quite mistaken. You
174
Correspondence
will fmd my treatment of this point written out separately. Hobson and
Burnside seem to have succumbed to the temptation to make an illicit
inference from the case of any v to that of No.
I should be obliged if you would write out for me the leading
formulae for the construction of your No series. I can't reproduce them
consistently from memory.
It would be interesting to know whether Hobson would refuse to
admit that he himself has a shape unless he can find something else of
exactly the same shape. The uniqueness of certain aggregates is no reason
for refusing to assign them types and cardinal numbers. There is no
objection to gaining new types by arranging old types, as long as you
don't count types as having been made before they are arranged.
When the postulating method is sound, it can always be converted
into the genetic method. The only way to prove that a set of attributes is
self-consistent is to produce something which combines them. Pieri's
postulating treatment of geometry depends on the instance obtained by an
arithmetic meaning of 'point', for its consistency, and he might have
began with that definition and made the treatment genetic. The objection
to Cantor's treatment of ordinals is not that it is genetic, but that some
untrue principle has been mistakenly admitted into the body of "principles
already" recognized as true. If the arithmetical continuum has also been
defined by an unsound genetic construction, as Hilbert suggests, there
can be no hope in the postulatory method.
The difficulty about finite definability is relevant to this order of
ideas, because at first sight Cantor's theorem appears to prove that some
real numbers are not finitely definable. When we look into the proof we
see that it comes to this; in terms of any Cio of real numbers a real number may be defined not belonging to that Cio So that if the theorem were
taken as proving that indefinable real numbers exist, it would be by
defining a particular indefinable real number in terms of a supposed
progression of all the definable ones. If we are right in concluding that
'finitely definable' is not a determinate predicate, then the contradiction
disappears, not at the same time Cantor's theorem ceases to be a reason
175
for believing that there are real numbers which are in no sense finitely
definable. My suggestion at present is that 'finitely definable' is a common expression for a class of predicates; that this class is determinate in
intention, so that we know what predicates belong to it; but indeterminate
in extension, so that the predicates have no logical sum, and there is no
class of things definable as the class of things that have some one or the
other of the predicates.
Your new No is also a link that it is possible for a class of classes
to exist without having a logical sum.
A subject that stands in need of clear statement is that of mathematical existence. A exists mathematically, if the assumption that A exists, along with other assumptions, leads to no contradiction. But what
other assumptions? If primary ones, then A and B may each exist, while
the hypothesis that A and B both exist may be self-contradictory. There
may be no such thing as a mathematical universe: i. e., a system of entities such that the supposition of an entity not in the system necessarily
leads to a contradiction. I am not sure that the solutions of the contradictions are not to be found on these lines.
Having now assimilated all that I can from Hardy and Hobson, I
am sending them back. The next thing I want to see is Du Bois Reymond's discussion of the lawless decimal; probably the book is in one or
other of the libraries.
With kind regards
Yours sincerely
G. G. Berry.
176
Correspondence

Hamilton Road.
Sep 24.1
Dear Russell,
Thanks for your letter: I am just writing to communicate what
seems a discovery viz the proposition
I have also proven that No is well-ordered in respect both of f and

and that every well-ordered series is similar to a segment of it.
~,
Of course these results can't be said to be true because they involve so many contradictions. But I think I have proven them subject to
such reservations. You will probably find it more amusing to make
proofs of your own than to send mine, especially as I am so inexperienced a Peanist.
I didn't know Hobson had a more idiomatic shape than anybody
else; it was a random phrase. As for his pamphlet, I took down the reference to Du Bois Reymond, and that is all I should want to look at it
again for. Thanks all the same for the offer of it.
Does he call certain aggregates 'sphinxlike' on the ground that
there was only one Sphinx?
Yours sincerely,
G. G. Berry.
1. After reading the next missive, it is possible to conjecture that both letters (Sept 24 and Sept 26)
were written in 1905.
177
1908?'
Hamilton Road.
Sep 26
Dear Russell,
I sent you a statement of the proposition I have been meditating on.
The alternative definition of No which I was trying to get out appears to
be less simple than I had thought at first. The third and, I hope, true
version of this result is given on pp. 4 & 5. The idea which I have expressed by J leads to a contradiction without the introduction of any
ordinal idea. You claim, I think, the same for your No, but I don't myself
see how to prove No!x.:::> .x - EX without using ordinal notations of some
kind. [The notion I used for this purpose was that of epsilonic and non-episolonic classes. u, is episolonic if in the progression
we always have u,+,
U,].
These investigations, I think, demolish the 'consistent aggregate'

explanation of Burali-Forti, since we get the contradiction without assigning 'types' to anything.
If you propose to solve the puzzle by admitting properties which do
not define classes, you have to face the possibility of restating the contradiction in terms of properties only, eliminating the notion of class, as can
be done with the x E x contradiction.
I hope I haven't caused you to waste time by sending mistaken
1. Russell dated this letter in his own hand. The !heme of the letter suggests !hat it was written in
1905, immediately after !he previous letter was sent to Russell. If this is the case, then I believe that
Russell dated !his letter some years after it was written.
178
Correspondence
versions of my proposition. This time, as I also send proofs, it will matter less if there is an error.
Yours sincerely,
G. G. Berry.
Hamilton Road
Oct 8, 1905
Dear Russell,
I am glad the Ms interested you. The proof you sent me that
No v Ix. :::> .x - ex
is not unlike one I had already hit on, viz. that if
<p!x. 55 :yex.:::> .y-ey,

<p is an inductive property.
If you are still interested in the possibilities of nothing, you may
like to see the enclosed which is the result of further meditation on the
same fertile topic. I took for starting-point the definition I had before
mistakenly given for No and tried to find what it really does define. If I
am right, it covers everything which can be made out of nothing. A point
which I haven't cleared up yet is how to distinguish, by a not ordinal
criterion, between what I have called derivative objects and standard
objects in general. As to functions and properties in general, I don't think
I have any new ideas.
Yours sincerely,
G. G. Berry.
179
Dear Russell,
Hamilton Road,
Nov 2,06
Thanks; I will come up on Tuesday.

Since I saw you last, I have meditated a little on the subject of
lying, partly because I have had uncomfortable personal relations with a
real liar-a young woman who was in my service but now is not.
Now supposing this young woman were seeking a new situation in
your house, might not a communication pass from my wife to yours
containing a proposition of the following form:
Not all the propositions affirmed by Elsie Coombes are true?
Now, Elsie Coombes herself does not assert this proposition, and if
she did it could not be said that all her other statements are true. No
difficulty arises on ordinary views. But in your view, have you not to
translate the above quotation from an imaginary letter into the following
form:
It is not false for every value of x and y that x is a finite integer,
that y is a proposition of type x, that Elsie Coombes asserts y, and that y

is false;
which is equivalent to authorizing the class of preproposition of

finite type.
You may perhaps say that your present view allows types to be
lumped together up to a specific figure, and that you would translate
rather like this:
Correspondence
180
There is a proposition of one of the types 1, 2, 3, ... , 1000000

which is false and asserted by Elsie Coombes.
If so, I should say that here 1000000 does not mean 'a million',
but a 'sufficiently large finite number'? That is the higher type of propositions that are not apprehended but only denoted; and the validity of
their denotation seems to me to stand or fall with the possibility of abstracting from the type-number. An 'and so on' which leaves nothing to
whim must have a determinate meaning of some kind.
Perhaps you will have some light to show on Tuesday.

Yours sincerely.
G. G. Berry.
Dear Russell,
Southside,
Hamilton Road,
Oxford.
April 17, 1910
Your problem is one that interests me, partly because I once tried
to solve it myself. After vainly trying to prove that your solution is right
I went on the other track and soon succeeded in finding what seems to be
a proof that it is wrong.
The point at which your plan breaks down is fl; and the reason it
breaks down there is that if p. ;e 1, the ultimate position of p. in the
representations of ordinals less than fl' is nothing else than fl itself. That
is, in these predecessors of f I, all the natural numbers, except 1, have the
same ultimate position, and the only operative rule of priority is the one
which makes the smaller precede the greater integer. Then the series
which ought to be of type fl' is only of type w.
181
way
Now, how do I know that the ultimate position of JL is

is the limit of
l?
In this
....
w,w '" ,w "," ,w "," , ...

and I find that in the interval between any two successive terms of this
series every integer except 1 receives promotion at least once. For example, between WW and WW there is a series, whose type is w W, if ordinals
each of which is immediately preceded by 1 in the highest derivative to
which it belongs; let these ordinals be denoted by ~l' ~2' , ~W' , ~~,
... . Suppose, too, that in the representation of wW the first terms are 1,
JL2, JL3, , JLv' now, in representing wW+ 1, wW+2 etc, the number JL2,
JLI" receive promotion; between ~l & ~l +w, we promote JL~, , JLv'
... ; in general, if " is of form (2r+ 1)25 + 1, JLv finds promotion between
the formation of ~5 and that of ~5+W. Hence between forming wW and ~W'
all the terms in the representation of wW which have a finite position
become promoted to positions greater than WW.
In the representation of ~W' the finite positions will be occupied by
numbers which in the representation of WW had positions 1, w, w + 1, ... ,
w+ JL, These will all be shifted between ~w and ~2w' When we get to
~w2' all the terms of WW between positions 1 &
will have been promoted; and so it will go on; if 11 is a mUltiple of w, by the time we get to ~1/'
all the terms in the representation of WW between the 1st and the 11th will
have been promoted to positions in advance of the wWth. If then, JL is any
integer but 1, I can take 11 between the position of JL in WW and w W, JL
must therefore be promoted beyond the wWth place, and the ultimate
position of JL in arrangements preceding fl is higher than wW. But I might
here take
and
w'"
....
182
Correspondence
and
and the result would have been the same. The ultimate position of p., thus
exceeds every term of
(a), (a)
(,)
,(a)
(,).,
This is an incomplete statement of the argument, but I think it will

enable you to construct a complete one for yourself, if you have not done
so already. If you find anything inconvincing [sic], or any lacuna which
you cannot fill I should be glad if you would mention it; I may have
assumed something without real proof.
By "derivatives" I understand series of ordinals; the (~+ l)th derivative contains those which in the ~th have no immediate predecessor; if ~
is a limit-ordinal, the ~th derivative consists of those which are common
to all derivatives prior to the ~th; the arrangement is always in ascending
order of magnitude & the 1st derivative consists of those which have no
immediate predecessor in the complete series of ordinals. The ~th derivative begins with 1, ~, 2~, 3~, & consists of the multiples of ~.
The arrangement you give for wand WW seem to me faulty; in
the even numbers should all occupy places between wand w.
I think I shall now try to prove that the problem can't be solved-if
I find anything further I will write.
Yours sincerely,
G. G. Berry.
183
Dear Russell,
Hamilton Road,
Oxford.
April 20, 1910.
I have hit on another argument which it may save your labour to

have at once. It goes further than the first argument & shows that you
cannot make your method adequate by adopting a new method of constructing the representations for 'Y + 1 while retaining the old one for
limit-ordinals.
Suppose then you have a successful plan for representing all 2w
class ordinals in which ordinals of the 2w species (i. e., having no immediate predecessor) are dealt with by your method of the ultimate position. I
am going to deduce a contradiction.
Your arrangements representing ordinals of the 1st species are Nl
in number. Classify these arrangements according to their last term; we
have one class of arrangements ending with 2, another ending with 3, &
so on; since Nl members are distributed among No classes, some class has
Nl members; let the arrangements of this class all end with p..
Let the types of the arrangements ending with p. be arranged in
ascending order, tJh tJ2, ... , tJ~, ... The suffixes exceed any 2w class
ordinal; hence we have the progression,
Let the limit of this progression be O. There are three alternatives.

(1) 0 is of form 'Y+ 1. This is impossible since 0 is the limit of a
progression.
(2) 0 is of form -y+w. This is impossible because there would be
then an No of B' s exceeding 'Y and falling short of 'Y + w. Two
184
Correspondence
consecutive {J's would differ by a finite amount; but between

{Jvwand {J(r+l)w there is an infinity of {J's if of no other ordinals.
(3) 0 is of neither of the form -y+ 1 nor of form -y+w. Now consider your rule for the series which is to represent O. The
position of p. is determined by its "ultimate position" in the
series representing ordinals prior to O. Its positions are successively greater than all {J's and :. have 0 for their limit.
No other integer has a higher "ultimate" position. Hence if-y
is the type of the series of p.'s predecessors in the series
representing 0, 0 is of type -y + c5 when c5 is formed of p. and
its successors. But these successors, having all the same
"ultimate position" viz 0 are in ascending order of magnitude
and the type is either ." or W hence we have 0 = -y + 1 +." or
o = -y+ 1+w; i.e., 0 is of type -y+ 1 or of type -y+w.
The reason I don't think it likely that the problem can be solved is
this. Suppose the solution is found, its sufficiency has to be proved. That
is, we have to prove that, if ~ is of the 2nd class, the method provides a
representation of ~. Now the method must fail for ordinals ~ WI; and the
problem of proof will assume this form: show that if the method fails for
~, ~ ~ WI; and this seems to me the opposite of the hopeful task. It is
conceivable that the fact of ~ being < WI is implied in the construction,
by making it start from a supposed arrangement in a progression of terms
which are also arranged in type ~; but then the construction would have
to be such that starting from anyone of the possible progressions we
should always arrive at the same result; and this, too, seems improbable.
Where is the work of Hausdorff, to which you refer, to be found?
Yours sincerely,
G. G. Berry.
185
6.6 Alfred N. Whitehead.- Whitehead played a fundamental

role in Russell's development as a mathematician and as a philosopher.
They first met when Whitehead examined Russell for admission to Trinity College at Cambridge University. Afterwards, Whitehead taught Russell 'statics' at the same university, and Whitehead also advised Russell
on his fellowship dissertation. They became colleagues as philosophers of
mathematics when they decided to join efforts and publish Principia
Mathematica. Their friendship cooled because of their differing opinions
on the First W orId War.
LETTER I: From Alfred N. Whitehead to Bertrand Russell

The Mill House Grantchester
Nov 16, 1900.
Dear Bertie,
I have been studying Peano's arithmetic - Formulaire, 1899. He
seems to have prematurely identified his symbols with those of ordinary
mathematics. The result is that he is led into some inconsistencies. Thus
- [the references on left are to the formulaire, on right are for my convenience of quotation].
9 *1.1 a,b,c
cls.C:.
20. *5.4 b
No.::>. +b
23 *2.4 a
No.::>.a
25 *1.0 a,b
a, b
a& b = :x
a.Cxx u
b ... (1)
No & No ... (2)
= +a ... (3)
No.::>. axb = ab = O[(+a)b] ... (4)
No:::>:. (1)
n (2) n (3) n (4).::>.a+b = a b = axb!!
186
Correspondence
My belief is that the development according to his own principles is

much more lenghty than the one he gives-I have been working at it in
my interval of leisure.
Yours ever,
A.N.W.
LETTER 2: From Alfred N. Whitehead to Bertrand Russell
1902.1
Everything, even the object of the book, has been sacrificied to making
the proofs look short and neat. It is essential, especially in the early parts
that the proofs be written out fully.
[A. N. Whitehead].
[Whitehead's criticism of my first

draft of *1- 5 of
Principia Mathematica]. 2
LETTER 3: From Alfred N. Whitehead to Bertrand Russell
The following is a reproduction of a telegram Whitehead sent
Russell, when Whitehead thought that Russell had solved his contradiction.
1. The date was added by Russell, first in pencil at the time he added Whitehead's name in square
brackets.
2. After some time, Russell went over the date in blue ink and added the explanatory note in the same
ink.
~} ~,\g,.v,",,,f.. ~~
,
0-",,-."",",,~-,- ~
TO{
--f......AN~
I't / '
Plate 11. A telegram from Whitehead to Russell.
187
188
Correspondence
6.7 G. H. Hardy. - Hardy was, perhaps, the mathematician

from Cambridge University that Russell admired the most. In 1942,
Hardy wrote a pamphlet [Bertrand Russell and Trinity], as part of the
campaign to get Russell reinstaled, in which he discussed the particulars
of Russell's dismissal from his Lectureship in 1916. Later on, in 1944,
he also played an important role in Russell's reinstalement. Some of the
following letters were written as a consequence of the polemics generated
by Zermelo's proof of the Well-Ordering Theorem and Hobson's comments on it and other articles on this topic.
LEITER 1: From O. H. Hardy to Bertrand Russell
Trinity College,
Cambridge.
Oct 14'03.
Dear Russell,
I enclose a letter I received this morning from Jourdain (whom I
think you know), which I hope will interest you. It seems to me now that
my criticism was not sound mathematics, though it certainly seemed like
common sense.
In the case I took (the universe consisting of a large solid spherical
Earth-a particle moving in a line with the centre) it is plain that 2 observations do in general [?] the motion. But if (in Jourdain's notation)
we have 2 distinct solutions. viz.

x = -lhgt2
&
which are functionally distinct x
+ Ihgttl
x=O
= 0 not being a particular case of x =
189
-Ihgt2 + CIt + c2 And I seem to have been misled by this & by not
knowing enough about the nature of the general differential equations of
dynamics. If we have an yO
y
+ Ay + B = 0
(say A, B constant for implicity) we have the solution (i) if we have any
particular values of y, Y (ii) if we have 2 values of y; so it seems the 2
things are the same really, and my argument is wrong in spite of its
apparent plausibility-for certainly the datum of y does seem a priori to
contain much more.
The other point about the a's is also interesting but I don't think
Jourdain is write [sic]. For if anything seems obvious it is that the series
of a's does exist. I had tried myself when writing my Q. J. paper to set
an argument for every a being in the series (but came to the conclusion
that I couldn't) out of Burali-Forti's contradiction. My line of argument
was to show that adding one more to the whole series meant nothing if
(and only if) it was already similar to the aggregate of all Entities. But
even then the series
being ordinally the same as
there seems no reason why one should not put a o on at the other end.
However if you have (as Whitehead says) solved all the difficulties
about the greatest cardinal etc I suppose you have solved Burali-Forti's
contradiction too.
Yours sincerely
G. H. Hardy.
190
Correspondence
LETTER 2. From: Bertrand Russell to G. H. Hardy

Bagley Wood, Oxford. June 22, 1905.
Dear Hardy,
How odd that our letters about Hobson should have crossed. I am
glad to know his particular criticism of you is wrong; I was not clear
enough about your way of defining your limiting terms to be sure.
The other point of criticism is not, I think, quite what you take it
to be. The question is this: Any No which has no predecessor is the limit
of a number of different progressions. In order to fix what sequence is to
represent (say) WW in your correlation, you must settle which of all the
progressions whose limit is WW is to be chosen as the one out of which
you construct your diagonal. That is, you must pick out one, & only one,
progression from each class composed of all progressions having a given
limit. Since the number of ordinals having no predecessor is infinite, you
can't simultaneously pick one term out of each of the above classes of
progressions unless you have a rule. It is, as you will perceive, a case of
Whitehead's multiplicative class. Put
= class of progressions whose limit is a Of.

Prog(a) = p 3 [(3bJ3). {1 < a . p = w ((1)] Of
w(a)
i.e., Prog (a) is composed of all classes w({1) where {1
<
a.
For your construction , you require a term of
where xCk
p C 3 {p uCk : u
k .:J. P
n u I} Of.
i.e., 11 is got by picking one term out of each member of k.
191
The norm is required for this process of picking out one term from
each member of Prog(cxl).
You might suppose (as I did till lately, & as most people seem to
do) that if a set of classes consisted of mutually exclusive classes, none
of which were null, it must be possible to form a class composed of one
term out of each class in the set. This may be true, but I know of no
reason for believing it. The point is, that there is not necessarily any
definition of such a class as we want. The results are very awkward.
E. g., one can't (at least I can't) identify the two Dfs of the finite, by
induction & by non-similarity of whole & part. I can't prove that if you
have an infinite number of pairs of boots, you must have an even number
of boots, unless (as is likely) your right boots can be distinguished from
your left boots, in which case the number of right boots = the number of
left boots. But if you are faddy, & have no difference, there is no way of
selecting one out of each pair (no norm in Hobson's phrase); thus if k is
the class of pairs, xCk may be null, & you can't prove that the number of
boots is even.
Hobson's norm always concerns multiplicative classes or close
analoges of them. If k is a class of classes, & fu a function of classes, f
is a norm in his sense if u k.:::> fu u. In this case, the class of terms
of the form fu gives us what we want. This is Zermelo's assumption in
proving all classes can be well-ordered. He makes his assumption universally, in which case it is equivalent to (3f): 3Cu.:::> u. fu U, i. e., "there is
a certain operation f which transforms every existent class into a member
of itself". This is a rather tall for aRp.
Yours sincerely,
B. Russell.
U.
P.S. I see I have forgotten to answer your specific question about Du

Bois Reymond's decimal. I think it is true that a decimal in which the
digits obey no law is indefinable, & that there is no reason to believe that
it is an entity at all. Taking the scale of notation Z, every class of finite
numbers defines a decimal, namely E 2-x where Y. is the said class. For
this it is necessary & sufficient that there should be a class Y.. But a class,
when it is infinite, can't be defined by extension, but must be defined by
192
Correspondence
a propositional function satisfied by the members of the class & by no

other terms. Any such propositional function constitutes a law or norm;
in the absence of a law, therefore, we get no class & no decimal. [For
the existence of a class, a propositional function is necessary, but not
sufficient; the propositional function must be of the sort that I call ~
~, i.e., equivalent to "x has the predicate so-&-so". This results
from my contradiction; & it solves Burali-Forti & all other paradoxes I
have been able to discover.]
Hobson is not clear, but I think he is in the main right; except in
the following points: N 4, (2), is vague, & NS 6, 7, seems to me mistaken. His criticism of you, also, as you point out, is mistaken in one
point; in the others, for ought I know, you may be able to supply what is
required. I think of answering him, but not very controversially: much of
what I have to say is out putting his views more clearly.
B.R.
LETTER 3: From G. H. Hardy to Bertrand Russell
Cranleigh
Surrey
June 30
[1908].1
Trinity College,
Cambridge.
Dear Russell,
I have thought about what you said in your letter and discussed the
points with Jourdain, whom I saw recently. I see the point now quite
clearly, I think, & I see that the 'norm' which I suggested in my former
1. Russell dated this letter some years after receiving it. It is clear from the topics under discussion
that this letter was written in 1905. In addition, there are other elements that confirm this date [see:
Grattan-Guinness 1977a, 54]. This letter has already been extensively quoted in: Grattan-Guinness
1978a.
193
letter does not meet the real difficulty. I think that further analysis of the
f-numbers might do so (in the particular case) but I can't say I think this
very probable.
The following points still puzzle me:
(i) Is definition by extension logically restricted to finite classes. In
'Pro of Math.' (71) you seem to say, No. Do you adhere to this, and if
so, doesn't it affect the root of the multiplication class difficulty?
2. Is Cantor's proof of 20<0 > <Xo open to the same objection. His
'diagonal class' is defined by Cn = 1 if Cn n = 0, = 0 if Cn n = 1. Isn't
this perfectly definite i.e., without 'freedom of choice'? If so how does
denying the X' class solve the difficulty of the greatest cardinal?
3. If Hobson is right, is there any reason for supposing that the
continuum is not enumerable, or even if it is not, for supposing that C >
20<0 is not all we can say about its cardinal? I am not even clear we could
not prove that C is = 20<0.
4. In any case, how does denying the X' solve Burali-Forti's contradiction? This at present, I don't see at all.
5. Do you regard Burali-Forti's argument as unassailable, apart
from denying the X' class? I am half inclined to agree with Schoenflies
that it is at bottom unmeaning, though I can't put it even to myself in a
satisfying way. Isn't it possible that your original solution was right
(denying that W is well-ordered) in the sense that although W (as Jourdain has shown) has undenially some of the properties of its members, it
has not all, and in particular not those which make addition of 1 logically
possible. I daresay you have some more general and fundamental way of
stating B-F's argument, but I don't know what it is.
6. Jourdain showed me a letter in which you say Konig proves 200
~ a l I don't think he does; he proves 200 ~ ap where {3 = 'Y + W, in
particular ~ aw ' This, so far as it goes confirms Cantor's view, according to which
Correspondence
194
for then every 8.y is a
unless'Y is a limit number. I fancy Konig's argument can be intended to

prove 2&0 ao, {3 = any limit number.
7. Is the case of the decimals really a parallel to my particular
difficulty? In the case of the decimals, the X' class ~ exist because we
can point to particular members of it, e.g., 11111 .... My difficulty is
to specify ~ member of the class I want.
Doesn't this show that Hobson hasn't really got hold of the point;
for he makes the two cases parallel? The difficulty for Jourdain, Zermelo, and so, is exactly the same as for me. It also applies to Konig,
Bernstein, etc.-and it seems to me that if Hobson is right his remarks
about 'the debt which we owe to Cantor's genius etc' must be taken cum
pano.
8. Is your agreement with Hobson on the 'norm' question~? I
rather gather (from your excepting 6.7) that you still uphold the aleph-series but I am not clear.
9. (A minor point). Did Du Bois really deny the gesetzioses Decill1fl1.. I haven't done more than look at his book, but he puts the objections into the mouth of his Realist (and I thought he rather sided with the
Idealist).
My present position is rather like this:
I admit that it seems impossible to prove the existence of the multiplicative class in general, and I don't see how to do it in the particular
case I want. I seem to see pretty clearly that for any 'Y there is a special
195
fund-sequence, one which anybody but a born fool would choose, but I
certainly can't formulate a finite set of rules for it.
On the other hand to deny the class 1) seems paradoxical, as an
idealist would say 'does not satisfy the intellect', 2) does not seem logically necessary, 3) does not seem to me (though I am open to conviction)
to solve in an obvious way difficulties otherwise insoluble, 4) seems to
make hay of a lot of the most interesting mathematics.
tes'?
In connection with 5 do you believe at all in 'inconsistent aggrega-
I'm sorry to have to bombard you with such an alarming list of

questions.
Yours sincerely,
G. H. Hardy.
Correspondence
196
LEITER 4. From: Bertrand Russell to G. H. Hardy
8~
(J vvJ.,
<'-'<
Ih (-.J.
'Jr-lj J
I )' dT .
fhv.4
y~~ <' ~;....~ .... ..A -(~". )~~I G.r ( ..... d "4>f?rf....b.
7 ~ ~ v-" tv-.:. ~ -- .,.../...

(I) . .ff C; ~....-~ tci-<~ I
a,...a.
q,
___ !l:
~ ..
of
Of
,.b....
~z.a. L ....
<>'>.
,dt-f)..-.tZ,
u.>w"...... <~ ~O...CX?. .l.,J.;-'h:""-1 ~U>. ~ r~

......J~~ J} ~ 'r~ .
~~,.."" I!.
t <Co ...
Q.
C- - ; f)c. '=' a. . ~ . >< " Ci .2>f
~~ c..~;Jc. .......-L-.,
~ c.t... ,,$i.V1-Wck, J1-...
vA . $'fPlrk.;f .Df ,(, ':;', -r,.................tu h.;. ~
c.oJ.......J. L
10
.;,,,~ t~ ~(. fL,
t, :.. ~J ; te... ...:....f~~ ,.u..-J. ,O..Q. ...,.:,tJ: ~
u-It.
~ ~ ~ ~ Q.... ~ .n f"J),~It~, ~ Jl- c..Jt..f..; If

r':"'" t-- ~ a. 'Z.J- c:t.- ...d.......v, t,Q.~ ~ h:>.L ......-.t ~ O~
0/'
)<
0.-
';"~<n--) ~ ("G..12..-.
~ E:R'>. fa:. r .J.
1.--- ~ .,; (,.

tt.c..t .L ~~
,g
90.):
t';i"
~<c.:;. ';,
AA-
,......."'"' ~ ~~ '" G--.. ............."
ll",a.
r:.:.- ....-..f.;..i
":'fga.)'f"G.:l.~aJ
.:-
c.. _ _
~~...9. ~ D)- tt- t~~j}f
~ f fC;,c.f .
~ cJ. ___ ;.,. ____ o..-f);.- ~
(2).
'prt ,.:
M "l-
_ .,...c. ci-e....-
? .-,.., ...... J
c.,..a.. .
-.c.-,
0/(11<: J. 2'" >CI ...... ~.k rb,..,'/d

~ a- ....-.Y> .,.....:~ ~ j-oJ.;j ~
Cl-{~~ .:( 0( = Ale: T ;) ,_./~ V= ;~, - ~;> = c ~~ ' 7. . . "i.. C""?ool
.. 2 "~ >"Clot l-L r~
nI f6r'.r, ....L t A. ,<: ....-----,
.~~cJ...~ .J1. ~.0'
'
o t...:.-~ tt:.,.. ~rl, ~&. ~ ~"~..4 r~,
Cc...J-'s
c ___r7r-e-c~/IL.,
('r~ tf- ..,R~'

_-
,fI,;-V '" ~
e,,"-r ...... V
~ P.~
~
,7_ Co,
S...c...r..
So
Ale' T =: 2 pc "":"
&~
.JL<o -
e.-.-.
.... ctb . ..::>'"

I ~':- V. Ie cL.
c.J...e..." 'rltC)
' "", n-I , _, 1-0 , '- ( .~o'""\o.c.~ .
~ V-s':""" c:..c: .... <{'4 l'r~"

j
r,,; V
"'~ ~~ ~ ~~ 8. (L ~.k. _
Plate 12. Letter: Russell to Hardy, July 2, 1905.
197
I ~ .... S',, ~ ~rnf In

'Z...<> tL
~ I l-Jo?'''''':',
J-. - ~ ~ ~ .
l~.-t. ' o.o.-/r d
/,Ik.C.
........
G .
" c~'fv,
pO'
aQ2.
11....
1..........
,0
u.-~v-:> urL...a.. ~
'i?~ J
'- CJ1,,'
(jl~ U. c..~
(L.
\.A.f~
" ... r
7<
fjfc~cj , '-'
-t:t. ~f. "1 ~ " . ~
'-;~ o~ '"
'CAU 'J.:-o./~
""4 c...
~ l.!
L~~ S'-...I
R.
n..
~)
2' ct.. < 2 , Of.
Q(.
CI.o 't.,
LI
o.l...-~
~~~
pen- ~ cd;:;;.
(t. Q.
S'o..-
lie. Cu. :::
Alc~7:
~~.
.....1<>->-........
~~ r~"'" Jj ~~ ~ '-~~
.;J.-'"...,..LeI.
~rn.n.J...:.~... ,.., q: u........:....... ~ ~
,~ ~ - f.-~;'
~~.J.~ ~r~9-
"') _ 0I? ot
r.
-Jlh .
v. ~f:Xf 2l.
Yu-.- ~ t:.. (L ,;...J~ 3 u.re...:..o..

~ ~ f)<. an c;.. ~J.."... . S'~ ~b
~ 'H.c..
t...
II>
{f~, 'f ,"?

~ ~ ~ u.N:f. '~'1- C-O"""J a;:.
~ j-, J...
S CA..-I:-- ~ ~
"
~~ V,." o.,. ~~ ....
m-
OIl ~ fL ...-d-.,.J '2
~~ c.~,
, 7j x.
~,~
-~ ~l C k;".J')~
],,)c cJ..
9 ~ ~-L. ~ <k.o....&... iT>

r~ i..,..,. ;
.... ~~
C'<>.(,~ .~ ~f., ~ ~ ~- 'jJ-..J).L
r1~c.L ~i '\'- sof.- <>'>~, ~ ..; ...... ~9' k,...1.. a..w ".' I 7---4J
a.-:......... ~ ~- .;... ~b.!{;J. g...1.f 4.b ~{';J 4 u .J7-Ca. ft-.. V2...i- ~ ~
S\
.Jo
t.
o.r ro
Qwn..A,
(1__
a....
c.
<
( .. .:...I<:..-<n.b (~ ~rk'=""" .. r~, ~ ~.;,
<1
~L(-,:.rt.-J>
I
So ~c--\-
do- .n.t..-
rl tb...-
Vl..
f1.....- ~ c..-.,
J,
0\.-
L........ So:.t.. . ~ ~
~ ~
l4i C
~~ ~Alc,,,Q.,rlL
j
"' ...............
f1.... J... ~
> ''It
~f- .....
"0.--:"'" "'-
'3~.
tL. .......-.4-li,ftR... ~/":""~ .... ' 2~ t.O ~~/,,.JI,.
4;.c..k. ..c.. ~
).
4. ~ <>--- ~
n....-ll cJ~
...
so.. era.... Q h ?
2 oNe ' ..
ct:&
c;.lI ,?~/~L ... 1..:...
..._._0/<'___....,
c..Ca..,..,. s~ ~ - i-J.i. 'II- v,
<:.a..... .Lb, ..,..,......,. If.,...(
~c:..J.. ~4
tL
...n- lL, U- ':'-f~
I")...:,..
. . , - d::.. ...-...J..,
~ ~ 7 c..oJ2Q.. ~ r:~":"" ~ ~ .en ct.. ~'4' ~."..

I~d r:""""'" t- )'0 UCI\oo
fJ --.Il.. W . l.)' . "1...I;/'<) 0'\ i?... r~ ' r.?
..Jr a. ~o.-~ ~
~ "'" it'; or 0/ > .rt__ .... 1 ......... .......L......I~
I
I)
;.., _"-...Il- a...~~~~ ~n.
0)
rI .. 'Z .
Correspondence
198
.r ) Ll ..
A~ ~ ~.bQ.-"'a ) Q"-f
~
(,. I"
.7~
~ .L. .J1. ~ ~ "f....,....~.....:.......
7.k.,
r,
...........
~ 1'~ ~ f-..h, 0
J'
t ,n..
~ -.../......" f) .......... ''''' _ ",',""" 10....../7

"... , (~ "-<> ~~rf'Loo) 7
Ja,,--,> '-- a....J.. ........ ..,."
? I.- ..... ~
cn--"-
...... c.-/
,, ~ -
f....C
. - I:f' r ' -
Co--
~ -.. ~
cn-fJ.
. :. . y.--- ~ J. a...o Iu
c..o~ ~
'l;.....C.
'r:J. .
Wi 77;,
q...5f- G... ~, c.....t- ~ -r-~""..,. -;.b
J.. ~~ ( ~c..Jl..
.
/1
[':l
' - "~ft!...--J.
~'9l' III".
~~ .7 0......- <, on ": ( w ~v.c.-c-=- a., rJ?'=i.,ct<f/
r ____ (....
~ J. t.i.
.J7.
t..u.
~",--of1
c...... -A,.
~
?-..1 .~ ~ ... ',ef ~ ~f~ 01~"

'h~'f,:<n-.> c..:~ W-.
(,. 9 -....-...J. L~ ... -,.
0.....
a..~ L-'
G!c..:".f':"
oQ ....
~t-L
a... :. ,.
"
cJJ-.".J-
p8
_.', "
~~
L&..~
1-
Ko"~,f- q. .2 -<*" ~ -'
~~ 'l"'<? oJ.;.~ .
;
..
NO t
~ 1- C- "L....c.......:.c.J2. ..:.(' t-J)- ~
I a.e~ ~~.;.
. . ,. .
a... .. a.'t"..."
t.. V'-'~.
rOLIl~ (L J.....-o'-~.
""rcA.
4....r
~ scx.f- ~. J'.-f;~v- n;
a- 0
M cf). t!:..
k. ':"'~, i.I- ~
~ ~ I- tc. .Q~
cY. 'J' ,sJ,.,.,-...~ 0 ....... " .,.~ Fe. ~ e:..........,.., i . G. . ~~
L... v,
i.d ~,.o--
0-;
0....
CA.o-
Q.-">
tl-
~ J. c- ~-.:rI'I.. f J'~ l J ~ ~ t!-.... ~ .....
((;'\..~, +~ M-k..):'" f e " .

t~{~ ~~~> lU4..c..-@~ ~~'t-cJ.c-..,
.:- ~ D:) ~~
~
Uo"L
~ ~~
~...:... ~J- L
:.. cJ.,," ;
I.
f'.n.-
:7.... ,----,--
~ fn<11-c-
~oCr ... ,~~.
!J. a.J):.~ cL.t.:.~ u:.. f..;....i../-........-
L.... to....- ~ ~ t:t .;...~~ v-f J.J'~~ ~
r-:/-
~'I-< ~ . <iL:.,9- fo.- C04 .t~, ~ ':r"'c~,

/~
tb..t-c. ~ "'I ~ C1 .. ...Q,... .. , L._I- ~ Q~ ~ fn
Q
~II\-~ ,f! . ~. ~~ f< ~
cla..t--~ ~ ./Z.;...;I.. ~ ".

,. (2u~ ~ ~
.........0..1 ~~ ~a
c.Pw-
9 ,f..:...~J(
":"{-~JAI> ~
Co"" ~ ~.9~.
199
'9 o.er-- d:.c:..+ ~ J 7 ~ o12-Jl. ~ kOA- (;"(~ ~ ~

~ .Lc ~ ~..,:u. e..:.. df~"~'
~.
'tP, -.J....:o
cJ=..A-9<Af>fV"-' ~I:t (~, ~~

L:0l..,..." i Of- ti-. C-_ _J').4>.L.JI,.,...." ~ ct.
~ uo-d;i..."~
~ p. ~ {...
.....-a...
SbL- ~ ~ Q... ~ c:t... '.~ -('~ ~k ..:.~. J, IEtJ../'t- ~_k~~~.. 1'l~JJ.:.f'l.(~
~ ~ ~ ~ V5~ - ~ 'r ......-.a....,. ct..:. cLOJo. 7r/uu

'"1',C...e.a. .e::.. P.J2'1C~ ,,: cg~ ~.:. a...rv.:dd~ ~ thc..f.
~ c:: f #-r~ ~ ~~ ~ ch- tft- ~ ~!)It~ .;.. ~ d)- ~~ I,.J} Ch.-.... 0...
a.t- Cl.0>, 1'..
tr/!!
t!w ~ ~
&.:,~~ ;.,... ~h ~ lw2,, " ~ ~a.{,
rfl-~....u-~&..~ ~aJ)cL.~q ~~
~ a.>tA. Q:k... tA.. "..6Jk"", It L. 01h(? ::. o'l7r,4lo, d' < go .
9( &.-..... c.b""", (7" L~ ~ ...... ' a...u:. -- ~.:..........t2, 7 .
") ~ ~ ~ ~ R-C,...-J.. cL..--s f~ ...........
<&..0
7.
".,...,,~_<
')
~",,' C,ia/l.:- ~~L e.,-...J.
tL.
. ;. .
0-..
-...Q. ........ ~ ~ ~ j, u '" f ~ ~ c....... ~

tl.-. / f. ~ c...r-L....: ..... .;.. ~ ~-rJ....t, -i' ..... 0... ~ ~~~~- ~~~
~r?lc.-IfC!o-1
Lu~'h~ @~ ~a.~1cD....)< ... ~ .. 5t ~:"
c..a.A M
'!...
Q.
/?r J2. c' Rc
Q: P, qC''R,
?Nr}.
cCp,., c <i>
~ /I: ::J.
~.,
,"(j)-fj)" C<R c
[~
...
1""'''
xCH;e "c C
ifr:s><:>7C.C1l.'J~ f"lll{'
~ """"l )~ Ct:>..O.A..>"
/-: CtaJ,,'2.a-.e.:).
~ ~ ~
Go
'/~
(h..a..
'3< ~c h..-c..o......G....-~.
x Cf1 "4
C
~ 1-0 ~ ~ f).. h.
,,)
r J' VA- {--'
( ~ rC 'k
~ u...o-Il-~c.JI~ c~ a.oc.:..~ .
200
Correspondence
f"d?
<x.
=:.:
7< ..
't"(r'-,,)
'=-:
~<'f' ~f~'l~~1'''( t''f)
J.
f<,.
ot.c'f"x
~r4{ _ ~
:>: t"Cf:O fc.4 . :J."" t('(

:> : " 'to -t''t
e()
~....-.~ ...c.'f" r(b~ s:gf).fcCf"t~(r(f('f)

~
r<.f
~,,~
,elf" f'~; "'''(tp"x. !fx)
N-'
<.(.. til
r~ f4('t),"'c(Cr<t)
~cJJ... ~ ~ ckJ, { 'f1- f-
+J
tL ~ cJ.a.u.....
~ ~ e ~ cp''' ..... ,.< X , fIe-.Ua. i.f: <qJ -=. cf1cvv.

~ ~)
CCC~ e- 0. ~ ~)CA.IJ,..,..Lp'~/~~'f~~.
l
M ~) ft'"-(/.K
~~. ~
l....
X ..
WCoJ.. (..
q>.~ ...
c....~'W1 ~tn.....~ ~
C'4'C)C
()."I
J.'2..r'.fr':J ... ~cJ(
~ ~)Cf'~ =-~"1 ~ It.... o.Ill~~~'>J w ~:.. ~
kdi-,. . 1J1~j ~.:..... ~~ tt.. ~ ~ ,c=~..J ....C~C). ~ h.

201
"-
c.f-,)
':1 't
:. if)('
C
~> ..:t.
= CP'"f. . . c.f
...,
Jc.
~~/tV- ~~~fAH~."~~~~
~'~.;..~&J.
2<t'>ri.
~- dh. ~~~
~ ,,~~.~.~C.((l)<.';' p~~ ~~: i..I--~ ~

~~ ~ ~ ~ Of !l. S'~ : (/) ~ cSt til.-. (L-- ~(k.)
c.a..o&
~~-.o..~cn--L~~ ~~n-~(~j~n. k.Jl.~;
~) ~
Of w...:&
~ ~ "'<d"/Jc.
c.D
o....-...:..-~
. . :. . ~J.
~
~ ~ ~I-o -:0: .. Co.. ~ ~~. _9 ~ or~, (J.J;':'-I
~-(,e.. o 0",'0 ~ 2 ~~ ,~ ~ ~ :..- CJtt ,cJ1s.
1l.c. ~ (,)~~ t:Aak ~coOJ..~~,l~
9~ Ct.) 9 CQ.Ct2.. ~~
Ik ~ (0 ~
:x
P"'""'"'
~~M";'" ~ ~>~c:t:.~CZ)c.....J2.D~<L."'~
iJ. ~ (.. a.. ~.-~Oa..-A- I. en-. JT&' ~ U~. ?I-JJ 1I/.c-
~h ~tl. ~~In.,e..,.. )~~ ~ ~0AI1
'7-
~Q4~' S'u:....:.~~ ~Of ?rvL,;.J.~~
~~'-f--7.I-o~~ ~.r..,...o..~~~
~L~-"'-"'-'1; ;f.::+ "'~6e~d1;..~~c.~

~ ~tJJ.J1.
'IV"-:.-.-d(~ .:-~~~~~.
~ 7
0..-
~ tL:... ~ ~ ~ ~~ d..
~~t"-~G"4~ f,-rW) .'.
..
y ~~ ~
CZ-Ih-a...-A
~a .
.......
o?-
Correspondence
202
::
,qh-).~ru
'"' i (x. t?.:r ')
7$<:1
fg,.))cC?~
;(~ti})
~ lta:r)' x (?~
"'~$fi( ...
=:
UvV.
~f,)c,r'J.<.
'V"Ck
CJlnC
fA.
,"tt.
= tiC...
('X:'
U-t..r
QJ-iJrR
to =
1\
~~
CL"
'"
cr
5 (j=-1<..) " cL.J.......... ~~~ ~ j~
:"f"r~'''''-('tr)l
==-
..
. '0,: ~"rr:-:J. :>c('&t.
!}- 'LI. = ,,";.(r.r~ CA. vc: 'U..)

o.-g
(l ..
~
::
'V.Jc.fv-f
..........;,.f...;.5~6J-
Q:><~)-(D<R) [<;:r~, ':-Cl-~]

,...Jla..?~ ~ ~- ~ ~
--etl- c.O..::.....v.. '" ;(:Jt;. rio x)
::.
~
11ftc cL,..
e..
~. u1- tr. :J. lJ.

II}
~ f.!.. f".'''t-4tt.? ~ ~, (" ('
V' r
(I
'lr '"
~v
~/~
;f tf'>c. u ~ .. L. f cf)- ~ ",)
tp"u, lob ?'<>c.J" ~ 'f.'~ t;. 'lI.'~'J
fl...,,(fflr. ': ~ ~cTt:lc.) cp"o,........;....,Jl ~ ~~ cI). u.-..:...~.
,2... o.prlh-r' .... :-
'5
J. .:: : f
Cp'}C
Jj-f ~t
203
LETTER 5: From G. H. Hardy to Bertrand Russell

Cranleigh
Surrey
July 5
Dear Russell.
Many thanks for your explanations, which relieve me not a little. I

am particularly glad to hear what you say about the question of 3'x'k
existing not being involved in the explanation of the contradictions. I
haven't fully mastered all your letter yet (my symbolism is 3 years out of
date) but I think I see all the general points, there are still one or two
things about which I am not clear.
1. As to Of. by extension. I don't see that

l'a V l'b = x '{x = a. V .x = b}
really does away with it. I mean that the disjunction A = a. V . x
really defined by extension. Then as regards the disjunction
= b is
A = a . V . x = b . V . x = c. V ...
exactly the same point seems to arise as in the case of the ordinary notion
of Of by extension. Why should not x'{x = a. V. etc} be a valid Of
without a 'a/\ [definable in finite terms]? (the number of x = a 's not
being finite). And doesn't the case of the continuum seem to indicate that
this must be so? If its cardinal is 20:0, and if Cantor's first proof of 200 >
Ci o holds, I agree that it seems we must recognize 'classes of integers not
defined in finite terms'. [In fact is it not necessary to suppose that a Of
is a finite no. of repetitions etc' is too narrow].
C
2. Does the theory of cardinals o! proceed as of old by way of

x'k? If so, taking the instance you give, in which the k's have no xve
class, but k + 1's can be defined which have, can one avoid
204
Correspondence
k = Nc'k. k = k
1,
~ =
0, (k
l)a
1,
supposing for simplicity the cardinals of every k all equal & infinite. In
my case every k = 200 P < O. k = w(JJ) = els. of progtns whose limit
is P1.
I should not like to say that anything so abstract and general as Zermelo's axiom strikes me as obvious, exactly: but I will go as far as saying that the more one thinks of it the more paradoxical the contrary
seems, so that, unless it appeared to lead to contradictions I should (in
default of proof) be disposed to assume it and hope for the best. I suppose that Jourdain's & Zermelo's arguments are not unassailable even if
3'x'k, depending as they do on Burali-Forti's contradiction. I always felt
with Zermelo that it was the end and not the beginning of his argument
that I found hard to swallow. Bernstein & Konig on the other hand depend simply on 3'x'k. I thought for a while that Bernstein's proof of 200
~ 0/ 1 was independent of it; but he selects one out of all the ways in
which an w can be arranged in a given {J, so that he seems to be in much
the same position, as myself.
I daresay I shall want to ask more questions when I have been
through your letter again.
Yours sincerely,
G. H. Hardy.
205
6.8 Eliakim Hastings Moore.- E. H. Moore was, perhaps,

the most influential mathematician in the United States of America at the
turn of the century. At the time he wrote this letter, he was the head of
the Mathematics Department of the University of Chicago and, a year
later, he was the editor of the Transactions of the American Mathematical
Society.
LETTER 1. From E. H. Moore to Georg Cantor
September 5, 1898
My dear Sir and Colleague From Mr. Professor James of our University you know that our
Mathematical Club has been given during the last year considerable
attention to your Mengenlehre. - - - In particular, with reference to the
well-ordered totalities and the corresponding ordinals (carry out a hint of
your recent paper "II"). I presented in the spring the two theorems
(which you probably have yourself established):
ex) The totality {7} of ordinals 7 < a given ordinal as arranged in
natural order, is a well-ordered totality whose ordinal 7' is
7-1, if 7 is finite
7,
if 7 is transfinite.
And, conversely;
(3) The ordinal 7' of any (definite) totality {7} of ordinals 7, full to
the bottom (i. e. such that every ordinal 7' < an ordinal 7 is itself an
ordinal 7 of the totality {7}), as arranged in natural order is
I
the largest ordinal 7, if the totality {7} is finite

an ordinal 7', if the totality {7} is transfinite,
206
Correspondence
where 10)
T'
> every T,
20) no ordinal
T"
exists, such that
T'
>
T"
> every T.
Now, barring error in these theorems, (3) shows that the totality of ill
ordinals is not a definite totality in the sense of your fundamental explanation of paper "I". - - At least, this is the only explanation I see of the
paradox resulting from taking for {tl the totality of all ordinals.
[My theorems have a certain contact (of content not of origination)
with the two papers of BuraIi-Forti in the 1897-8 Palermo Rendinconti-- though he is of course working independently of your recent papers].
I am hoping you will honor me by statements as to your solution of
the paradox above. - - If my suggestion is correct, does it not become
necessary to introduce (besides the analytical logical processes) ~
strained synthetic processes as basal elements in the Mengenlehre) - and thus to take a general position (analogous to that of Kr6necker on
"finiteness") and to build up a theory for each distinct group of constructive processes utilized?
Believe me, with very high respect,
Yours most truly,
E. H. Moore
Professor Dr. Georg Cantor.
Halle a. S.
Tables
Table I.
" An Analysis of Mathematical Reasoning"
Table II.
"The Fundamental Ideas and Axioms of Mathematics"
Table III.
"Principles of Mathematics"
77
Table IV.
"The Principles of Mathematics (Nov-Dec, 1900)"
93
Table V.
"The Principles of Mathematics", Part I: (April 1901)
103
Table VI.
"The Principles of Mathematics", Part I & II (ca. May 1901)
104
Table VII.
"The Principles of Mathematics" (May 1901)
Table VIII.
"The Principles of Mathematics", Part I (April 1902)
Table IX.
The Principles of Mathematics (Final draft, January 1903)
66
72-73
106-107
118
124-126
Plates
Plate 1.
Bertrand Russell (1907)
Plate 2.
Russell's contradiction
Plate 3.
Lord Russell
44
Plate 4.
Lady Russell
44
Plate 5.
Russell (ca. 1883)
46
Plate 6.
Alys Russell (Dec. 1893)
57
Plate 7.
Part VI Chapter III Folio 54a
99
Plate 8.
Part I Chapter IX Folio 22 (May 1901)
111
Plate 9.
Part I Chapter IX Folio 23 (May 1901)
112
Plate 10.
'The Contradictions' in Principia Mathematica
133
Plate 11.
A telegram from Whitehead to Russell
187
Plate 12-18.
Letter: Russell to Hardy, July 2, 1905
196-202
frontispiece
xiii
Bibliography
Abbagnano, N. 1967a. Positivism, contained in: The Encyclopedia of Philosophy 6:

414-419.
Acton, H.B. 1967a. Bernard Bosanquet (1848-1923), contained in: The Encyclopedia
of Philosophy 1: 347-350.
_ _ . 1967b. Francis Bradley (1846-1924), contained in: The Encyclopedia of Philosophy 1: 359-363.
Adamson, R. 1911a. A shon history of logic. Edinburg & London: William Blackwood & Sons.
Agassi, J. 1963a. Towards a historiography of science. Monton: La Haya.
American Heritage Dictionan, The. 1982a. Boston: Houghton Mifflin Co. (Second
College Edition, 1985).
Anellis. 1. 1984a. Russell's earliest reactions to Cantorian set theory. Contemporary
Mathematics 31: 1-11
Anonymous. 1877a. Douglas A. Spalding. Nature 11: 35-36
_ _ . 1878a. News: Douglas A. Spalding. Mind 3: 153-154.
_ _ . 1897a. Review of Russell 1897a. The Times (June 24) 13.
_ _ . 1898a. Review of Russell 1897a. The Athenaeum 25-26.
_ _ . 1900a. Note on Russell 1900a. Westminster Review (December) no page
number, copy at the BRA.
_ _ . 1901a. Note on Russell 1900a. Revue de Metaphysique et de Morale 9 (supl.)
9.
_ _ . 1901b. Note on Russell 1900a. The Saturday Review (January 12) 53.
_ _ . 1901c. Note on Russell 1900a. The Guardian (April 3) no page number, copy
at the BRA.
_ _ . 1901d. Note on Russell 1900a. Weekly Register (April 26) no page number,
copy at the BRA.
212
Bibliography
_ _ . 1901e. Note on Russell 1900a. The Athenaeum (July 21) 114-115.

_ _ . 1901f. Review of Russell 1897a. Bulletin des Sciences Mathematiques 25:
79-80.
_ _ . 1902a. The mind of Leibniz (review of Russell 1900a). Literary Guide (July
1) no page number, copy at the BRA.
_ _ . 1902b. Note on Russell19OOa. American Encyclopedia o/Theology. (January)
no page number, copy at the BRA.
_ _ . 1903a. Review of Russell 1903a. Spectator 91: 491-492.
_ _ . 1903b. Review of Russell 1903a. TImes Literary Supplement 2: 263. [perhaps
written by G. H. Hardy, according to a letter written from Jourdain to Hardy
dated October 23, 1903].
_ _ . 1904a. Review of Russell 1903a. The Athenaeum 214.
_ _ . 1904b. Les principes de la theorie des ensembles. Revue Generale des Sciences Pures et Appliquees 17: 209.
_ _ . 1938a. Review of Russell 1938b. Nature 142: 662-663.
_ _ . 1938b. Review of Russell 1938b. TImes Literary Supplement 37: 14.
_ _ . 1939a. Review of Russell 1937a. Revue de Philosophie 8: 68-69.
_ _ . 1939b. Review of Russell 1938b. Revue de Philosophie 8: 77-78.
Asquith, P. D. & Kyburg, H. E. 1979a. (Editors). Current research in philosophy
science. Michigan: Edwards Brothers.
0/
Autonne, L. 1901a. Review of Russell's 1897a. Revue Generale des Sciences Pures
et App/iquees 12: 850-851.
Ayer, A. J. 1971a. Russell and Moore: the analytical heritage. Great Britain: Robert
Maclehose and Co.
_ _ . 1972a. Russell. New York: The Viking Press.
_ _ . 1982a. Philosophy in the Twentieth Century. London: Weidenfeld and Nicolson.
Ball, W. W. R. I 880a. The origin and history
bridge: E. Johnson.
0/ the mathematical tripos.
Cam-
Bar-hillel, Y. 1958a. See: Fraenkel, Bar-Hillel & Levy 1958a.

Barker, S. F. 1964a. Philosophy o/mathematics. New Jersey: Englewood Cliffs.
_ _ . 1967a. Number, contained in: The Encyclopedia 0/ Philosophy 5: 526-530.
Bell, E. T. 1937a. Men o/mathematics. New York: Simon & Shuster.
_ _ . 1945a. The development o/mathematics. New York: McGraw-Hill. (2nd ed).
_ _ . 1951a. Mathematics, queen and servant o/science. New York: McGraw-Hill.
213
Benacerraf, P. & Putnam, H. 1964a. (Editors). Philosophy of mathematics: selected

readings. New Jersey: Prentice Hall.
Benardete, I. A. 1964a. Infinity: an essay in metaphysics. Oxford: Clarendon Press.
Benthem, I. F. A. K. von. 1978a. Four paradoxes. Journal of Philosophical Logic 7:
49-72.
Bernstein, F. 1905a. Uber die Reihe der transfiniten Ordnungszahlen. Mathematische
Annalen 60: 187-193.
_ _ . 1905b. Zum Kontinuum problem. Mathematische Annalen 60: 463-464.
_ _ . 1905c. Untersuchungen aus der Mengenlehre. Mathematische Annalen 61:
117-155. [Dissertation Gottingen, 1901].
_ _ . 1938a. The continuum problem. Proceedings of the National Academy of
Sciences 24: 101-104.
Beth, E. W. 1936a. Some remarks on Dr. Perelman's essay on logical antinomies.
Mind 45: 487-488.
_ _ . 1959a. The foundations of mathematics. Amsterdam: North Holland Publishing Co.
Bingley, G. A. 1941a. See Cantor 1941a.
Bishop, E. 1975a. The crisis in contemporary mathematics. Historia Mathematica 2:
507-517.
Black, M. 1933a. Nature of mathematics. London: Kegan Paul & Trench.
Blackwell, K. (no date) (*) A bibliography of Bertrand Russell's periodical writings
1895-1923. BRA. Ts.
_ _ . 1969a. The importance to philosophers of The Bertrand Russell Archive.
Dialogue 7: 608-615.
_ _ . 1971a. (*) The genesis of Russell's first theory of denoting. BRA. Ts.
_ _ . 1971b. (*) The text of Russell's The Principles of Mathematics. BRA. Ts.
_ _ . 1971c. (*) Revisions to Ch. X, The Contradiction of The Principles of Mathematics. BRA. Ts.
_ _ . 1972a. An essay on the foundations of geometry. Russell # 6: 3-4 .
. 1972b. Russell's American lecture courses. Russell # 6: 8-9.
_ _ . 1972c. How Russell wrote [including a letter from Russell responding to an
academic questionnaire]. Russell # 8: 13-15
_ _ . 1973a. Our knowledge of Our knowledge. Russell # 12: 11-13.
_ _ . 1975a. A non-existent revision of Introduction to Mathematical Philosophy.
214
Bibliography
Russell # 20: 16-17.

_ _ . 1984a. Part I of The Principles of Mathematics. Russell (N. S.) 4: 271-288.
Blackwell, K. & Eames, E. R. 1975a. Russell's unpublished book on the theory of
knowledge. Russell # 19: 3-14 & 18.
_ _ . 1984a. (Editors). See: Russell 1984a.
Blackwell, K. 1976a. See: Thomas & Blackwell 1976a.
Blackwell, K. et al. 1983a. (Editors). See: Russell 1983a.
Bochenski, I. M. 1961a. A history offormal logic. USA: University of Notre Dame.
Bocher, M. 1904a. The fundamental conceptions and methods of mathematics. Bulletin of the American Mathematical Society 11: 115-135.
Bochner, S. 1973a. Mathematics in cultural history, contained in: Wiener 1973a I,
177-185.
Boggio, T. 1924a. See: Burali-Forti & Boggio 1924a.
Bolzano, B. 1950a. Paradoxes of the infinite. London: Routhledge and Kegan Paul.
Translated by: D. A. Steele. [Paradoxien des Unendlichen. 1851].
Bonevac, D. 1984a. Mathematics and metalogic. The Monist 67: 56-71.
Borel, E. 1905a. Quelques remarks sur les principes de la thoorie des ensembles.
Mathematische Annalen 60: 194-195.
_ _ . 1907a. La logique et l'intuition en mathematiques. Revue de Metaphysique et
de Morale 15: 273-283.
_ _ . 1908a. Les paradoxes de la thoorie des ensembles. Annales Scientijiques de
l'Ecole Normale Supirieure 25: 443-448.
_ _ . 1912a. Cing lettres sur la thoorie des ensembles, contained in: Theorie des
Fonctions. Paris: Hermann. p 150-160. (and in subsequent editions). [This
correspondence was originally published in Bulletin de la Societe Mathematique de France 33 (1905) 261-273. There is now an English translation contained in: Moore 1982a, 311-320].
_ _ . 1947a. Les paradoxes et l'axiome des choice. Comptes Rendus Hebdomadaires des Seances de l'Academie des Sciences (paris) 224: 1537-1538.
Bosanquet, B. 1888a. Logic or the morphology of knowledge. London: Oxford University Press. 2 vols.
215
Bourbaki, N. 196Oa. Elements d'histoire des Math.ematiques. Paris: Hermann.

Bowne, G. D. 1966a. The philosophy of logic, 1880-1908. The Hague: Moulton &
Co.
Boyer, C. 1949a. The concepts of the calculus. New York: Hafner Publishing Company.
_ _ . 1968a. A history of mathematics. New York: Wiley & Sons.
Bradley, F. 1883a. The principles of logic. London: Oxford University Press. 2 vols.
(2nd ed, 1922).
_ _ . 1893a. Appearance and reality. London: Swan Sonnenschein & Co.
Braithwaite, R. B. 1931a. See: Ramsey 1931a.
Broad, C. D. 1948a. Alfred North Whitehead (1861-1947). Mind 57: 139-145.
_ _ . 1973a. Bertrand Russell as a philosopher. Bulletin of the London Mathematical Society 5: 328-341.
_ _ . 1978a. Kant. An introduction. London: Cambridge University Press.
Broadbent, T. A. 1975a. Bertrand Arthur William Russell (1872-1970), contained in:
Dictionary of Scientific Biography 12: 9-17.
Brouwer, L. E. J. 1907a. On the foundations of mathematics, (Dissertation Amsterdam), contained in: Brouwer 1975a I, 11-101. [Over de Grondslagen der
Wiskunde. Amsterdam: Maas & van Suchtelen].
_ _ . 1908a. The unreliability of the logical principles, contained in: Brouwer
1975a I, 107-111. [De onbetrouwbaarheid der logische principes. Tijdschrift
voor wijsbegeerte 2 (1908) 152-158].
_ _ . 1913a. Intuitionism and formalism. American Mathematical Society Bulletin
20: 81-96. Translated by Arnold Dresden. [Also contained in: Brouwer 1975a
I, 123-138].
_ _ . 1923a. On the significance of the principle of excluded middle in mathematics, especially in function theory, contained in: van Heijenoort 1967b, 334-341. Translated by Stefan Bauer-Mengelberg and Jean van Heijenoort. [fiber
die Bedeutung des Satzes vom ausgeschlossenen dritten in der Mathematik,
insbesondere in der Funktionentheorie. Journal jar die reine und angewandte
Mathematik 154: 1-7].
_ _ . 1927a. On the domains of definitions of functions, contained in: van Heijenoort 1967b, 447-463. Translated by Stefan Bauer-Mengelberg. [aber Definitiosbereiche von Funktionen. Mathematische Annalen 97: 60-75].
_ _ . 1928a. Intuitionism reflections on Formalism [section 1], contained in: van
216
Bibliography
Heijenoort 1967b, 490-492. Translated by Stefan Bauer-Mengelberg. [Intuitionistische Betrachtungen fiber den Formalismus. Koninklijke Akademie van
wetenschappen te Amsterdam, Proceedings of the section of sciences, 31:
374-379.].
_ _ . 1940a. Consciousness, philosophy and mathematics, contained in Benacerraf
& Putnam 1964a, 78-84. [Originally published in: 10th International Congress
of Philosophy (Amsterdam, 1940). Proceedings I, fascicule II. Amsterdam:
North Holland, p 1243-1249].
_ _ . 1952a. Historical background, principles and methods of intuitionism. South
African Journal of Science 49: 139-146. [Also contained in: Brouwer 1975a I,
508-515].
_ _ . 1975a. Collected Works.. Amsterdam: North Holland. 2 vols. Editor: A.
Heyting.
Brown, R. & Rollins, C. D. 1969a. (Editors). Contemporary philosophy in Australia.
London: George Allen & Unwin.
Brunschvicg, L. 1912a. Les etapes de la philosophie mathematique. Paris: Blanchard.
(1972a, reprinted).
Bunn, R. 1977a. Quantitative relations between infinite sets. Annals of Science 34:
177-191.
_ _ . 1980a. Developments in the foundations of mathematics, 1870-1910, contained in: Grattan-Guinness 1980b, 220-255.
Burali-forti, C. 1894a. Logica matematica. Milano: Hoepli. (1919, 2nd. ed.).
_ _ . 1897a. A question on transfinite numbers, contained in: van Heijenoort 1967b, 104-111. [Una questione sui numeri transfiniti. Rendiconti del circolo
matematico di Palermo 11: 154-164.]
_ _ . 1897b. On well-ordered classes, contained in: van Heijenoort 1967b, 111-112. [Sulle clasH ben-ordinate. Rendiconti del circolo matematico di Palermo
11: 260.]
Burali-forti, C. & Boggio, T. 1924a. Espaces courbes. Critique a la relativite. Turin.
Burau, W. & Scriba, C. J. 1978a. Hermann Gunther Grassman (1809-1877), contained in: Dictionary of Scientific Biography 15: 192-199.
Byrd, M. 1987a. Part II of The Principles of Mathematics. Russell N. S. 7: 60-70.
Cajori, F. 1915a. The history of Zeno's arguments on motion [part 1 of 9]. American
Mathematical Monthly 21: 1-5.
217
_ _ . 1915b. The history of Zeno's arguments on motion [part 2 of 9]. American

_ _ . 1915c. The history of Zeno's arguments on motion [part 3 of 9]. American
_ _ . 1915d. The history of Zeno's arguments on motion [part 4 of 9]. American
_ _ . 1915e. The history of Zeno's arguments on motion [part 5 of 9]. American
_ _ . 1915f. The history of Zeno's arguments on motion [part 6 of 9]. American
_ _ . 1915g. The history of Zeno's arguments on motion [part 7 of 9]. American
_ _ . 1915h. The history of Zeno's arguments on motion [part 8 of 9]. American
_ _ . 1915i. The history of Zeno's arguments on motion [part 9 of 9]. American
_ _ . 1919a. A history of mathematics. New York: Chelsea Publications Co. (3rd
ed, 1980).
Cantor, G. 1883a. Sur una propriete du systeme de tous les nombres algebraiques
reels. Acta Mathematica 2: 305-310 .
. 1883b. Une contribution a la theorie des ensembles. Acta Mathematica 2:
311-328.
_ _ . 1883c. Sur les series trigonometriques. Acta Mathematica 2: 329-335.
_ _ . 1883d. Extension d'un theoreme de la theorie des series trigonometriques.
Acta Mathematica 2: 336-348.
_ _ . 1883e. Sur les ensembles in finis et lineares de points I. Acta Mathematicai 2:
349-356.
_ _ . 1883f. Sur les ensembles infinis et lineares de points II. Acta Mathematica 2:
357-360.
_ _ . 1883g. Sur les ensembles infinis et lineares de points III. Acta Mathematica
2: 361-371.
_ _ . 1883h. Sur les ensembles infinis et lineares de points IV. Acta Mathematica
2: 372-380.
_ _ . 1883i. Fondaments d'une theorie generale des ensembles. Acta Mathematica
2: 381-408.
_ _ . 1883j. Sur diverses theoremes de la theorie des ensembles de points situes
dans un espace continu a N dimensions (part I). Acta Mathematica 2: 409-414 .
. 1895a. Sui numeri transfiniti. Rivista di Matematica 5: 104-109.
_ _ . 1899a. Sur lesfodements de la tMorie des ensembles transfinis. Paris: Librairie Scientifique A. Hermann. Translated by F. Marotte.
218
Bibliography
_ _ . 1899b. Letter to Richard Dedekind, 28 July 1899. Printed in van Heijenoort

1967b, 114-117; and in Cantor 1932a, 443-447.
_ _ . 1915a. Contributions to the founding of the theory of transfinite numbers.
Chicago: Open Court. Edited and translated by P. Jourdain.
_ _ . 1932a. Gesammelte Abhandlungen mathematischen und philosophischen [nhaIts. Berlin: J. Springer. Editor: E. Zermelo.
_ _ . 1941a. (*) Transfinite numbers. Annapolis. Edited and translated by G. A.
Bingley.
_ _ . 1976a. Foundations of a general theory of manifolds. The Campaigner 9:
89-96.
Carnap, R. 1931a. The logicist foundations of mathematics. Erkenntnis 2: 91-121.
[Also contained in: Benacerraf & Putnam 1964a, 31-41 and reprinted in Klemke 1970a, 341-354].
Carus, P. 1908a. The foundations of mathematics: a contribution to philosophy of
geometry. Chicago: The Open Court Publishing Co.
_ _ . 1910a. The nature of logical and mathematical thought. The Monist 20: 3375.
Cassirer, E. 1902a. LeibniZ' system in seinen wissenschqftlichen Grundlagen. Marburg: N. G. Elwetsche Vedagsbuchhandlung.
_ _ . 1950a. The problem of knowledge, philosophy, science and history since
Hegel. New Haven: Yale University Press. Translated by W. H. Woglom &
C. W. Hendel.
Cavailles, J. 1962a. Philosophie Mathematique. Paris: Hermann.
Chihara, Ch. 1984a. The semantic paradoxes: some second thoughts. Philosophical
Studies 45: 223-229.
Church, A. 1934a. The Richard paradox. American Mathematical Monthly 41: 356-361.
_ _ . 1936a. A bibliography of symbolic logic. Journal of Symbolic Logic 1: 121-218.
_ _ . 1938a. Additions and corrections to Church 1936a. Journal of Symbolic Logic
3: 178-212.
_ _ . 1942a. Logical paradoxes, contained in: Runes 1942a, 224-225.
Clark, R. 1975a. The life of Benrand Russell. London: Weidenfeld & Nicolson .
. 1981a. Benrand Russell and his world. London: Thames & Hudson.
219
Clifford, W. K. 1885a. The common sense of the exact sciences. New York: A. A.
Knopf. [Reprinted 1946].
Coffa, A. 1979a. The humble origins of Russell's paradox. Russell # 33-34: 31-38.
_ _ . 1980a. Russell as a platonic dialogue: the matter of denoting. Synthese 45:
43-70.
Copi, I. 1958a. The Burali-Forti paradox. Philosophy of Science 25: 281-286.
_ _ . 1971a. The theory of logical types. London: Routhledge & Kegan Paul.
_ _ . 1978a. Introduction to logic. New York: Macmillan Publishing Co. (5th ed).
Copilowish, 1. 1948a. The logical paradoxes from 1897 to 1904. Ph.D. dissertation.
The University of Michigan.
Copleston, F. 1966a. A history of philosophy: Bentham to Russell. USA: Burnes &
Oates LTD. (Vol VIII).
Couturat, L. 1896a. De ['injini mathematique. Paris: A Blanchard. (Reprinted 1973).
_ _ . 1898a. Essais sur les fondements de la geometrie par Bertrand Russell. Revue
de Metaphysique et de Morale 6: 354-380.
_ _ . 1899a. La logique mathematique de M. Peano. Revue de Metaphysique et de
Morale 7: 616-646.
_ _ . 1899b. Review of Russell 1897a. Bulletin des Sciences Matematiques 23:
54-62.
_ _ . 1900a. L' Algebre universelle de M. Whitehead. [Review of Whitehead 1898a.] Revue de Metaphysique et de Morale 8: 323-362.
_ _ . 1901a. La logique de Leibniz d'apres des documents inedits. Paris: Alcan.
_ _ . 1903a. Opuscules et fragments inedits de Leibniz extraits de la Bibliotheque
Royale de Hanovre. Paris: Alcan.
_ _ . 1904a. Review of Russell 1903a. Bulletin des Sciences Mathematiques 28:
129-147.
_ _ . 1904b. La philosophie des mathematiques de Kant. Revue de Meraphysique et
de Morale 12: 321-383.
_ _ . 1904c. Logique et philo sophie des sciences. Revue de Metaphysique et de
Morale 12: 1037-1077.
_ _ . 1904d. Les principes des mathematiques [Part 1 of 7]. Revue de Meraphysique
et de Morale 12: 19-50.
_ _ . 1904e. Les principes des mathematiques [parts 2 & 3 of 7]. Revue de Metaphysique et de Morale 12: 211-240.
_ _ . 1904f. Les principes des mathematiques [parts 4 & 5 of 7]. Revue de Metaphysique et de Morale 12: 664-698.
_ _ . 1904g. Les principes des mathematiques [part 6 of 7]. Revue de Metaphysique
220
Bibliography

_ _ . 1905a. L 'algebra de la logique. Paris. [Reprinted by Hildesheim: Georg Olms
Verlagsbuchhandlung. 1965].
_ _ . 1905b. Les principes des mathematiques [part 7 of 7]. Revue de Metaphysique
_ _ . 1905c. Les principes des mathematiques, avec un appendice sur la philosophie
des mathematiques de Kant. Paris: Felix Alcan.
_ _ . 1906a. Pour la logistique. Revue de Meraphysique et de Morale 14: 208-250.
_ _ . 1906b. La logique et la philosophie contemporaine. Revue de Metaphysique et
de Morale 14: 318-341.
_ _ . 1906c. Logique et moralisme. Revue de Metaphysique et de Morale 14: 873-876.
_ _ . 1912a. For logistics. The Monist 22: 484-523. [Abridged English translation
of Couturat 1906a. See Jourdain 1912b] .
. 1960a. Lafilosofla de las matemaricas en Kant. Mexico: U.N.A.M. [Spanish
translation by M. Bueno of Couturat 1904b. Reprinted in Mathesis 1 (1985)
79-131].
Cranston, M. 1974a. John Stuart Mill (1806-1873), contained in: Dictionary of Scientiflc Biography 9: 383-386.
Crossley, J. N. 1973a. A note on Cantor's theorem and Russell's paradox. Australasian Journal of Philosophy 51: 70-71.
Curry, H. B. 1951a. Outlines of formalist philosophy of mathematics. Amsterdam:
North Holland .
. 1963a. Foundations of mathematical logic. New York: McGraw-Hill.
Dambska, I. 1974a. An essay on the foundations of geometry of B. Russell. Organon
10: 245-253.
Dan, P. 1986a. Russell's first theory of denoting and quantification. Notre Dame
Journal of Formal Logic 27: 133-166.
Dauben, J. W. 1971a. The trigonometric background of Georg Cantor's theory of
sets. Archives for the History of Exact Sciences 7: 181-216.
_ _ . 1974a. Denumerability and dimension: the origins of Georg Cantor's theory
of sets. Rete 2: 105-134.
_ _ . 1975a. The invariance of dimension: problems in the early development of set
theory and topology. Historia Mathematica 2: 273-288.
_ _ . 1977a. Georg Cantor and Pope Leo XIII: mathematics, theology and the
infinite. Journal of the History of Ideas 38: 85-108.
221
_ _ . 1977b. C. S. Peirce's philosophy of infinite sets. Mathematics Magazine 50:

123-135.
_ _ . 1978a. Georg Cantor: the personal matrix of his mathematics. Isis 69: 534-550.
_ _ . 1979a. Georg Cantor: his mathematics and philosophy of the infinite. Camb,
Mass: Harvard University Press.
_ _ . 1980a. The development of Cantorian set theory, contained in: Grattan-Guinness 1980b, 181-219.
_ _ . 1983a. Georg Cantor and the origins of transfinite set theory. Scientific American 248 (June 1983, # 6) 122-131.
_ _ . 1984a. Conceptual revolutions and the history of mathematics, contained in:
Mendelsohn 1984a, 81-103.
_ _ . 1988a. Review of Purkert & Ilgauds 1987a. Isis 79 (1988) 700-702.
Davis, H. T. 1939a. Review of Russell 1903a. Isis 30: 298-302.
Dedekind, R. 1888a. Essays on the theory of numbers. Chicago: Open Court Publishing Co. 1901. Translated by W. W. Beman. [Was sind und was sollen die
Zahlen? Braunschweig: Vieweg].
Demaio, F. P. 1976a. A unique endeavour. Russell # 21-22: 56-60.
Denon, L.E. 1961a. (Editor). See: Russell 1961a.
Dictionary of scientific biography, The. See: Gillispie 1973a.

Dixon, A. C. 1906a. On well-ordered aggregates. Proceedings of the London Mathematical Society 4: 18-20.
_ _ . 1907a. A question in the theory of aggregates. Proceedings of the London
Mathematical Society 4: 317-319.
Dorward, A. 1951a. Bertrand Russell. London: Benham & Co.
Dou, A. 1968a. E1 infinito en matematicas. Madrid .
. 1970a. Fundamentos de las matematicas. Barcelona: Labor.
Dresden, A. 1928a. Some philosophical aspects of mathematics. Bulletin of the American Mathematical Society 34: 438-452.
Du Bois-Reymond, P. 1887a. 'l11iorie generale des fonctions. [Vol I]. Nice: Imprimerie Nicoise. Translated by G. Milhaut & A. Gisot. [Die Allgeneine Functionentheorie. Dormstadt: Wissenchafliche].
222
Bibliography
Dummett, M. 1967a. Gottlob Frege (1848-1925), contained in: The Encyclopedia

Philosophy 3: 225-237.
0/
Duncan, G. M. 1901a. Review of Russell 1900a. Philosophical Review 10: 288-297.

Eames, E. R. 1969a. Benrand Russell's theory 0/ knowledge. London: George Allen
& Unwin.
_ _ . 1975a. See Blackwell & Eames 1975a.
_ _ . 1975b. Philip E. B. Jourdain and the Open Court papers. [CarbS 2: 101-112.
Eames, E. R. & Blackwell, K. 1984a. (Editors). See: Russell 1984a.
Edwards, P. 1967a. (Editor). The Encyclopedia
Millan. (8 vols).
0/ Philosophy.
New York: Mac-
Egner, R. E. & Dennon, L. E. 1961a. (Editors). See: Russell 1961a.

Encyclopedia
0/ philosophy,
The. See: Edwards 1967a.
Enriques, F. 1929a. The historic development o/logic. New York: Henry Holt & Co.
Translated by J. Rosenthal.
Euclid. 1956a. The thineen books 0/ Euclid's geometry. New York: Dover. 3 vols.
Editor: Sir Thomas L. Heath.
Eves, H. 1983a. An introduction to the history
Rinehart & Winston. (5th ed).
0/ mathematics.
New York: Holt,
Eves, H. & Newson, C. V. 1965a. An introduction to the/oundations andfundamental concepts o/mathematics. New York: Holt, Rinehart & Winston.
Farley, C. & Hodgson, D. 1972a. The life 0/ Benrand Russell in pictures and in his
own words. London: Russell Press Limited.
Feferman, S. 1982a. See: Kleene & Feferman 1982a.
Feigl, H. & Maxwell, G. 1962a. (Editors). Scientific explanation, space and time.
Minneapolis: University of Minneapolis Press.
Feinberg, B. 1967a. (Editor). A detailed catalog
London: Goodwin Press .
. 1969a. See: Russell 1969b.
0/ the archives 0/ Benrand Russell.
223
Feyerabend, P. R. 1962a. Explanation, reduction and empiricism, contained in: Feigl

& Maxwell 1962a, 28-97.
Flexner, J. Th. 1984a. An American saga: the story of Helen Flexner & Simon Flexnero Boston: Little Brown & Co.
Forder, H. G. 1938a. Review of Russell 1938b. Mathematical Gazette 22: 300-301.
Fraenkel, A. A. 1922a. The notion 'definite' and the independence of the axiom of
choice, contained in: van Heijenoort 1967b, 284-289. Translated by Beverly
Woodword. [Der Begriff "definit" und die Unabhlngigkeit des Auswahlsaxioms. Sitzungsberichte der Preussinschen Akademie der Wissenschqften,
Physikalisch-Mathematische Klasse, 253-257].
_ _ . 1933a. On modern problems in the foundations of mathematics. Scripta Mathematica 1: 222-227.

_ _ . 1947a. The recent controversies about the foundations of mathematics. Scripta
Mathematica 13: 17-36.
_ _ . 1951a. On the crisis of the principle of the excluded middle. Scripta Mathematica 17: 5-16.
_ _ . 1953a. Abstract set theory. Amsterdam: North Holland.
_ _ . 1966a. Set theory and logic. New York: Addison Wesley.
_ _ . 1967a. Georg Cantor (1845-1918), contained in: The Encyclopedia of Philosophy 2: 20-22.
Fraenkel, A. A., Bar-Hillel, Y. & Levy, A. 1958a. Foundations of set theory. Amsterdam: North Holland. (2nd ed, 1973).
Frege, G. 1879a. BegrijJsschriJt, aformula language, modeled upon that of arithmetic, for pure thought, contained in van Heijenoort 1967b, 1-82. Translated by
S. Bauer-Mengenlberg. [BegrijJsschriJt, eine der arithmetischen ... Halle:
Formelespracha des Reihen Denkens].
_ _ . 1884a. Los fundamentos de la aritmetica: una investigaci6n 16gico-matematica
sobre el concepto de numero, contained in: Frege 1972a, 105-126. [Die Grundlagen der Arithmetik, eine logisch-mathematische ... Breslau].
_ _ . 1893a. The basic laws of arithmetic. California: University of California
Press. 1964. Translated by M. Furth. [Grundgesetze der Arithmetik, .... Vol
I. Jena: Verlag Hermann Pohle.]
_ _ . 1902a. Letter to Bertrand Russell, June 22, 1902. Printed in: van Heijenoort
1967b, 127-128; and in: Frege 1980a, 131-132.
_ _ . 1903a. Grundgesetze der Arithmetik, .... Vol II. Jena: Pohle.
_ _ . 1952a. Translationsfrom the philosophical writings ofGottlob Frege. Oxford:
Basil Blackwell. Translated by M. Black and P. T. Geach.
224
Bibliography
_ _ . 1972a. Conceptografla. Los fundamentos de la aritmetica. Otros estudios

filos6ficos. Mexico: U.N.A.M. Edited and translated by H. Padilla.
_ _ . 1979a. Posthumous writings. Chicago: Chicago University Press. Translated
by P. Long & R. White.
_ _ . 1980a. Philosophical and mathematical correspondence. Chicago: Chicago
University Press. Translated by H. Kaal.
Freguglia, P. 1981a. La logica matematica di Peano: un'analisi. Physis 23: 325-336.
Freudenthal, H. 1975a. The cradle of modem topology, according to Brouwer's
inedita. Historia Mathematica 2: 495-502.
Gandy, R. O. 1973a. Bertrand Russell as mathematician. Bulletin of the London
Mathematical Society 5: 342-348.
Garciadiego, A. 1977a. La teom de conjuntos: sus origenes, desarrollo y consecuencias. Undergraduate dissertation. Universidad Nacional Aut6noma de Mexico.
_ _ . 1981a. See Moore & Garciadiego 1981a.
_ _ . 1983a. Bertrand Russell and the origin of the set theoretic paradoxes. Ph. D.
dissertation. University of Toronto. [Dissertations Abstracts International 75
(1984) 2864-A].
_ _ . 1984a. Bertrand Russell y los fundamentos de la fisica. Revista Mexicana de
Ffsica 30: 567-584.
_ _ . 1985a. The emergence of the non-logical paradoxes of the theory of sets,
1903-1908. Historia Mathematica 12: 337-35l.
_ _ . 1986a. On rewriting the history of the foundations of mathematics at the tum
of the century. Historia Mathematica 13: 39-4l.
_ _ . 1986b. L'influence de Georg Cantor sur Bertrand Russell. Cahiers du seminaire d'histoire des mathematiques de Toulouse # 8: 1-14.
_ _ . 1987a. Cantor y las paradojas. Aportaciones Matemdticas (Comunicaciones)
4: 139-148.
_ _ . 1987b. La precision du langage de Bertrand Russell. Cahiers du seminaire
d'histoire des mathematiques de Toulouse. # 10: 1-12.
_ _ . 1988a. Bertrand Russell y el origen de las paradojas de la teoria de conjuntos.
Mathesis 4: 113-130.
_ _ . 199Oa. Un pionero de la etologia: Douglas A. Spalding. Ciencia y Desarrollo
# 90. Enero-Febrero 1990. p 49-57.
_ _ . 1991a. Una introspectiva: cuestionando la influencia de las paradojas. Mathesis 7 (1991) 507-529.
_ _ . 1993a. The set-theoretic paradoxes, contained in Grattan-Guinness 1993a, (to
appear).
225
Gascoigne, J. 1984a. Mathematics and meritocracy: the emergence of the Cambridge

Mathematical Tripos. Social Studies of Science 14: 547-584.
Geach, P. T. 1970a. Two paradoxes of Russell's. Journal of Philosophy 67: 89-97.
GOdel, K. 1944a. Russell's mathematical logic, contained in: Schilpp 1944a, 125-153. [Also contained in: Benacerraf & Putnam 1964a, 211-232].
_ _ . 1947a. What is Cantor's continuum problem? American Mathematical Monthly 54: 515-525. [Also contained in: Benacerraf & Putnam 1964a, 258-273].
Goldstein, L. 1983a. Wittgenstein and the logico-semantical paradoxes. Ratio 25:
137-153.
Gomperz, T. 1913a. Greek thinkers. A history of ancient philosophy. London: J.
Murray. 4 vols. (Vol II-IV translated by G.G. Berry) [Griechische Denker.
1896.].
Goodman, N. 1984a. The knowing mathematician. Synthese 60: 21-38.
Goodstein, R. L. 1962a. The foundations of mathematics. Aristotelian Society Supplement 36: 145-154.
_ _ . 1965a. (Editor). Essays in the foundations of mathematics. USA: Leicester
University Press.
Gottschalk, H. 1967a. Benrand Russell, a life. London: Unwin Books. Translated by
E. Fitzgerald.
Grattan-Guinness, I. 1971a. The correspondence between Georg Cantor and Philip
Jourdain. Jahresbericht der Deustchen Mathematiker-Vereinigung 73: 111-130.
_ _ . 1971b. Towards a biography of Georg Cantor. Annals of Science 27: 345-391.
_ _ . 1972a. Bertrand Russell on his paradox and the multiplicative axiom. An
unpublished letter to Philip Jourdain. Journal of Philosophical Logic 1: 103-110.
_ _ . 1972b. University mathematics at the turn of the century. Unpublished recollections of W. H. Young. Annals of Science 28: 369-384.
_ _ . 1972c. Russell and Philip Jourdain, a study of their relationship. Russell # 8:
7-12.
_ _ . 1973a. Russell and Jourdain an exchange. Russell # 9: 20-21.
_ _ . 1974a. Jourdain's copy of Whitehead and Russell's Principia Mathematica.
Historia Mathematica 1: 187.
226
Bibliography
_ _ 1974b. The rediscovery of the Cantor-Dedekind correspondence. Jahresbericht der Deustchen Mathematiker-Vereinigung 76: 104-139.
_ _ . 1974c. The Russell Archives: some new light on Russell's logicism. Annals of
Science 31: 387-406.
_ _ . 1974d. Achilles is still running. Transactions of the Charles S. Peirce Society
10: 5-16.
_ _ . 1975a. Preliminary notes on the historical significance of quantification of the
axioms of choice in the development of mathematical analysis. Historia Mathematica 2: 475-488.
_ _ . 1975b. Russell's logical progress: some new light from manuscript sources.
Historia Mathematica 2: 489-493.
_ _ . 1975c. The Royal Society's financial support of the publication of Whitehead
and Russell's Principia Mathematica. Notes and Records of the Royal Society
of London 30: 89-104.
_ _ . 1975d. Wiener on the logics of Russell and SchrOder. An account of his
Doctoral thesis, and of his discussion of it with Russell. Annals of Science 32:
103-132.
_ _ . 1976a. On the mathematical and philosophical background to Russell's The
Principles of Mathematics, contained in: Thomas & Blackwell 1976a, 157-173.
1977a. Dear Russell - Dear Jourdain. London: Duckworth.
_ _ . 1978a. How Bertrand Russell discovered his paradox. Historia Mathematica
5: 127-137.
_ _ . 1980a. Georg Cantor's influence on Bertrand Russell. History and Philosophy
of Logic 1: 61-93.
_ _ . 1980b. (Editor). From calculus to set theory, 1630-1910. An introductory
history. London: Duckworth.
_ _ . 1981a. Are there paradoxes of the set of all sets? International Journal of
Mathematical Education, Science &: Technology 12: 9-18.
_ _ . 1982a. Psychology in the foundations of Logic and Mathematics: the cases of
Boole, Cantor and Brouwer. History and Philosophy of Logic 3: 33-53.
_ _ . 1984a. Notes on the Fate of Logicism from Principia Mathematica to GOdel's
Incompletability Theorem. History and Philosophy of Logic 5: 67-78.
1985a. Bertrand Russell's logical manuscripts: an apprenhesive brief. History
- - and Philosophy of Logic 6: 53-74.
1986a. Russell's logicism versus Oxbridge logics, 1890-1925: a contribution
- - to the real history of tweentieth-century English philosophy. Russell N. S. 5:
101-131.
_ _ . 1986b. Review of Lowe 1985a. Transactions of the Charles S. Pierce Society
22: 61-68.
_ _ . 1986c. What do theories talk about? A critique o( Popperian fallibism with
especial reference to ontology. Fundamenta Scientiae 7: 177-221.
227
_ _ . 1990a. Bertrand Russell (1872-1970), by acquaintance and by description.

Notes and Records of the Royal Society of London 44: 280-306.
_ _ . 1990b. Peirce between logic and mathematics, contained in: Roberts et al.
_ _ . 1991a. Boole y su semi-seguidor Jevons. Mathesis 7: 351-362.
_ _ . 1992a. Peirce entre la 16gica y las Matematicas. Mathesis 8: 55-72.
_ _ . 1993a. (Editor). The Encyclopedia of the History and Philosophy of the Mathematical Sciences. London: Routledge (to appear).
Gray, J. 1991a. Did Poincare say "Set theory is a Disease'? The Mathematicallntelligencer 13: 19-22.
Gray, P. H. 1962a. Douglas Spalding: the first experimental behaviorist. ]oumal of
General Psychology 67: 299-307.
_ _ . 1967a. Spalding and his influence on research in developmental behavior.
] oumal of the History of Behavioral Sciences 3: 168-179.
Gregory, J. C. 1920a. The nature of number [review of Russell 1919a]. Science
Progress 14: 141-146.
Grelling, K. 1936a. The logical paradoxes. Mind 45: 481-486.
Grelling, K. & Nelson, L. 1908a. Bemerkungen zu den Paradoxieen von Russell und
Burali-Forti. Abhandlungen der Fries 'chen Schule 2: 300-324.
Greenstreet, W. J. 1901a. Review of Russell 1897a. Mathematical Gazette 2: 121.
Griffin, N. 1980a. Russell on the nature of logic (1903-1913). Synthese 45: 117-188.
_ _ . 1981a. A choice set of letters (review of Grattan-Guinness 1977a). Russell #
37-40: 75-86.
_ _ . 1981b. The acts of the Apostles (review of Levy 1979a). Russell N. S. 1:
71-82.
_ _ . 1985a. Wittgenstein's criticism of Russell's theory of jugdement. Russell N.
S. 5: 132-145.
_ _ . 1988a. The Tiergarten programme. Russell N. S. 8: 19-34.
_ _ . 1991a. Russell's idealistic apprenticeship. London: Oxford.
Griffin, N. & Lewis, A.C. 1990a. Bertrand Russell's mathematical education. Notes
and Records of the Royal Society of London 44: 51-71.
_ _ . 1990b. (Editors). See: Russell 1990a.
Gross, T. J. 1974a. The Russell - Poincare debate. Master of Arts dissertation. McMaster University.
228
Bibliography
Grover, D. 1982a. Berry's paradox. Analysis 43: 170-176.

Guthrie, E. R. 1915a. The paradoxes of Mr. Russell with a brief account of their
history. Lancaster: The New Era Printing Co. [ph. D. dissertation. University
of Pennsylvania, 1914].
Hadamard, J. 1897a. Sur certain applications possibles de la theorie des ensembles.
Verhandlungen des Ersten Intemationalen Mathematiker Kongresses in Zurich

vom 9. bis 11. August 1897 [Leipzig, 1898] P 201-202.
[Hadamard, J.] 1904a. Le troisieme congres international des mathematiciens. Revue

Generale des Sciences Pures et AppJiquees 15: 961-962.
_ _ . 1905a. La theorie des ensembles. Revue Generale des Sciences Pures et Appliquees 16: 241-242.
_ _ . 1905b. Les principes des mathematiques et Ie probleme des ensembles. Revue
Generale des Sciences Pures et AppJiquees 16: 541-543.
_ _ . 1906a. Les principes de la theorie des ensembles. Revue Generale des Sciences Pures et Appliquees 17: 209.
Hagstrom, H. G. 1914a. Note sur l'antinomie Burali-Forti. Arkiv for Matematik,
Astronomi och Fysik 10: 1-4.
Haldane, J. B. S. 1954a. Introducing Douglas Spalding. British Journal of Animal
Behavior 2: 1.
Halsted, G. B. 1897a. Review of Russell 1897a. Science (September) # 143: 487-491.
Hannequin, A. 1895a. Essai critique sur l'hypothese des atomes dans la science
contemporaine. Paris.
Hardy, G. H. 1904a. A theorem concerning the infinite cardinal numbers. Quarterly
Journal of Pure and Applied Mathematics 35: 87-94.
_ _ . 1906a. The continuum and the second number class. Proceedings of the London Mathematical Society 4: 10-17.
[Hardy, G. H.] 1903a. Review of Russell 1903a. Times Literary Supplement 2: 263.
Harrah, D. 1959a. The influence of logic and mathematics on Whitehead. Journal of
the History of Ideas 20: 420-430.
229
Hart, W. P. 1983a. Russell and Ramsey. Pacific Philosophical Quarterly 64: 193-210.
Hartcher, W. S. 1968a. Foundations of Mathematics. Philadelphia: W. B. Saunders
Company.
Hawskwell-Briggs, R. Ch. 1972a. T. J. Cobden-Sanderson (1840-1922), contained
in: Encyclopedia Britannica 5: 991.
Heath, P. L. 1967a. George Boole (1815-1864), contained in: The Encyclopedia of
Heath, P. L. et al. 1967a. History of Logic: the Boolean Period, contained in: The
Encyclopedia of Philosophy 4: 541-548.
Heath, T. 1956a. See Euclid 1956a.
Hegel, G. W. H. 1892a. The science of logic. London: Oxford University Press.
Impression of 1931, originally published in two volumes 1812-1816. Translated (from the 2nd ed, revised and augumented) by W. Wallace from The Ency-
clopedia of the Philosophical Sciences).
Heijenoort, J. van. See Van Heijenoort, J.

Henkin, L. 1952a. Are logic and mathematics identical? An old thesis of Russell's is
reexamined in the light of subsequent developments in mathematical logic.
Science 138: 788-794.
1971a. Mathematical foundations of mathematics. American Mathematical
Monthly 78: 463-487.
Heyting, A. 1931a. The intuitionist foundations of mathematics, contained in: Benacerraf & Putnam 1964a, 422-49. [Die intuitionistische Grundlegung der Mathematik. Erkenntnis 2 (1931) 91-121].
1971a. Intuitionism. An introduction. Amsterdam: North Holland.
1975a. See Brouwer 1975a.
_ _ . 1978a. History of the foundations of mathematics. Nievw Archief voor Wiskunde 24: 1-21.
Hilbert, D. 1899a. The foundations of geometry. Chicago: The Open Court Publishing Co. 1902. Translated by E. J. Towsend. [GrundZagen der Geometrie.
Leipzig].
_ _ . 1900a. Uber den Zahlbegriff. lahresbericht der Deustchen Mathematiker--
230
Bibliography
Vereinigung 8: 180-184.
_ _ . 1900b. Mathematical problems. American Mathematical Society Bulletin 8
(1902) 437-479. Translated by M. W. Newson. [Mathematische Probleme.
Archiv der Mathematik und Physik HI 1: 44-63 & 213-237; also contained in:
Hilbert 1935a, 290-323].
_ _ . 1904a. On the foundations of logic and arithmetic. The Monist 15: 338-352.
Translated by G. B. Halsted. Also contained in: van Heijenoort 1967b, 129-138 and translated by B. Woodward. [Uber die Grundlagen der Logik und der
Arithmetik. Verhandlungen des Dritten Internationalen Mathematiker-Kongresses in Heidelberg vom 8. bis 13. August 1904. [Teubner, Leipzig
(1905)].
_ _ . 1925a. On the infinite, contained in: van Heijenoort 1967b, 369-392. Translated by S. Bauer-Mengenlberg. [Uber das Unendliche. Mathematische Annalen 95: 161-190.
_ _ . 1927a. The foundations of mathematics, contained in: van Heijenoort 1967b,
464-479. Translated by S. Bauer Mengenlberg. [Die Grundlagen der Mathematik. Abhandlungen aus dem Mathematischen Seminar der Hamburgischen
Universitat 6: 65-85].
_ _ . 1935a. Gesammelte Abhandlungen. New York: Chelsea Publishing Co. Vol
HI.
Hobson, E. W. 1905a. On the general theory of transfinite numbers and order types.
Proceedings of the London Mathematical Society II 5: 170-188.
_ _ . 1906a. On the arithmetic continuum. Proceedings of the London Mathematical
Society II 5: 21-28.
_ _ . 1921a. The theory of junctions of a real variable and the theory of Fourier
series. London: Cambridge University Press. Vol I.
_ _ . 1926a. The theory of junctions of a real variable and the theory of Fourier
series. London: Cambridge University Press. Vol II.
Hodgson, D. 1972a. See: Farley & Hodgson 1972a.
Houser, N. 1991a. See: Roberts, Houser & van Ebra 1991a.
Hudson, R. W. H. T. 1903a. Review of Russell 1903a. Nature 68: 410-412.
Hylton, P. 1990a. Russell, idealism and the emergence of analytical philosophy.
London: Clarendon Press.
Ilgauds, H. J. 1987a. See: Purkert & Ilgauds 1987a.
231
Irvine, W. B. 1984a. Russell's construction of space from perspectives. Synthese 60:

333-347.
Ishigoro, H. 1972a. Leibniz's philosophy of logic and language. Ithaca: Cornell
University Press.
Iyanaga, S. & Kawada, Y. 1977a. (Editors). Encyclopedic Dictionary of Mathematics. Camb, Mass: The MIT Press. 2 vols.
Jager, R. 1972a. The development of Bertrand Russell's philosophy. London: George
Allen & Unwin.
Jorgensen, J. 1931a. A treatise offormallogic. New York: Russell & Russell Inc. 2
vols.
Jourdain, P. 1904a. On the transfinite cardinal numbers of number classes in general.
Philosophical Magazine VI 7: 294-303.
_ _ . 1904b. On the transfinite cardinal numbers of well-ordered aggregates. PhilosophicaJ Magazine VI 7: 61-75.
_ _ . 1905a. On a proof that every aggregate can be well-ordered. Mathematische
Annalen 60: 465-470 .
. 1906a. De infinito in mathematics. Rivista di Matematica 8: 121-136.
_ _ . 1906b. The development of the theory of transfinite numbers [part I of 4].
Archiv der Mathematik und Physik III 10: 254-281.
_ _ . 1907a. On the question of the existence of transfinite numbers. Proceedings of
the London Mathematical Society II 4: 266-283.
_ _ . 1907b. Essays on logic by the Hon. B*rtr*nd R*ss*ll. Granta 20: 158 .
. 1908a. Review of Russell 1907b. Mathematical Gazette 4: 306-307.
_ _ . 1909a. The development of the theory of transfinite numbers [part 2 of 4].
_ _ . 191Oa. A mathematical paradox. The Monist 20: 134-135.
_ _ . 191Ob. The development of theories of mathematical logic and the principles
of mathematics [part 1 of 3]. Quarterly Journal of Mathematics 41: 324-352.
_ _ . 191Oc. The development of the theory of transfinite numbers [part 3 of 4].
_ _ . 191Od. Transfinite numbers and the principles of mathematics [part 1 and
only]. The Monist 20: 93-118.
_ _ . 1911a. Some modem advances in logic. The Monist 21: 564-566.
_ _ . 1911b. The philosophy of Mr. B*rtr*nd R*ss*ll. The Monist 21: 481-508.
_ _ . 1912a. The development of the theories of mathematical logic and the principles of mathematics [part 2 of 3]. Quarterly Journal of Mathematics 43: 219-314.
232
Bibliography
_ _ . 1912b. Introductory note to English translation of Couturat 1906a. The Monist

22: 481-483.
_ _ . 1912c. Henri Poincare. The Monist 22: 611-615.
_ _ . 1912d. Mr. Bertrand Russell's first work on the principles of mathematics.
The Monist 22: 149-158.
_ _ . 1913a. A correction and some remarks. The Monist 23: 145-148.
_ _ . 1913b. The development of the theories of mathematical logic and the principles of mathematics [part 3 of 3]. Quanerly Journal of Mathematics 44: 113-128.
_ _ . 1913c. The development of the theory of transfinite numbers [part 4 of 4].
Archiv der Mathematik und Physik TIl 22: 1-21.
_ _ . 1913d. Tales with philosophical morals. The Open Coun 27: 310-315.
_ _ . 1915a. Edition, translation, introduction and notes to Cantor 1932a, 282-351.
_ _ . 1915b. Mathematicians and philosophers. The Monist 25: 633-638.
_ _ . 1916a. Richard Dedekind (1833 [sic] - 1916). The Monist 26: 415-427.
_ _ . 1916b. The logical work of Leibniz. The Monist 26: 504-523.
_ _ . 1916c. The philosophy of Mr. B*rtr*nd R*sS*ll. The Monist 26: 24-62.
_ _ . 1917a. Function and symbolism in mathematical logic. Scientia 21: 1-12.
_ _ . 1918a. The philosophy of Mr. B*nr*nd R*ss*ll. London & Chicago: George
Allen & Unwin.
_ _ . 1919a. Cause and effect. The Monist 29: 453-467.
_ _ . 1920a. Logic and mathematics. Science Progress 14: 669-674.
_ _ . 1920b. Review of Russell 1919a. Mathematical Gazette 10: 46.
_ _ . 1922a. A proof that every aggregate can be well-ordered. Acta Mathematica
43: 239-261.
Kant, I. 1787a. Critique ofpure reason. New York: St Martin's Press. (1965). Translated by N. K. Smith from the 2nd ed.
_ _ . 1793a. Critique of judgement. London: Macmillan and Co. (1914, 2nd ed).
Translated by I. H. Bernard.
Karrils, R. 1969a. See Russell 1969b.
Kattsoff, L. 1948a. A philosophy of mathematics. Iowa: The Iowa State College
Press.
Kaufmann, F. 1978a. The irifinite in mathematics. Netherlands: D. Reidel Publishing
Co.
Kawada, Y. 19TIa. See: Iyanaga & Kawada 1977a.
Keen, C. N. 1971a. The interaction of Russell and Bradley. Russell # 3: 7-11.
233
Kennedy, H. C. (no date) An appreciation of Giuseppe Peano. Pi Mu Epsilon Journal

3: 107-113.
_ _ . 1963a. The mathematical philosophy of Giuseppe Peano. Philosophy of Science 30: 262-266.
_ _ . 1970a. Cesare Burali-Forti (1871-1931), contained in: Dictionary of Scientific
Biography 2: 593-594.
_ _ . 1972a. The origins of modern axiomatics: Pash to Peano. American Mathematical Monthly 79: 133-136.
_ _ . 1973a. Edition, notes, bibliographical sketch and a chronological list of publications of Peano 1973a.
_ _ . 1973b. What Russell learnt from Peano. Notre Dame Journal of Formal Logic
14: 367-374.
_ _ . 1974a. A prospective biography of Giuseppe Peano. Historia Mathematica 1:
87-89.
_ _ . 1974b. Giuseppe Peano (1858-1932), contained in: Dictionary of Scientific
Biography 10: 441-444.
_ _ . 1974c. Peano's concept of number. Historia Mathematica 1: 387-408.
_ _ . 1974d. Giuseppe Peano. Berlin: Birkhauser.
_ _ . 1975a. Nine letters from Giuseppe Peano to Bertrand Russell. Journal History
of Philosophy 13: 205-220.
_ _ . 1980a. Peano: his life and work. New York: North Reidel.
Keyser, C. J. 1903a. Concerning the axiom of infinity and mathematical induction.
Bulletin of the American Mathematical Society 9: 424-434.
_ _ . 1904a. The axiom of infinity: a new presupposition of thought. Hibben Journal 2: 523-552.
Kilmister, C. W. 1984a. Russell. Great Britain: The Harvester Press.
King, J. (no date) (*) A report on the manuscripts for" An analysis of mathematical
reasoning", "The fundamental ideas and axioms of mathematics", and "The
principles of mathematics." BRA. Ts.
Kleene, S. C. 1952a. Introduction to metamathematics. Amsterdam: North Holland.
Kleene, S. C. & Feferman, S. 1982a. Foundations of mathematics, contained in:
Encyclopedia Britannica (15th ed). [Macropedia] 11: 630-639.
Klein, F. 1896a. The arithmetization of arithmetic. American Mathematical Society
Bulletin 2: 241-249. Translated by I. Maddison. [Arithmetisierung der Mathematik. (Lecture delivered on 2 November 1895). Also contained in: Felix
Klein. Gesammelte Mathematische Abhandlungen. Berlin: Springer. 1921-23.
234
Bibliography
Vol II. P 232-240].

_ _ . 1921-1923a. Gesammelte mathematische Abhandlungen. Berlin: Springer.
Klemke, E. D. 1968a. (Editor). Essays on Frege. Chicago: University of Illinois
Press.
_ _ . 1970a. (Editor). Essays on Benrand Russell. Chicago: University of Illinois
Press.
Kline, M. 1972a. Mathematical thought from ancient to modern times. New York:
Oxford University Press.
_ _ . 1980a. Mathematics: the loss of certainty. New York: Oxford University
Press.
Kneale, M. 1967a. See: Kneale & Kneale 1967a.
Kneale, W. 1979a. Russell's paradox and some others, contained in: Lewis 1979a,
34-51.
Kneale, W. & Kneale, M. 1967a. The development of logic. Great Britain: Oxford
University Press.
Knorr, W. 1975a. The evolution of the euclidean elements. Boston: D. Reidel.
Konig, J. 1905a. Zum Kontinnum Problem. Verhandlungen des Dritten lnternationalen Mathematiker-Kongresses in Heidelberg vom 8. bis 13. August 1904.
Leipzig: Teubner (1905) 144-147.
_ _ . 1905b. Zum Kontinuum Problem. Mathematische Annalen 60: 177-180.
_ _ . 1OO5e. On the foundations of set theory and the continuun problem, contained
in: van Heijenoort 1967b, 145-149. Translated by S. Bauer-Mengelberg.
[Uber die Grundlagen der Mengenlehre und das Kontinuum-Problem [part 1 of
2]. Mathematische Annalen 61: 156-160].
_ _ . 1906a. Sur la theorie des ensembles. Comptes Rendus Hebdomadaires des
Seances de l'Academie des Sciences (paris) 143: 110-112.
_ _ . 1OO7a. Uber die Grundlagen der Mengenlehre und das Kontinuum-Problem
[part 2 of 2]. Mathematische Annalen 63: 217-221.
_ _ . 1914a. Neue Grundlagen der Logik, Arithmetik und Mengeniehre. Leipzig:
Veit.
Korner, S. 1955a. Kant. London: Penguin Books Ltd.
_ _ . 196Oa. The philosophy of mathematics. London: Hutchinson University Library.
235
Kowalewski, G. 1950a. Bestand und Wandel. Munchen: R. Oldenbourg.

Kramer, E. E. 1970a. The nature and growth of modem mathematics. New York:
Hawthorn.
Krazer. 1903a. Bericht uber die lahresversammlung in Kassel. lahresbericht der
Deutschen Mathematiker-Vereinigung 12: 517-520.
Kreisel, G. 1958a. Hilbert's program. Dialectica 12: 346-372; also contained in:
Benacerraf & Putnam 1964a, 157-180.
_ _ . 1973a. Bertrand Russell, F. R. S. Biographical Memoirs of Fellows of the
Royal Society 19: 583-620.
_ _ . 1984a. Frege's foundations and intuitionist logic. The Monist 67: 72-91.
Kremer, M. 1985a. Frege's theory of number and the distinction between function
and object. Philosophical Studies 47: 313-323.
Kyburg, H. E. 1979a. See: Asquith & Kyburg 1979a.
Lackey, D. P. 1972a. The Whitehead correspondence. Russell # 5: 14-16 .
. 1973a. See: Russell 1973a.
Laisant, C. A. 1898a. Review of Russell 1897a. Revue Generale des Sciences Pures
et App/iquees 9: 35.
Langlois, Ch. & Seignobos, Ch. 1898a. Introduction to the study of history. New
York: Henry Holt. 1926. (The preface signed by F. York Powell is dated
1898). Translated by G. G. Berry. [Introduction aux etudes historiques. Paris].
Larguier, E. H. 1940a. Brouwerian philosophy of mathematics. Scripta Mathematica
7: 69-78.
Latta, R. 1901a. Critical notices (review of Russell 1900a). Mind 10: 525-533.
Legett, H. W. 1950a. Benrand Russell: a sketch of his life and work. New York:
Philosophical library.
Lehman, H. 1984a. Mathematical proofs: gaps and postulativism. The Monist 67:
108-113.
236
Bibliography
Levin, G. D. 1982a. Dialectics and the paradoxes of set theory. Soviet Studies in
Levy, A. 1958a. See Fraenkel, Bar-Hillel & Levy 1958a.
Levy, P. 1979a. G. E. Moore and the Cambridge Apostles. New York: Holt, Rinehart & Winston.
Lewis, A.C. 1990a. See: Griffin & Lewis 1990a.
. 1990b. See: Griffin & Lewis 1990b.
Lewis, C. I. 1918a. A survey of symbolic logic. Berkeley: University of California
Press.
Lewis, H. D. 1979a. (Editor). Bertrand Russell memorial volume. London: George
Allen & Unwin.
Lilley, S. 1949a. Social aspects of the history of science. Archives Intemationales
d'Histoire des Sciences 2: 376-443.
Lind, R. W. 1984a. Microphenomenology and numerical relations. The Monist 67:
29-45.
Lorenzen, P. 1960a. Constructive mathematics as a philosophical problem, contained
in: Logic and the foundations of mathematics. The Netherlands: Walters Noordhoff Pub!.
L[owe], A. E. H. 1897a. Review of Russell 1897a. Nature 56: 417-418.
Lowe, V. 1962a. Understanding Whitehead. Maryland: John Hopkins University
Press.
_ _ . 1974a. Whitehead's 1911 criticism of The Problems of Philosophy. Russell #
13: 3-10.
_ _ . 1975a. A. N. Whitehead on his mathematical goals: a letter of 1912. Annals
of Science 32: 85-101.
_ _ . 1985a. Alfred North Whitehead: the man and his work, 1861-1910. London:
The Johns Hopkins Univbersity Press. Vol I.
Lucas, B. J. 1971a. Moore's influence on Russell. Russell # 1: 9-10.
Luchins, A. 1961a. See: Luchins & Luchins 1961a.
237
Luchins, E. & Luchins, A. 1961a. Logicism. Scripta Mathematica 27: 223-243.

Maccoll, H. 1899a. Review of Whitehead 1898a. Mind 8: 108-113.
_ _ . 1906a. Symbolic logic and its applications. London.
Macgill, V. J. 1938a. Review of Russell 1938b. Nation 146: 733-734.
Mackenzie, N. 1978a. See: Webb & Webb 1978a.
Maddy, P. 1984a. Mathematical epistemology: what is the question? The Monist 67:
46-55.
Mathews, G. B. 1898a. Review of Whitehead 1898a. Nature 58: 385-386.
Maxwell, G. 1962a. See: Feigl & Maxwell 1962a.
May, K. O. 1970a. Russell: the rebel aristocrat. Videocassette. Toronto: Ontario
Department of Education. 30 mins. BIN. 3/4 inc. AUL Videocass 001408.
McClearly, J. 1989a. See: Rowe & McClearly 1989a.
McFarlane, A. 1899a. Review of Whitehead 1898a. Science 9: 324-328.
McMichael, A. 1982a. A set theory with Frege-Russell cardinal numbers. Philosophical Studies 42: 141-149.
Meder, A. E. 1940a. Review of Russell 1938b. Scripta Mathematica 7: 138-141.
Mehrtens, H. 1976a. T. S. Kunh's theories and mathematics: a discussion paper on
the 'new historiography' of mathematics. Historia Mathematica 3: 297-320.
Mendelsohn, E. 1984a. (Editor). Transformation and tradition in the sciences. Camb,
Mass.: Cambridge University Press.
Menzel, Ch. 1984a. Cantor and the Burali-Forti paradox. The Monist 67: 92-106.
Meschowski, H. 1964a. Ways of thought of great mathematicians. San Francisco:
Holden Day. Translated by J. D. Bennet.
_ _ . 1965a. Evolution of mathematical thought. San Francisco: Holden Day.
_ _ . 1971a. Georg Cantor (1845-1918), contained in: Dictionary of Scientific
Biography 3: 52-58.
238
Bibliography
Mill, J. S. 1848a. A system of logic. London: Longsmans, Green, Reader & Dyer.
(7th ed., 1868, 2 vols).
_ _ . 1924a. Autobiography. New York: Columbia University Press.
Mirimanoff, D. 1917a. Les antinomies de Russell et de Burali-Forti et Ie probleme
fondamental de la throrie des ensembles. L 'Enseignement Mathematique 19:
37-52.
_ _ . 1917b. Remarques sur la throrie des ensembles et les antinomies cantoriennes
[part 1 of 2]. L 'Enseignement Mathematique 19: 209-217.
_ _ . 1920a. Remarques sur 1a throrie des ensembles et les antinomies cantoriennes
[part 2 of 2]. L 'Enseignement Mathematique 21: 29-52.
Moore, A. W. 1984a. Possible words and diagonalization. Analysis 44: 21-22.
Moore, E. H. 1903a. On the foundations of mathematics. Bulletin of the American
Mathematical Society 9: 402-425. [Also published in: Science 17 (1903) 401-416].
Moore, G. E. 1899a. Review of Russell 1897a. Mind 8: 399-400.
_ _ . 1899b. The nature of judgement. Mind 8: 176-193.
_ _ . 1900a. Necessity. Mind 9: 288-304.
_ _ . 1905a. (*) Review of Russell 1903a. BRA. Ts.
_ _ . 1931a. Preface to Ramsey 1931a.
Moore, G. H. 1976a. Ernst Zermelo, A. E. Hadamard and the axiomatization of set
theory. Historia Mathematica 3: 206-208.
_ _ . 1978a. The origins of Zermelo's axiomatization of set theory. Journal of
Philosophical Logic 7: 307-329.
_ _ . 1980a. Review of Dauben 1979a. History and Philosophy of Logic 1: 238-240.
_ _ . 1981a. The Russell-Peano connection [review of Kennedy 1980a]. Russell #
37-40: 71-74.
_ _ . 1982a. Zennelo's axiom of choice. New York: Springer-Verlag.
_ _ . 1988a. The roots of Russell's paradox. Russell N. S. 8: 46-56.
_ _ . 1989a. Towards a history of Cantor's Continuum Problem, contained in:
Rowe & McClearly 1989a, 79-121.
Moore, G. H. & Garciadiego, A. 1981a. Burali-Forti's paradox: a reappraisal of its
origins. Historia Mathematica 8: 319-350.
Moritz, R. 1958a. On mathematics. New York: Dover.
239
Muirhead, J. H. 1935a. Bernard Bosanquet and his friends. Letters illustrating the
sources and development of his philosophical opinions. London: George Allen
& Unwin.
Murata, T. 1958a. French empirism 1. Commentari Mathematici Universitati Sancto
Pauli 6: 93-114.
Myhill, J. 1984a. Paradoxes. Synthese 60: 129-143.
Nelson, L. 1908a. See: Grelling & Nelson 1908a.
Newmann, J. See: Van Newmann, J.
Newson, C. V. 1965a. See: Eves & Newson 1965a.
Nidditich, P. H. 1962a. The development of mathematical logic. Illinois: The Free
Press.
_ _ . 1963a. Peano and the recognition of Frege. Mind 72: 103-110.
o 'Briant , W.
H. 1979a. Russell on Leibniz. Studia Leibnitiana 11: 159-222.
Oxford English Dictionary, The. 1989a. Oxford: Clarendon Press. (2nd.ed.). 20 vols.
Padilla, H. 1972a. See: Frege 1972a.
Passmore, J. 1967a. Logical positivism, contained in: The Encyclopedia of Philosophy 5: 52-57.
_ _ . 1968a. A hundread years of philosophy. New York: Penguin Books.
_ _ . 1969a. Bradley and Russell, contained in: Brown & Rollins 1969a, 21-30.
Peano, G. 1888a. The geometrical calculus according to the Ausdehnunggslehre of H.
Grassmann, preceded by the operations of deductive logic, contained in:
Peano 1973a, 75-100. [Excerpts from Calcolo Geometrico secondo l'Aus ....
Turin: Boccal].
_ _ . 1889a. The principles of arithmetic, presented by a new method; contained in:
Peano 1973a, 101-134; also contained in: van Heijenoort 1967b, 83-97. [Arithmetices Principia, nova methodo exposita. Turin].
_ _ . 1891a. The principles of mathematical logic, contained in: Peano 1973a,
153-161. [Princippi di logica matematica. Rivista di Matematica 1: 1-10].
_ _ . 1894a. Formulaire de mathematiques. Turin: Bocca. 4 vols (1894 - 1908).
_ _ . 1896a. The most general question in the mathematical sciences, contained in:
Peano 1973a, 189. [Reply to question # 719. L'lntermediaire des Mathemati-
240
Bibliography
ciens 3: 169].
_ _ . 1897a. Studies in mathematical logic, contained in: Peano 1973a, 190-205.
[Studii di logica matematica. Atti della Reale Accademia delle Scien:ze di
Torino 32: 565-583].
_ _ . 19OOa. Les definitions mathematiques. 1st International Congress of Philosophy. Paris: Colin. Vol III p 279-287.
_ _ . 1906a. Supplement to "On the Cantor-Bernstein Theorem", contained in:
Peano 1973a, 206-218. [Super theorema de Cantor-Bernstein et additione.
Rivista di Matematica 8: 136-157].
_ _ . 1913a. Review of Whitehead & Russell 1910a. Bolletino di Bibliografia e
storia delle scien:ze mathematiche 15: 47-53 & 75-81.
_ _ . 1915a. The importance of symbols in mathematics, contained in: Peano 1973a, 227-234. [Importanza del simboli in matematica. Scientia 18: 165-173].
_ _ . 1921a. Definitions in mathematics, contained in: Peano 1973a, 235-246. [Le
definizioni in matematica. Periodico di matematiche 1: 175-189].
_ _ . 1973a. Selected works of Giuseppe Peano. Toronto: University of Toronto
Press. Editor: H. C. Kennedy.
Pears, D. F. 1972a. Benrand Russell: a collection of critical essays. New York:
Doubleday & Co.
_ _ . 1975a. Review of Clark 1975a. New Review 2: 63-66.
Peirce, C.S. 1881a. On the logic of number. American Journal of Mathematics 4:
85-95.
Perelman, Ch. 1936a. Les paradoxes de la logique. Mind 45: 204-208.
Pitt, I. 1980a. Russell and Moore. Russell # 37-40: 9-16.
Poincare, H. 1894a. Sur la nature du raisonnement mathematique. Revue de Metaphysique et de Morale 2: 371-384.
_ _ . 1898a. On the foundations of geometry. The Monist 9: 1-43.
_ _ . 1899a. Des fondements de la geometrie a propos d'un livre de M. Russell.
Revue de Metaphysique et de Morale 7: 251-279.
_ _ . 1899b. La logique et la intuition dans la science mathematique et dans l'enseignement. L'Enseignement Mathematique 1: 157-162.
_ _ . 1902a. Review of Hilbert 1899a. Bulletin of the American Mathematical
Society 10 (1903) 1-23. Translated from French by E. V. Huntington. [Bulletin des Sciences Mathematiques 26: 249-272].
_ _ . 1905a. Les mathematiques et la logique [part 1 of 3]. Revue de Metaphysique
_ _ . 1905b. Science and hypothesis. London: The Walter Scott Publ.
241
_ _ . 1906a. Les mathematiques et la logique [part 2 of 3]. Revue de Meraphysique

_ _ . 1906b. Les mathematiques et la logique [part 3 of 3]. Revue de Metaphysique
_ _ . 1906c. A propos de la logistique. Revue de Meraphysique et de Morale 14:
866-868.
_ _ . 1906d. M. Poincare's Science et Hypothese [note to Russell]. Mind 15: 141-143.
_ _ . 1908a. L'avenir des mathematiques. Atti del IV Congresso Intemazionale dei
Matematici. Rome 6-11 April. Rome: Tipografia della R. Accademia dei Lincei, C. V. Salviucci. P 167-182.
_ _ . 191Oa. The future of mathematics. The Monist 20: 76-92. English translation
of Poincare 1908a by G. B. Halsted and contained in: Poincare 1913a, 369-382.
_ _ . 1912a. The new logics. The Monist 22: 243-256. Abridged English translation
of Poincare 1905a by G. B. Halsted and contained in: Poincare 1913a, 460-471.
_ _ . 1912b. The latest efforts of the logicitians. The Monist 22: 524-539. Abridged
English translation of Poincare 1906b by H. B. Halsted and contained in:
Poincare 1913a, 472-485.
_ _ . 1913a. The foundations of science. Lancaster: The Science Press. Translated
by G. B. Halsted.
_ _ . 1964a. On the nature of mathematical reasoning, contained in: Benacerraf &
Putnam 1964a, 382-389; excerpted and reprinted from Poincare 1905b, 1-19.
Pollock, J. L. 1970a. On logicism, contained in: Klemke 1970a, 388-396.
Posy, C. J. 1984a. Kant's mathematical realism. The Monist 67: 115-134.
Purkert, W. 1986a. Georg Cantor und die antinomien der Mengenlehre. Bulletin de
la Sociere MatMmatique de Belgique 38: 313-327.
_ _ . 1989a. Cantor's view on the foundations of mathematics, contained in: Rowe
& McClearly 1989a, 49-64.
Purkert, W. & Ilgauds, H. J. 1987a. Georg Cantor, 1845-1918. Basel: Birkhauser.
Putnam, H. 1964a. See: Benacerraf & Putnam 1964a.
_ _ . 1967a. The thesis that mathematics is logic, contained in: Schoen man 1967a,
273-303.
_ _ . 1979a. Philosophy of mathematics: a report, contained in: Asquith & Kyburg
1979a, 386-398.
242
Bibliography
Quine, W. 1941a. Russell's paradox and others. Technology Review 44: 16-17.
_ _ . 1941b. Whitehead and the rise of modern logic, contained in: Schilpp 1941a,
125-163.
_ _ . 1953a. From a logical point of view. Cambridge, Mass: Harvard University
Press.
_ _ . 1966a. The ways of paradox and other essays. New York: Random House.
Radner, M. 1972a. Index of names of the foundations of geometry. Russell # 6: 20.
_ _ . 1975a. Philosophical foundations of Russell's logicism. Dialogue 14: 241-253.
Ramsey, F. P. 1925a. Review of Whitehead & Russell 1910a. Mind 34: 506-507.
_ _ . 1925b. Review of Whitehead & Russell 1910a. Nature 116: 127-128.
_ _ . 1926a. The foundations of mathematics. Proceedings of the London Mathematical Society 25: 338-394 .
. 1927a. Achilles and the tortoise. The Forum 78: 430-434.
_ _ . 1931a. The foundations of mathematics and other logical essays. London:
Kegan Paul. Editor: R. B. Braithwaite.
Rang, B. 1981a. See: Thomas & Rang 1981a.
Ready, W. B. 1969a. Necessary Russell. Canada: Clark Publications Co.
Reese, W. L. 1980a. Dictionary of Philosophy and Religion. New Jersey: Humanities
Press.
Reid, C. 1970a. Hilben. Berlin: Springer-Verlag.
Rempel, R. 1979a. From imperialism to free trade: Couturat, Halevy, and Russell's
first crusade. Journal of the History of Ideas 40: 423-443.
Rempel, R., et al. 1979a. (*) The Bertrand Russell editorial project: an application
for a major editorial grant from the Social Sciences and Humanities Research
Council. Hamilton, Ontario: McMaster University .
. 1985a. See: Russell 1985a.
Resnik, M. D. 1980a. Frege and the philosophy of mathematics. New York: Cornell
University Press.
Richard, J. 1905a. The principles of mathematics and the problems of sets, contained
in: van Heijenoort 1967b, 142-144. Translated by J. van Heijenoort. [Les
principes des mathematiques et Ie probl~me des ensembles. Revue Generale
243
des Sciences Pures et AppUquees 16: 541].

_ _ . 1907a. Sur un paradoxe de la theorie des ensembles et sur l'axiome de Zermelo. L 'Enseignement Mathematique 9: 94-98.
Richards, J. L. 1986a. Projective geometry and mathematical progress in mid-Victorian Britain. Studies in History and Philosophy of Science 17: 297-325.
_ _ . 1988a. Bertrand Russell's Essay on the Foundations of Geometry and the
Cambridge mathematical tradition. Russell N. S. 8: 59-80.
_ _ . 1988b. Mathematical visions: the pursuit of geometry in Victorian England.
New York: Academic Press.
Roberts, D., Houser, N. & Van Ebra, J. 1991a. (Editors). Studies in the logic of
Charles S. Peirce. Indiana: Indiana University Press.
Robinson, A. 1968a. Some thoughts on the history of mathematics, contained in:
Logic and the foundations of mathematics. The Netherlands: Walters-Noordhoff Publ.
Rodriguez-Consuegra, F. 1987a. Bibliografia de Bertrand Russell en espanol (primaria y secundaria). Mathesis 3: 183-187.
_ _ . 1987b. Russell's logicist definitions of numbers 1898-1913. History and Philosophy of Logic 8: 141-169 .
. 1988a. Bertrand Russell 1898-1900: una filosofia de la matematica inedita.
Mathesis 4: 3-76
_ _ . 1988b. La obra logicista de Peano y su escuela. Mathesis 4: 221-299.
_ _ . 1988c. Bertrand Russell 1900-1913: Los Principios de la Matematica (Parte
1). Mathesis 4: 355-392.
_ _ . 1988d. Bertrand Russell 1900-1913: Los Principios de la Matematica (Parte
2). Mathesis 4: 489-521.
_ _ . 1989a. Russell's theory of types, 1901-1910: its complex origins in the unpublished manuscripts. History and Philosophy of Logic 10: 131-164.
_ _ . 1990a. Bertrand Russell 1895-1899: una filosofia prelogicista de la geometria.
Dfalogos 55: 71-123 .
. 1991a. The mathematical philosophy of Bertrand Russell: Origins and Development. Basel: Birkhauser.
Rollins, C. D. 1969a. See: Brown & Rollins 1969a.
Rosser, B. 1942a. The Burali-Forti paradox. Journal of Symbolic Logic 7: 1-17.
Rowe, D. & McClearly, J. 1989a. (Editors). The history of modern mathematics.
New York: Academic Press. 2 vols.
244
Bibliography
Runes, D. D. 1942a. (Editor). Dictionary of philosophy. New York: Philosophical

Library Inc.
Russell, B. 1888a. Greek Exercises. BRA. Ts. Contained in: Russell 1983a, 5-21.
_ _ . 1890a. Diary 1890-1894. BRA. Ms. Contained in: Russell 1983a, 41-67.
_ _ . 1893a. Paper on epistemology II. BRA. Ts. Contained in: Russell 1983a,
124-130.
_ _ . 1895a. The a priori in geometry. Proceedings of the Aristotelian Society 3:
97-112. [Also contained in: Russell 1983a, 291-304].
_ _ . 1895b. Introduction [to fellowship dissertation on the foundations of geometry]. BRA. Ts. Contained in: Russell 1983a, 258-259.
_ _ . 1895c. Alternative or supplementary introduction [to 1895b]. BRA. Ts. Contained in: Russell 1983a, 260-262.
_ _ . 1896a. Review of Hannequin 1895a. Mind 5: 410-417. [Also contained in:
Russell 1990a, 36-43].
_ _ . 1896b. The logic of geometry. Mind 5: 1-23. [Also contained in: Russell
1983a, 267-286].
_ _ . 1896c. (*) Various notes on mathematical philosophy. BRA. Ts.
_ _ . 1896d. On some difficulties of continuos quantity. BRA. Ts. Contained in:
Russell 1990a, 46-58.
_ _ . 1897a. An essay on the foundations of geometry. Cambridge: Cambridge
University Press.
_ _ . 1897b. Review of Couturat 1896a. Mind 6: 112-119. [Also contained in:
Russell 1990a, 60-67].
_ _ . 1897c. On the relations of number and quantity. Mind 6: 326-341. [Also
contained in: Russell 1990a, 70-82].
_ _ . 1897d. The philosophy of matter. BRA. Ms. Contained in Russell 1990a, 84.
_ _ . 1897e. Why do we regard time, but not space, as necessarily a plenum?
BRA. Ts. Contained in: Russell 1990a, 92-97.
_ _ . 1897f. Motion in a plenum. BRA. Ms. Contained in: Russell 1990a, 89-90.
_ _ . 1897g. On the conception of matter in mixed mathematics. BRA. Ms. Contained in: Russell 1990a, 86-87.
_ _ . ca 1897h. (*) On quantity with distinction of sign. BRA. Ms.
_ _ . ca 1897i. (*) On pure extensive (scalar) quantity. BRA. Ms.
_ _ . ca 1897j. (*) Calculus of extension. BRA. Ms.
_ _ . ca 1897k. (*) On equality and inequality. BRA. Ms.
_ _ . ca 18971. (*) Positional manifold. BRA. Ms.
_ _ . 1898a. Notebook on mathematical articles. BRA. Ms.
_ _ . 1898b. Les axiomes propes a Euclide. Sont-ils empiriques? Revue de Metaphysique et de Morale 6: 759-776.
_ _ . 1898c. On causality as used in dynamics. BRA. Ms. Contained in: Russell
1990a, 108-110.
245
_ _ . 1898d. (*) Note on order. BRA. Ms.

_ _ . 1898e. (*) An analysis of mathematical reasoning[,] being an inquiry into the
Subject-Matter, the Fundamental Conceptions, & the Necessary Postulates of
Mathematics. BRA. Ms. Contained in: Russell 1990a, 162-242.
_ _ . 1898f. On the principles of arithmetic. BRA. Ms. Contained in: Russell 1990a, 247-270.
_ _ . 1899a. The classification of relations. BRA. Ms. Contained in: Russell 1990a,
138-146 .
. 1899b. The fundamental ideas & axioms of mathematics. BRA. Ms. Contained in: Russell 1990a, 265-305.
_ _ .1899c. Sur les axiomes de la geometrie. Revue de Mhaphysique et de Morale
7: 684-707.
_ _ . 1899d. (*) On the difference of metrical and projective geometry. BRA. Ms.
_ _ . 189ge. (*) On the quadrilateral construction. BRA. Ms.
_ _ . 1899f. (*) Definition of a plane. BRA. Ms.
_ _ . 1899g. (*) A critical exposition of the philosophy of Leibniz. BRA. Ms.
_ _ . 1899h. (*) Principles of mathematics. BRA. Ms.
_ _ . 1900a. A critical exposition of the philosophy of Leibniz, with an appendix of
leading passages. Cambridge: Cambridge University Press.
_ _ . 1900b. (*) Necessity and possibility. BRA. Ms.
_ _ . ca 1900c. (*) Notes on Peano. BRA. Ms.
_ _ . 1900d. (*) On identity. BRA. Ms.
_ _ . 1900e. (*) The principles of mathematics. BRA. Ms. [Parts III to VI], contained in: Russell 1902f.
_ _ . 1901a. L'idee de l'ordre et la position absolue dans l'espace et Ie temps. 1st.
International Congress of Philosophy. Paris: Colin. 3: 241-277.
_ _ . 1901b. Sur la logique des relations avec des applications a la theorie des
series. Rivista di Matematica 7: 115-148. [Translated by R. C. Marsch and
contained in: Russell 1956c, 3-38] .
. 1901c. On the notion of order. Mind 10: 30-51.
_ _ . 1901d. Is position in space and time absolute or relative? Mind 10: 293-317.
_ _ . 1901e. Recent work on the principles of mathematics. International Monthly
4: 83-101. [Reprinted as: Mathematics and the metaphysicians, contained in:
Russell 1957a, 70-92].
_ _ . ca 1901f. (*) Recent Italian work on the foundations of mathematics. BRA.
Ms.
_ _ . 1901g. (*) Theorie generale des series bien-ordonnees. BRA. Ms.
_ _ . 1901h. (*) On well-ordered series. BRA. Ms.
_ _ . 1901i. (*) Part I: The variable. BRA. Ms. [Reconstructed according to Blackwell 1971b, 7].
_ _ . 1901j. (*) Part II: Number. BRA. Ms. [Contained in: Russell 1902f, Part II].
_ _ . 1902a. Geometry, non-Euclidean, contained in: Encyclopedia Britannica (10th
246
Bibliography
ed). 4: 665-674.
. 1902b. On finite and infinite cardinal numbers. American Journal of Mathematics 24: 378-383. [Section III od Whitehead 1902a].
_ _ . 1902c. Theorie generale des series bien-ordonnees. Rivista di Matematica 8:
12-43.
_ _ . 1902d. Letter: Russell to Frege, June 16, 1902. [printed in: van Heijenoort
1967b, 124-125, also in Frege 1980a, 130-131].
_ _ . 1902e. The teaching of Euclid. The Mathematical Gazette 2: 165-167.
_ _ . 1902f. (*) The principles of mathematics. BRA. Ms. [Manuscript "finished"
May 23, 1902].
_ _ . 1902g. (*) Journal: November 12, 1902 to April 3, 1903. BRA. Ms. [Contained in: Russell 1985a, 3-28].
_ _ . 1903a. The principles of mathematics. Cambridge: Cambridge University
Press.
_ _ . 1903b. Recent work on the philosophy of Leibniz [reviews of Couturat 1901a
and Cassirer 1902a]. Mind 12: 177-201.
_ _ . 1904a. The axiom of infinity. Hibbert Journal 2: 809-812.
_ _ . 1904b. Non-euclidean geometry. The Athenaeum # 4018 (October 29, 1904)
592-593.
_ _ . 1904c. Review of Couturat 1903a. Mind 13: 131-132.
_ _ . 1904d. (*) Plan for a general introduction to vol II [of 1903a]. BRA. Ms.
_ _ . 1904e. (*) Notes for a general introduction to vol II [of 1903a]. BRA. Ms.
_ _ . 1905a. On denoting. Mind 14: 479-493. [Reprinted in: Russell 1956c, 39-56].
_ _ . 1905b. Sur la relation des mathematiques a la logistique. Revue de Metaphysique et de Morale 13: 906-917. rrranslated by D. P. Lackey and printed in
Russell 1973a, 260-271].
_ _ . 1906a. On some difficulties in the theory of transfinite numbers and order
types. Proceedings of the London Mathematical Society 4: 29-53.
_ _ . 1906b. On the substitutional theory of classes and relations. BRA. Ms. [Read
before the London Mathematical Society in April 1906 and published for the
first time in: Russell 1973a, 165-189].
_ _ . 1906c. The theory of implication. American Journal of Mathematics 28:
159-202.
_ _ . 1906d. Review of MacColl 1906a. Mind 15: 255-260.
_ _ . 1906e. Les paradoxes de la logique. Revue de Meraphysique et de Morale 14:
627-650. [See: Russell 1906j].
_ _ . 1906f. (*) The paradox of the liar. BRA. Ms.
_ _ . 1906g. The nature of truth. Mind 15: 528-533.
_ _ . 1906h. Review of Poincare. Mind 14: 412-418.
_ _ . 1906i. Reply to Poincare's criticism of Russell 1906h. Mind 15: 141-143.
_ _ . 1906j. On "Insolubilia" and their solution by symbolic logic, contained in:
Russell 1973a, 190-214.
247
_ _ . 1906k. (*) Multiplicative Axiom. BRA. Ms.

_ _ . 1907a. The regressive method of discovering the premisses of mathematics.
[Read before the Cambridge Mathematical Club on March 9, 1907. Published
for the first time in: Russell 1973a, 272-283].
_ _ . 1907b. The study of mathematics. New Quarterly 1: 29-44. [Originally written in September 1902].
_ _ . 1908a. "If' and "imply". Mind 17: 300-301.
_ _ . 1908b. Mr. Haldane on infinity. Mind 17: 238-242.
_ _ . 1908c. Mathematical logic as based on the theory of types. American Journal
of Mathematics 30: 222-262.
_ _ . 191Oa. See: Whitehead & Russell 1910a.
_ _ . 191Ob. Philosophical Essays. London: Longmans Green.
_ _ . 191Oc. La throrie des types logiques. Revue de Metaphysique et de Morale
18: 263-301. [See: Russell 191Oe].
_ _ . 191Od. Some explanations in reply to Mr. Bradley. Mind 19: 373-378.
_ _ . 1910e. The theory of logical types, BRA. Ms. [English translation of Russell
191Oc, and printed for the first time in Russell 1973a, 215-252].
_ _ . 1911a. L'importance philosophique de la logistique. Revue de Metaphysique
et de Morale 19: 281-291. [See, for English translation, Russell 1913b &
Jourdain 1913d].
_ _ . 1911b. Sur les axiomes de l'infini et du transfini. Comptes Rendus des Seances, Societe Mathematique de France 2: 22-35. [Reprinted in the same journal
39 (1967) 488-501. English translation contained in: Grattan-Guinness 1977a,
162-174].
_ _ . 1911c. Review of Carus 1908a. Mathematical Gazette 5: 103-104. [Reprinted
in Carus 191Oa] .
. 1912a. See: Whitehead & Russell 1912a.
_ _ . 1912b. The problems of philosophy. London: William & Norgate.
_ _ . 1912c. Response a M. Koyre. Revue de Metaphysique et de Morale 20: 725-726 .
. 1913a. See: Whitehead & Russell 1913a.
_ _ . 1913b. The philosophical importance of logicism. The Monist 23: 481-493.
[Reprinted in: Russell 1973a, 284-294].
_ _ . ca 1913c. (*) Notes on Wiener's thesis. BRA. Ms.
_ _ . 1914a. Our knowledge of the external world. London: Open Court.
_ _ . 1914b. Mysticism and logic. Hibbert Journal 12: 780-803.
_ _ . 1918a. Mysticsm and logic, and other essays. London: Longmans Green.
_ _ . 1918b. The philosophy of logical atomism [part 1 of 4]. The Monist 28:
495-527.
_ _ . 1919a. Introduction to mathematical philosophy. London: George Allen &
Unwin.
_ _ . 1919b. The philosophy of logical atomism [part 2 of 4]. The Monist 29:
248
Bibliography
33-63.
_ _ . 1919c. The philosophy of logical atomism [part 3 of 4]. The Monist 29: 190-222.
_ _ . 1919d. The philosophy of logical atomism [part 4 of 4]. The Monist 29:
344-380.
_ _ . 1920a. Mathematical philosophy. Science Progress 15: 101.
_ _ . 1926a. On education especially in early childhood. London: George Allen &
Unwin.
_ _ . 1927a. The analysis of matter. London: Kegan Paul, Trench & Trubner.
_ _ . 1927b. An outline of philosophy. London: George Allen & Unwin.
_ _ . ca 1927c. Selected papers of Bertrand Russell. New York: The modern library. [Introduction dated by Russell: March 1927].
_ _ . 1927d. Things that have molded me. The Dial 83: 181-186.
_ _ . 1928a. Bertrand Russell catches the tortoise. The Forum 79: 262-263.
_ _ . 1929a. How I came to my creed. The Realist 1 (Sept) 14-21.
_ _ . 1931a. Review of Ramsey 1931a. Mind 46: 476-482.
_ _ . 1932a. Review of Ramsey 1931a. Philosophy 7: 84-86.
_ _ . 1937a. A critical exposition of the philosophy of Leibniz, ... London: George
Allen & Unwin. (2nd ed, new introduction).
_ _ . 1938a. My religious reminiscences, contained in: Russell 1961a, 32-36.
_ _ . 1938b. The principles of mathematics. London: Cambridge University Press.
(2nd ed, new introduction).
_ _ . 1940a. An inquiry into meaning and truth. New York: W. W. Norton.
_ _ . 1942a. How to become a philosopher: the art of rational conjecture. London:
Halderman-Julius Publications. (The how to series. Vol 7). p 5-16.
_ _ . 1942b. How to become a logician: the art of drawing inferences. London:
Halderman-Julius Publications. (The how to series. Vol 8). p 16-27.
_ _ . 1942c. How to become a mathematician. London: Halderman-Julius Publications. (The how to series. Vol 9). p 28-40.
_ _ . 1944a. My mental development, contained in: Schilpp 1944a, 3-20.
_ _ . 1944b. Reply to criticism, contained in: Schilpp 1944a, 681-741.
_ _ . 1945a. History of westem philosophy. New York: Simon & Schuster.
_ _ . 1946a. Preface to Clifford 1885a, v-x.
_ _ . 1948a. Hwnan knowledge: its scope and limits. London: George Allen &
Unwin.
_ _ . 1948b. Whitehead and Principia Mathematica. Mind 57: 137-138.
_ _ . 1948c. A turning point in my life. Saturday Book 8: 142-146.
_ _ . 1951a. Is mathematics purely linguistic? BRA. Ms. [Printed for the first time
in: Russell 1973a, 295-306].
_ _ . 1954a. Nightmares of eminent persons. London: The Badley Head.
_ _ . 1956a. Portraits from memory. London: George Allen & Unwin.
_ _ . 1956b. An unaltered republication of Russell 1897a.
249
1956c. Logic and knowledge, 1901-1950. London: George Allen & Unwin.
1957a. Mysticism and logic. New York: Doubleday & Company, Inc.
1958a. Mathematical infinity. Mind 67: 385.
1959a. My philosophical development. London: Unwin books. (Paperback
edition).
_ _ . 1959b. The influence and thought of G. E. Moore. The Listener 61 (April 30,
1959) 755-756.
_ _ . 1959c. The future of science with a "self-portrait" of the author. New York:
Philosophical library.
_ _ . 1961a. The basic writings of Bertrand Russell. New York: Simon & Schuster.
Editors: R. E. Egner & L. E. Denon.
_ _ . 1967a. The autobiography of Bemand Russell, 1872-1914. London: George
Allen & Unwin.
_ _ . 1967b. False and true. The Observer (March 12, 1967) 33.
_ _ . 1968a. The autobiography of Bertrand Russell, 1914-1944. London: George
Allen & Unwin.
_ _ . 1969a. The autobiography of Bertrand Russell, 1944-1967. London: George
Allen & Unwin.
_ _ . 1969b. Dear Bertrand Russell ... A selection of his correspondence with the
general public, 1950-1968. London: George Allen & Unwin. Editor and introduction by: B. Feinberg & R. Karrils.
_ _ . 1971a. Addeundum to my reply to criticisms, contained in: Schilpp 1944a,
xvii-xx. (4th ed).
_ _ . 1972a. My own philosophy (a new essay). Hamilton: McMaster University
Press. [Originally written in 1946].
_ _ . 1973a. Essays in Analysis. London: George Allen & Unwin. Editor: D. P.
Lackey.
_ _ . 1983a. Cambridge Essays, 1888-99. London: George Allen & Unwin. Editors: Kenneth Blackwell et al. (The Collected Papers of Bertrand Russell. Vol
I).
_ _ . 1984a. Theory of knowledge. The 1913 manuscript. London: George Allen &
Unwin. Editor: Elizabeth R. Eames, in collaboration with Kenneth Blackwell.
(The Collected Papers of Bertrand Russell. Vol VII).
_ _ . 1985a. Contemplation & Action, 1902-14. London: George Allen & Unwin.
Editors: Richard Rempel et al. (The Collected Papers of Bertrand Russell. Vol
XII).
_ _ . 1990a. Philosophical Papers, 1898-1903. London: Unwin Hyman. Editors:
Nicholas Griffin & Albert C. Lewis. (The Collected Papers of Bertrand Russell. Vol II).
_ _.
_ _.
_ _.
_ _.
Russell, B. & Russell, P. 1937a. (Editors). The Amberley Papers. New York: W. W.
Norton. 2 vols.
250
Bibliography
Russell, P. 1937a. See: Russell & Russell 1937a.

Sacks, G. 1975a. Remarks against foundational activity. Historia Mathematica 2:
523-528.
Sainsbury, R. M. 1979a. Russell. London: Routhledge & Kegan Paul.
Schilpp, P. A. 1941a. (Editor). The philosophy of Alfred Nonh Whitehead. Illinois:
Open Court.
_ _ . 1942a. (Editor). The philosophy of G. E. Moore. Chicago: Northwestern
University.
_ _ . 1944a. (Editor). The philosophy of Benrand Russell. Evanston & Chicago:
Northwestern University.
_ _ . 1971a. Some recollections of Bertrand Russell. Journal of Thought 6: 68-79.
Schmid, A. F. 1983a. La correspondence inedite entre Bertrand Russell et Louis
Couturat. Dialectica 37: 75-109.
Schneenwind, J. B. 1967a. John Stuart Mill (1806-1873), contained in: The Encyclopedia of Philosophy 5: 314-323.
Schoenflies, A. 1905a. tiber wohlgeordnete Mengen. Mathematische Annalen 60:
181-186.
_ _ . 1906a. tiber die logischen paradoxien der Mengenlehre. Jahresbericht der
_ _ . 1911a. tiber die Stellung der Definition in der Axiomatik. Jahresbericht der
_ _ . 1922a. Zur Erinnerung an Georg Cantor. Jahresbericht der Deutschen Mathematiker-Vereinigung 31: 97-106 .
. 1927a. Die Krisis in Cantor's Mathematischen Schaffen. Acta Mathematica
50: 1-23.
Schoenman, R. 1967a. (Editor). Benrand Russell: philosopher of the century. London: George Allen & Unwin.
SchrOder, E. 1898a. On pasigraphy: its present state and the pasigraphic movement in
Italy. The Monist 9: 44-62.
Scriba, C. J. 1978a. See: Burau & Scriba 1978a.
Seignobos, Ch. 1898a. See: Langlois & Seignobos 1898a.
251
Shaw, J. B. 1918a. Lectures on the philosophy of mathematics. London: Open Court

& PUblishing Co.
Shearman, A. T. 1906a. The development of symbolic logic. London: Williams &
Norgate .
. 1907a. Review of Russell 1903a. Mind 16: 254-265.
Slater, J. G. 1976a. One hundred years of Bertrand Russell: an appraisal. Russell #
23-24: 4-25.
_ _ . 1982a. * Bertrand Russell, polymath, printed (for limited circulation) by the
Thomas Fisher Rare Book Library of the University of Toronto.
_ _ . 1988a. Russell's conception of philosophy. Russell N. S. 8: 163-178.
Snaper, E. 1979a. The three crisis in mathematics. Mathematics Magazine 52: 207-222.
Spadoni, C. 1975a. Russell in metareview (review of Thomas & Blackwell 1975a).
Russell # 20: 19-21.
_ _ . 1976a. Great God in boots!-the ontological argument is sound. Russell #
23-24: 37-41.
_ _ . 1977a. Russell's rebellion against neo-Hegelianism. Ph. D. dissertation. University of Waterloo.
_ _ . 1978a. Philosophy in Russell's letters to Alys. Russell # 29-32: 17-31.
_ _ . 1979a. Old Sidg. Russell # 33-34: 45-51.
_ _ . 1979b. Russell and the English Idealist heritage. Russell # 35-36: 50-54.
_ _ . 1980a. Russell in Italian. Russell # 37-40: 65-67.
Stigt, W.T. Van. 1979a. See: Van Stigt.
Struik, D. J. 1989a. A concise history of mathematics. New York: Dover. (4th.
revised ed).
_ _ . 1980a. Why study the history of mathematics. The Journal of the Undergraduate Mathematics 1: 3-28.
Student's Guide. 1880a. The student's guide to the University of Cambridge. Cambridge: Deighton, Bell, & Co. (4th revised ed).
_ _ . 1880b. The student's guide to the University of Cambridge. Cambridge: Deighton, Bell & Co. [Part II: Mathematical Tripos].
_ _ . 1881a. The student's guide to the University of Cambridge. Cambridge: Deighton, Bell & Co. [Part VIII: Moral Sciences Tripos].
_ _ . 1893a. The student's guide to the University of Cambridge. Cambridge: Deighton, Bell & Co. (5th rewritten ed). [The Mathematical Tripos 1893al; The
252
Bibliography
Moral Sciences Tripos 1893a2).
Tait, K. 1975a. My father, Benrand Russell. New York: Harcourt Brace Jovanovich.
_ _ . 1976a. A daughter's eye view [review of Clark 1975a). Russell # 21-22:
51-56.
Tannery, P. 1887a. La Geometrie Grecque. Paris.
Taub, A. H. 1961a. See: Van Neumann 1961a.
Thomas, J. E. & Blackwell, K. 1976a. Russell in review. Toronto: Samuel Stevens,
Hakkert & Co.
Thomas, W. & Rang, B. 1981a. Zermelo's discovery of the Russell paradox. Historia Mathematica 8: 15-22.
Tiles, M. 1984a. Mathematics: the language of science. The Monist 67: 3-17.
_ _ . 1989a. The philosophy of set theory: an historical introduction to Cantor's
paradaise. London: Basic Blackwell.
Trevelyan, J. 1967a. An old friendship, contained in: Schoen man 1967a, 26-31.
Urquhart, A. 1988A. Russell's zigzag path to the ramified theory of types. Russell N.
S. 8: 82-91.
Van Ebra, J. 1991a. See: Roberts, Houser & Van Ebra 1991a.
Van Heijenoort, J. 1967a. Logical paradoxes, contained in: The Encyclopedia of
_ _ . 1967b. (Editor). From Frege to GlJdeI: a source book in mathematical logic,
1879-1931. Cambridge, Mass: Harvard University Press.
Van Neumann, J. 1931a. The formalist foundations of mathematics, contained in:
Benacerraf & Putnam 1964a. 50-54.
_ _ . 1947a. The mathematician, contained in: van Neumann 1961a I, 1-9.
_ _ . 1961a. Collected Works. London: Pergamon Press. 6 vols. Editor: A. H.
Taub.
Van Stigt, W. P. 1979a. The rejected parts of Brouwer's dissertation on the foundations of mathematics. Historia Mathematica 6: 385-404.
. 1991a. Brouwer's Intuitionism. Amsterdam: North Holland.
253
Vivanti, 1. 1900a. Abstract-Review of Burali-Forti 1897a & 1897b. Jahrbuch uber

die Fortschritte der Mathematik (Berlin) 28: 62-63.
Von Wright, G.H. 1974a. Letters ffrom WittgensteinJ to Russell, Keynes and Moore.
Oxford: Blackwell.
Vuillemin, 1. 1908a.
Le~ons
sur la premiere philosophie de Russell. Paris: A. Collin.
Wang, H. 1958a. Eighty years of foundational studies. Dialectica 12: 466-497.

_ _ . 1974a. From mathematics to philosophy. London: Routhledge & Kegan Paul
LTD.
_ _ . 1984a. Wittgenstein and other mathematical philosophies. The Monist 67: 18-26.
Wavre, R. 1934a. Is there a crisis in mathematics? With reference to the notion of
the existence and a doubtful application of the law of the excluded middle.
American Mathematical Monthly 41: 488-499. [Summarized translation by A.
Ambrose. Y-a-t-il une crise des mathematiques? A propos de la notion d'existence et d'une application suspecte du principe du tiers exclu. Revue de Metaphysique et de Morale 31 (1924) 435-470].
Webb, B. 1978a. See: Webb & Webb 1978a.
Webb, S. & Webb, B. 1978a. The letters of .... London: Cambridge University
Press. 3 vols. Editor: N. Mackenzie.
Weyl, H. 1944a. David Hilbert and his mathematical world. Bulletin of the American
Mathematical Society 50: 612-654. [Also contained in: Weyl 1968a IV, 130-172].
_ _ . 1944b. Obituary: David Hilbert (1862-1943). Obituary Fellows Royal Society
4: 547-553. [Also contained in: Weyl 1968a IV: 120-129].
_ _ . 1946a. Mathematics and logic. American Mathematical Monthly 53: 2-13.
[Also contained in: Weyl 1968a IV, 268-279].
_ _ . 1946b. The philosophy of Bertrand Russell. American Mathematical Monthly
53: 208-214. [Also contained in: Weyl1968a IV, 599-605].
_ _ . 1968a. Gesammelte Abhandlungen. Berlin: Springer Verlag. 4 vols.
Whitehead, A. N. 1898a. A treatise of universal algebra with applications, 1 [and
only]. London: Cambridge University Press.
_ _ . 1901a. Memoir on the algebra of symbolic logic [part 1 of 2]. American
Journal of Mathematics 23: 139-165.
_ _ . 1901b. Memoir on the algebra of symbolic logic [part 2 of 2]. American
254
Bibliography
Journal of Mathematics 23: 297-316.

_ _ . 1902a. On cardinal numbers. American Journal of Mathematics 24: 367-394.
_ _ . 1905a. Note. Revue de Metaphysique et de Morale 13: 916-917.
_ _ . 1906a. Theorems on cardinal numbers. American Journal of Mathematics 26:
31-32.
_ _ . 191Oa. The philosophy of mathematics. Science Progress 5: 234-239.
_ _ . 1926a. Principia Mathematica. Mind 35: 130.
_ _ . 1941a. Autobiographical notes, contained in: Schilpp 1941a, 3-14.
Whitehead, A. N. & Russell, B. 1910a. Principia Mathematica. Cambridge: Cambridge University Press. Vol 1. (2nd ed, 1925).
_ _ . 1912a. Principia Mathematica. Cambridge: Cambridge University Press. Vol
II.
_ _ . 1913a. Principia Mathematica. Cambridge: Cambridge University Press. Vol

III.
Wiener, N. 1913a. (*) Notes on the theory of types. Institute Archives M.LT.: Ms.
[Xerox copy at the BRA].
_ _ . 1913b. (*) Bertrand Russell's theory of the nature of reality. Institute Archives M.LT.: Ms. [Xerox copy at the BRA].
Wiener, P. P. (Editor). 1973a. Dictionary of the History of Ideas. New York:
Charles Scribner's Sons. 4 vols.
Wilder, R. L. 1953a. The origin and growth of mathematical concepts. Bulletin of
the American Mathematical Society 59: 423-448.
_ _ . 1974a. Hereditary stress as a cultural force in mathematics. Historia Mathematica 1: 29-46.
Willis, K. 1976a. The critical reception of German Social Democracy. Russell #
21-22: 35-45.
Wilson, E. B. 1904a. Review of Russell 1897a. American Mathematical Society
Bulletin 11: 74-93.
Winchester, I. 1988a. The antinomy of dynamical causation in Leibniz and the Principles and Russell's early picture of physics. Russell 8: 19-34.
Wood, A. 1957a. Benrand Russell, the passionate sceptic. London: Unwin books.
(paperback edition).
_ _ . 1959a. Russell's philosophy: a study of its development, contained in: Russell
1959a, 189-205.
255
Woodbridge, B. A. 1977a. (Editor). Alfred North Whitehead: a primary - secondary

bibliography. Ohio: Philosophy documentation center.
Wright, G. H. von. 1974a. See: Von Wright 1974a.
Young, G.C. 1906a. See: Young & Young 1906a.
_ _ . 1928a. See: Young & Young 1928a.
Young, W. H. & Young, G. C. 1906a. The theory of sets of points. New York:
Chelsea Publishing Co. [Revised reprint 1972].
_ _ . 1928a. Review of Hobson 1921a. The Mathematical Gazette 14: 98-104.
Zermelo, E. 1904a. Proof that every set can be well-ordered, contained in: van Heijenoort 1967b, 139-141. Translated by S. Bauer-Mengelberg. [Beweis, dass
jede Menge wohlgeordnet Werden Kann. Mathematische Annalen 59: 514-516].
_ _ . 1908a. A new proof of the possibility of well-ordering, contained in: van
Heijenoort 1967b, 183-198. Translated by S. Bauer-Mengelberg. [Neuer
Beweis fur die Moglichkeit einer Wohlordnung. Mathematische Annalen 65:
107-128].
_ _ . 1908b. Investigations in the foundations of set theory I [and only], contained
in: van Heijenoort 1967b, 199-215. Translated by S. Bauer-Mengelberg.
[Untersuchungen uber die Grundlagen der Mengenlehre I. Mathematische
Annalen 65: 261-281] .
. 1932a. See: Cantor 1932a.
Zlot, W. 1957a. The role of the axiom of choice in the development of the abstract
theory of sets. Ph. D. dissertation. Columbia University.
_ _ . 1960a. The principle of choice in pre-axiomatic set-theory. Scripta Mathematica 25: 105-123.
Index
Absolute, The: 53.

Abstraction: 16, 76.
Acta Mathematica: 38, 62.
Aggregate (see: class, multiplicity, set): 75,
143-144, 168, 174, 195.
Aleph: 6, 7, 33, 35, 39, 143; exponentiation:
138; Null (It.,}: 6-8, 36, 39, 94, 134 note
3, 147, 149.
Algebra: 50, 74, 76; Boolean: 55, 68 note 1;
foundations of: 49 note 2; linear associate: 47.
'All': 34, 94, 95, 97, 119.
Analysis: 42 note 3; aritmetization of: 74, 76.
Antinomy (see also: contradiction, inconsistency, paradox): xi note 1, 33, 69, 74,
140; Cantor-Burali-Forti: 113, 139; infinite number: 71, 74, 78; philosophical:
xviii, 62, 68, 74, 92, 94, 130, 152.
'Any': 92, 95, 160.
Apostles, The: 51.
Arithmetic: 68, 74, 76, 78, 88, 98, 121, 144;
foundations of: 15, 49 note 2, 61-79.
Astronomy: 50.
Axiom: 45, 48, 54.
Axiom of Choice (see also: Multiplicative
Axiom, Well-Ordering Theorem): 132
note 1, 139, 139 note 2, 144-145, 191,
204.
Bacca laureale: 50.

Baer, Richard: 163.
Bagley Wood: 166.
Berkeley, George (1685-1753): 25 note 3.
Berlin: 56.
Bernstein, Felix (1878-1956): 138, 140, 141,

167,204.
Berry, George G. (1867-?): 131, 132, 134137, 148 note 3; correspondence with
Russell: 166-184; contradiction: 167-168;
paradox of the lier: 179-180; paradox
visiting card: 166.
Biology: xii.
Birth control: 43.
Blackwell, Kenneth: 102, 134 note 2.
Bochenski, I: 25.
Bodleian Library (Oxford): 134, 166.
Bolzano, Bernard (1781-1848): 89.
Boole, George (1815-1864) 55, 85, 85 note 3.
Boots: 145, 145 note 3, 191.
Bosanquet, Bernard (1848-1923): xvii, 53, 55,
58.
Bournemouth: 157
Bowne, G.D.: 67.
Boyle, Dr.: 157.
Bradley, Francis (1846-1924): xvii, 51, 53,
55,58,83.
Britain: 61.
Bulletin de la Societe Mathematique de
France: 139 note 2.
Bunn, R.: 25, 25 note 3, 33.
Burali-Forti, Cesare (1861-1931): xvi-xvii,
xviii, 19, 84, 113, 128-129, 177; admirer
of Cantor: 21; Analyse vectorielle generale: 164; article 1897a: 21, 23-24;
article 1897b: 29, 29 note 1, 30; background: 21-22; contradiction: 123, 127129; correspondence with Russell: 164165; credit: 26; Logica Matematica: 22;
paradox: 21-32, 123, 131-132, 134, 141,
146, 167, 189, 193,204,206; (aparent):
258
31; (consequence): 19; (dating): 20;
(priority): 26, 32; (origin): 20, lS2; (formulation) 22-24; (reaction): 25-32; student ofPeano: 21-22.
Burnside: 174.
Calculus: 78; differential: SO, 62; infinitesimal: 6S; logical: 67, 68, 92; ofpredicates: 68 note 1; variational: SO; vectorial: 164.
Cambridge University: xii, xvii, S2, S6, 86,
101, 117, lS2, lS6, lS7, 160, 164, 18S,

188.
Cambridge University Press: S8, 83, 122 note
I, lS7.
Cantor, Georg (184S-1918): xv, xx note I, I,
2,9,64, 7S, 82, 94, 9S, 96, 97, 100,
101, lOS, 110, 128, 129, 130, 139 note
2, lSI, 174, 203; "Beitrige ... ": I, 3,
11, 29-30, 36-39, 43, 101, 109, 134 note
3; concept of set: 3; concept of number:
2-S, 16; consistent & inconsistent multiplicities: 143, 143 note 2, 14S; correspondence with Peano: 16-17; diagonal
class: 193; influence on Russell: 1-2, 42,
82, 109; discovery of the 'paradox': xvixvii, 20, 109, 123, 132, 167; (existence):
3S; (priority): 32, 34, 3S note 2, 40,
128, lS2; (dating): xvii, 20, 33, 34, 3S40; (origins): 33-3S; (reaction): 37-39;
father: 13; health: 13; "Mannigfaltigkeitslehre": S2; Mengenlehre: 20S; 'paradox': xvii; 32-40, 88, 108, 119, 122,
123, 127, 129; reaction to Cantor: 13-1S;
second number class: 62,97, 134, 134
note 3, 146, 149, 167; second principle
of generation: 63, 78, 173; theological
believes: 36; theorem of power set: 3S,
96, 97-98, 101, lOS, 109, 110, 114, llS,
119, 127-128; theory of transfinite numbers: xv, xviii, I, 2-13,22, 30, 61, 63,
64, 78, 79, 109, 14S; 'w': 10-ll, 28, 6263, 78-79, 94, 108.
Carlyle, Mrs.: lS9.
Chisholm, Grace (see: Young, Grace).
Circolo Matematico di Palermo: 22; Rendiconti del ... : 26 note 2.
CIulracteristica Universalis: 71.
Index
Class (see also: aggregate, set): 108, 146,
17S; finite: 87; infinite: 87, 96; limitation: lOS, 14S; multiplicative: 194; of all
individuals: 87, 96, 102; of all classes:
lOS, 114-11S, 119; perfecdy-ordered: 2324, 27-30; universal: 87, 96.
Clifford, William K (184S-1879): 47-49;
Common sense of the exact sciences, 1he:
47; Linle People, 1he: 47.
Cobden-Sanderson, Thomas J. (1840-1922):
43-44.
Coffa, A.: 110, l1S.

Colette (see: Malleson, Lady Constance).
Collections: 62, 7S-76, 78-79.
College of the City of New York: xii.
Commutative Law: 10.
Comte, Auguste (1789-18S7): 49.
Conics: 48, S1.
Containment: 84, 86.
Continuity: 6S, 7S, 78-79, 94, 9S, 98, 101,
109.
Continuum hypothesis: 34, 34 note 2, 137,
146.
Continuum: ll, 11 note 2,34 note 2,62,98,
123, 138, 146-147, 148, 174, 193,203.

Contradiction (see also: antinomy, inconsistency, paradox): xi note I, xiv, 24, 128129; Cantor-Burali-Forti: 30, 31; class of
all classes: xiv; class of all classes which
are not members of themselves: xvii,
lOS, 110, 113, 120, 132, lSI; empty
space: 60; greatest cardinal number: 108,
114; infinity divisibility: S9; logical or
mathematical: xviii; natural: S9; Peano's
principle: 108-109; point: 60; polemics:
31; relativity: S9, 67; unbounded extension of space: S9.
Copi, I.: 26, 26 note 2, 29, 138 note 2, 148.
Counting: 48, S4, 76, 109.
Couturat, Louis (1868-1914): xviii, 29, 31,
31 note 1, 63, 67, 84, 86, 100, 102, 113,
llS, 129; De l'injini mathhnatique: 64.
Creighton, Beatrice: ISS.

Crisis: 2S, 2S note 3, 143, IS2-1S3, 169.
Darwin, Charles (1808-1882): 43, S2.

Darwinian evolution: 47.
Dauben, Joseph W.: 2, 3 note 1, 13, 17, 33,

38-39.
De Morgan, Augustus (1806-1871): 55, 58.
Dedekind, Richard (1831-1916): 7 note 2,9,
13, 20, 33-36, 38-39, 75, 143; "'Was
sind und ... ": 65, 74.
Deduction: 54.
Definability (Notion of): 140, 174.
Definitional: 54.
Denumerability: 75, 95, 147, 149.
Derivates: 181-182.
Descartes, Rene (1596-1650): 47, 52.
Dialectics: 56, 58, 60, 65, 69.
Differential Equations: 50.
Diversity: 67.
Dixon, Alfred C. (1865-1936): 131, 144,
146-150.
[Donnely], Lucy: 158, 159.
Du Bois Reymond, Paul (1831-1889): 175,
176, 191, 194.
Dynamics: 42 note 3, 50, 60 note 1, 61, 119,
120, 123, 189.
Economic Theory: Xli, 56.

Education: xii.
Electricity & magnetism: 50.
Elliot, Lady Frances Anna Maria (see: Russell, Lady).
Empiricism: 15, 54, 61, 109.
England: 44, 61, 85.
Enseignement Mathematique, L'.: 31.
Epimenides: 132; paradox of: 132.
Epistematic: 70.
Epistemology: 58.
Epistle to Corinthians: 1 note 1.
Equation theory: 50.
Equivalence relation: 4, 9-10, 109.
Ethics: 51.
Euclid: 45, 48 note 1, 50.
Eulenburg, Albert (1840-1917): 39.
Existence: 6, 175.
Experience: 54, 58.
Extramarital relations: 43.
Fallacy (see: mistakes): 62, 100, 105.

Fellowship dissertation: 42, 54, 56-61.
Fifth postulate: 45, 48.
259
First International Congress of Mathematicians: 83.

First International Congress of Philosophy: 2,
68, 79, 81, 83-87, 91, 120, 151. 160,
164.
First World War: 185
Flexner, Helen (1871-1956): 89-90.
Formalism: 20, 153, 162.
Fourier's theorems: 50.
Freedom: 49.
Frege, Gottlob (1848-1925): xvi, 20-21, 58,
116, 120-121; Begriffsschrift: 52; correspondence with Russell: 121-122; criticisms of Cantor: xvi, 15-16; Die Grundlagen der Arithmetik: 15, 122, 129, 136.
French philosophers: 38.
Functions: elementary elliptic: 50; Legendre
& Bessel: 50.
Geometry: 42 note 3,45,60 note 1, 98;

foundations of: 42, 47, 49 note 2, 56,
123; non-Euclidean: 48, 55, 59, 76;
philosophy of: 51; plane coordinate: 50;
solid: 50.
German Idealism (see: neo-Hegelianism).
German mathematics: 61.
Germany: 129, 163.
G5ttingen: 162-163.
Grammar: 42 note 3.
Grantchester: 157ss.
Grattan-Guinness, Ivor: xx note 1, 1, 13,33,
36, 108 note 1, 110, 115, 134 note 1.
Green, Thomas H. (1836-1882): 52.
Grelling, Kurt (1886-1941): 132, 134.
Hadamard, Jacques (1865- 1963): 15,27,

138-139, 141.
Hamilton, William (1788-1856): 55, 165.
Hannequin, Arthur (1856-1905): 52 note 1,
64, L'Hypoth ... : 61-63.
Happiness: 49.
Hardy, G. H. (1877-1948): 143-145, 173,
175; Bel1rand Russell and Trinity: 188;
correspondence with Russell: 188-204.
Hausdorff, Felix (1868-1942): 184.
260
Hegel, Georg Wilhelm (1770-1831): xviii, 52,
56, 64, 65; influence on Russell: 52.
Hegelianism (see: neo-Hegelianism).
Heidelberg: 137; congress: 138, 146.
Heine, Edward (1821-1881): 2.
Helmholtz, Hermann (see: von Helmholtz,
Hermann).
Hilbert, David (1862-1943): 36 note I, 136,
137, 138, 143, 143 note 2, 174; correspondence with Russell: 162-163; notion
of well-ordered set: 27 note 4.
History: xii, 48.
Hobson, Ernst W. (1856-1933): 143-144,
148, 173, 175, 176, 188, 190, 193, 194.
Human happiness: 49.
Hume, David (1711-1776): 51-52.
Hungarian Academy of Sciences: 146.
Hydrodynamics: 50.
Hydrostatics: 50.
Idealism (see: neo-Hegelianism).

Identity: 67.
Inclusion: 67.
Inconsistency (see: antinomy, contradiction,
paradox): xv, 70; greatest cardinal &
ordinal number: xiv.
Individual freedom: 49.
Induction (principle of): 17, 54, 87, 94.
Infinite: actual: 64, 70, 71; number: 63;
potential & actual: 63, 64, 75, 79, 123.
Infinity: 62, 91, 98, 101, 109.
Instants: 76.
Instinct: 76.
Intention: 67.
International affairs: xii.
Intenuuional Monthly, The: 100, 101.
Intuition: 3, 16, 54.
Intuitionism: 20, 153.
Jahresversanunlung: 37.
Jevons, W. S. (1835-1882): 55, 58.
Jourdain, Philip (1879-1919): xxii, 27,34,
39,52,52 note I, 89,90, 100, 110, 129,
141, 143, 145, 171-172, 188-189, 192,
193, 203; consistent & inconsistent sets:
143; well-ordered set notion: 27 note 4.
Index
Kant, Inmannuel (1724-1804): xvii, 15, 52,

53-55,61,64; second antinomy: 62, 62
note 1; sybthetic a priori: 54; trascendental aesthetic: 55.
Keisel: 37.
Kennedy, Hubbert: 25-26.
Kinematics: 60.
Klein, Felix (1849-1925): 31.
Knowledge: 48-49, 50, 71.
Konig, Julius (1849- 1913): 131, 132, 204;
paradox: 134, 146-150; Zermelo-paradox: 137-140, 193-194.
Kronecker, Leopold (1823-1891): 13.
Languages: 16, 48; symbolic: 86.

Leibniz, Gottfried W. (1646-1716): xviii, 51,
86, 135, 161; Russell's book on: 69, 83,
160-161; logic: 69-71.
Linguistics: xii.
Literature: 48, 129.
Logic: xv, 120, 144, 164; algebra of: 67;
algebraic: 55, 68; equational: 55; mathematical: 85; subject-predicate: 83; symbolic: 4 note 3, 71, 74, 81, 83, 120, 123.
Logicism (see: Logicist thesis).
Logicist thesis: xv, xviii, 2, 20, 68, 70, 81,
94,98, 100, 100 note I, 101, 117, 120,
122, 140, 142, 153.
London Mathematical Society: 143, 144, 148.
London School of Economics: 62.
Lotze, RudolfHerrnann (1817-1881): 58.
Magnitude: 67.
Malleson, Lady Constance (1895-1975): xxi.
Manifold: 65, 67, 68.
Marcolongo, Roberto (1862-1): 164.
Mathematische Annalen: 138, 141.
Mathematics: 47-48; defective: 60-61, 65, 67;
foundations of: xii, xiv, xv, xvi, 113;
Pure: xiv, xv, 71-72, 81, 92, 94, 102,
120.
Matter: 60, 69, 75; & Motion: 119, 158.
McTaggart, Ellis (1866-1925): 53, 166.

Mechanics: 48, so.
Membership (or pertenence): 84, 86, 116.
Metaphysics: 51, 55, 65, 70, 74.
Mill, John S. (1806-1873): IS, 43, 51, 52,
53-55, 64; autobiography: 49; System of
Logic, A: 49 note 2, 54; Three Essays on
Religion: 49 note 2; Utilitarianism: 49
note 2.
Mind: 64, 86.
Mistakes (also fallacies & difficulties): 40,
62, 82, 96, 100, 101, 102, lOS, 115,
128, 151.
Mittag-Leffler, Gosta M. (1846-1927): IS,
38.
Molocl: 158.
Monads: 70.
Monist, The: 84.
Moore, E. H. (1862-1932): xviii; correspondence with Cantor: 205-206.
Moore, G. E. (1873-1958): xx note I, 68, 69,
70, 71, 75, 83; correspondence with
Bertrand Russell: 160-161.
Moore, G. H.: 132 note 1.
Moral sciences: xii.
Motion: 75, 101; absolute vs relative: 60.
Multiplicative Axiom (see also: Axiom of
Choice, Well-Ordering Theorem): 144145, 191.
Multiplicities (consistent & inconsistent): 3437,39, lOS, 143.
Murray, Gilbert (1866-1957): 129.
Natural Sciences: xii.

Nelson, Leonard (1882-1927): 132, 134.
neo-Hegelianism 2, 42, 47, 49, 52-69, 74,
87, 96, lSI, 160.
Newton, Sir Isaac (1642-1727): 25 note 3;
Principia: SO.
Newtonian physics: 48.
Nobel Prize: xii.
No-Class Theory: 145.
Norm (see also: propositional function, set):
143-144, 191-192.
Number: 42,48,54,64,65,67,69,71,74,
75-76, 92, 110; all numbers of: 74, 78; a
priori: 54, 68; commutative law: 10;
concept of: xv, xvi, 2, 55, 60 note I, 68,
261
123, 151; distribuitive law: 11; finite cardinal: 4-7, 10, 11, 12, 16,22,35,39,
95, 109, 174; foundations: 1; greatest
cardinal: 36,97, 100, 101, 114-115, 119,
127-128, 144; greatest ordinal: 22; imaginary: 139; infinite: 55, 63-64, 68, 70,
74, 78, 92, 94, 109, 152; incommensurable: 139; irrational: 75, 78, 91-92;
last number: 74; largest number: 74; natural: xvi, 2, 55, 62, 75, 78; negative:
139; of numbers: 78, 96, 143; ordinal: 910, 12-14,23,31,35, 113, 128-129,
134,142,147,149,167,171,172,179181, 182, 183, 205-206; (type): 9, 128;
rational: 11, 13, 75; real: 34 note 2, 75,
174; transfinite cardinal: 6-9, 10, 16, 22,
95, 97, 108; transfinite ordinal: 22-24.
O'Briant: 70.
Olivier, Louis (1854-1910): 141.
One-to-one correspondence: 7, 9, 87, 97,
109-110.
Optics: so.
Order: 4, 9, 10, 12, 13, 16, 67, 75, 78, 79,
92,94.
Pacifism: xii.
Paradox (see: antinomy, contradiction, mistake); xi note I, 100-117, 130, 135; all
cardinal numbers (see also: Cantor): 34;
all ordinal numbers (see also BuraliForti): 34; autological & heterological:
132; Berry: 134-137; Burali-Forti: 116;
Cantor-Burali-Forti: 141; class of all
classes: xix, 81, 116; epistemological:
153; finite cardinal numbers: 32, 34;
greatest cardinal numbers: XIX, 81, 82,
114, 116, 193; greatest ordinal number:
21-24; infinite number: 65; Konig-Zermelo: 134, 137-140; linguistic: 153; logical: xix, 20, 130, 153; Moore: 205-206;
non-logical: 131-132, 134; response: xvi;
semantic: xix, 20, 131-150, 153; set of
all sets: 32; universal set: 32, 113; visitmg card: 135; Zermelo-Konig: 141.
262
Paris (see also: First International Congress of
Philosophy): 2, 81, 83, 89, 91, 151.
Peano, Giuseppe (1858-1932): xvi, 2, 21, 26
note 2, 31, 41, 58, 71, 81, 82, 89, 91,
94, 100, 101, lOS, 108, 117, 120, 121,
129, 142, 152, ISS, 160, 163; arithmetic:
185; criticism of Cantor: xvi, 16-17; influence on Russell: xviii, 1-2, 83-87,
100-101; notation: 166; postulates: 16;
principle: lOS, 108, 113, 116; program:
21.
Peirce, Charles S (1839-1914): 7 note 2, 58,
85 note 3.
Pembroke Lodge: 44,47.
Philosophy: xiv, SO; elements of: 51; history
of: 51; in general: xii.
Pieri, Mario (1860-1913): 84 note I, 174.
Physics: xii, 60; foundations of: xii, 42, 60
note I, 61.
Physiology: xiv.
Plato (428- 348 BC): 52, 69; influence on
Russell: 83.
Poincare, Henri (1854-1912): 13, 31 note I,
42, 60, 140, 142, 147 note 4.
Points: 59, 60, 76. 174; set of: 2 ..
Politics: xii.
Power (cardinal number): 3-4.
Plurality: 54.
Predicates: lOS, 108, 119, 121, 174-175, 192.
Prime Minister: 44, 48.
Principle of Excluded Middle: 37.
Probability: xii.
Proceedings of the London ... : 143.
Progress: 48-49.
Progression: 183-184, 190.
Propositional function (see also: norm): 144145,192.
Psychologism: 15-16.
Psychology: 15.
Quaker: 56.
Quantity: 42, 61-62, 64, 65, 74, 75, 92, 94.
Ramsey, Frank P. (1904-1930): 137.

Ravenscroft Estate: 43.
Realism: 82, 87, 96.
Index
Reductio ad absurdum: 24, 26, 30, 35, 37
note 4, 128.
Relations: 86.
Relativity theory: 32.
Religion: xii, 43, 47, 53.
Rempel, R: 56.
Revue de Metaphysique ... : 84.
Revue Generale des ... : 141, 146.
Richard, Iules (1864-1921): 132, 146, 148,
148 note 3; contradiction: 140-142; paradox: 140.
Richmond Park: 44.
Riemann, Bernard (1826-1866): 59.
Rivista di Matematica: 100, 101, 109, HO,
117, 123, 134, 135, 167.
Routh, E. I. (1831- 1907): 48; The elementary ... : 48.
Russell, Agatha (Aunt): 44, 159.
Russell, Alys (1868-1949): xx note I, xxi
note I, 56, 85 note 3, 89, 117, 119-121;
correspondence with Bertrand Russell:
155-159.
Russell Archives, The Bertrand: xix, 61, 71,
89, 90, H4 note 2, 166.
Russell, Bertrand (1872-1970): 19, 26, 130,
144; Analysis of mathematical reasoning,
An: 42, 65-71, 151; antinomies: 74;
autobiography: xx, xx note 1, 114; background: xvii-xviii, 41-79, 152; Cantor's
contradiction: 40, 88, 113, 119; Cantor's
influence: 42-43, 62, 63, 75-79, 82, 109,
151; childhood & adolesence: xvii, 4349; contradiction: xvii, 129; (consequences): 19; (dating): xxi, 1, 20, 105, 110,
113-116, 123-129; (first appearance) 105,
110; (origins): 1000, 102, 108; correspondence with Frege: 121-122, 127;
correspondence with Moore: 71, 160161; Critical Exposition ... , A.: 70; diary: 47-48, 117; discovery of the inconsistencies: xvi, xviii-xix, 20, 35 note 2,
114; (priority): 40; dismissal of Lectureship: 188; early education: 41, 45, 48-49;
egotism: xxiv; Essay on the Foundations
... : 56-61; Fellowship dissertation: 42,
54, 56-61, 185; Frege's influence on
Russell: 52 note 1, 123; "Fundamental
Ideas & Axioms ... ": 71-74; German Social Democracy: 56; indoctrination: 52;
Kantian influence: 76; Leibniz's book:

160-161; Mathematical Tripos: 50 note I,
51, 155; (first) marriage: 56, 155; mathematical studies: 50-51; "Metaphysical
basis of geometry": 58; Our knowledge
... : xxi; Peano's influence: 83-87, 160161; philosophical studies: 51-55; Principia Marhemarica: xv, xv note I, 132-133,
185, 186; "Principles of Mathematics":
xviii, 74-79, 83, 89, 119-120; Principles
of Mathematics, The: xv note 1, xvii, xx,
31, 134-135; (composition): 82; (cronology): xxii-xxiv, 90-92; (early drafts):
xviii, 88; (first final draft): 87-100;
(Preface): 91, 122-123; (Part I): 82, 88,
91, 92, 102, 109, 110, 113, 115, 152;
(Part II): 82, 88,91,92, 102, 109-110,
117, 123, 152; (Part III): 82, 88,90,90
note 2, 92, 152; (Part IV): 82, 88, 90,
90 note 2, 94, 152; (Part V): 82, 88, 90,
90 note 2, 94-98, 109, 127-128, 152;
(Part VI): 82, 88, 90, 90 note 2, 98- ,
152; (Part VII): 82, 88, 90, 91, 119,
152; (Appendixes): 82, 91, 122, 152;
Realism: 82, 87, 96, 108; "Recent developments ... ": 101; "Relations of sense
data to physics, The": xxi; Religious
doubts: 50; setback: 114, 117, 155; student at Cambridge: 42, 43, 49-56; tutors:
43, 45; What shall I read: 49 note 2, 51,
75; Whitehead's influence: 67-68; writing
methodology: xxii-xxiii.
Russell, Frank (1865-1931): 43-45, 54.
Russell, John (Viscount Amberley 1842-1876): 43-44.
Russell, Lady (1815-1898): 44-45, 56.
Russell, Lord John (1792-1878): 44-45.
Russell, Raquel (1868-1874): 43.
Russell, Rollo (Uncle): 44.
Sacks, G: xv note 1.
Schoenflies, Arthur (1853-1928): 37-38, 141,
193.
SchrOder, Ernst (1841-1902): 58, 84, 85, 85
note I, 85 note 3, 86.
Sciences: 48; social: xii; defective: 56, 59.
Second International Congress of Mathematicians: 162.
263
Segment: 12, 12 note 1.

Sensation: 55.
Sequences (see: series): 5-7, 8, 10,63,78.
Series: 7,22,75,76,78,79,94,97, 120.
129; trigonometric: 2; well-ordered: 176 .
Set (see also: aggregate, class, mUltiplicity):
142, 143, 146; actual infmite: 15, 68, 70,
142; all ideas: 6; all ordinals: 35; Cantor's concept: 3, 16; consistent & inconsistent: 143; equivalence: 9-10; finite: 6,
96; infmite: 6-8, 62, 64, 142; power set
of a set (cardinal number): 3, 8, 16, 32;
(cardinality): 105; real numbers: 140;
similarity of: 9-11; simply ordered: 9-11;
well-ordered: 11-14, 11 note 4, 27-30,
167-168.
Sex: 47.
Sidgwick, Henry (1838-1900): 51, 58.
Similarity: 9-10, 11,87,95.
Smith, Alys Pearshall (see: Russell, Alys).
Social progress: 48-49.
Social questions: xiv, 56.
Socrates: 86.
Soul: 53.
Space: 58, 59, 60, 69, 75, 98, 120.
Spadoni, Carl: 52, 64.
Spalding, Douglas A. (ca. 1840-1877): 43.
Species: 183.
Spencer, Herbert (1820-1903): 52.
Spinoza, Benedict (1632-1697): 52.
Standard Interpretation: xvi-xvii, 19-40, 19
note I, 152. .
Stanley Katherine (Viscountess Amberley
(1842-1874: 43.
Statics: 50, 5 I, 185.
Statistics: xii.
Stout, George F (1860-1944): 51, 86, 114
note 2.
Subsequence (see also subset): 12.
Substance: 53, 70.
Subset: 13.
Succession (see also: series, sequence, set):
55.
Switzerland: 155, 156.
Symbolism: 21-22
Synthesis: 56.
Taylor, Helen (1 - 1907): 43.
264
Theory of relations: 70, 86-87, 110, 120.
Theory of Types: 122, 132, 134, 149.
Third International Congress of Mathematicians: 137.
Thought: 3, 16.
Time: 69, 75, 120.
Transactions of the American Mathematical
Society: 205.
Trelleck (MoDlDOuthshire, Wales): 43.
Tricotomy law: 12,22-24,29, 31, 84, 97,
ll3, 128.
Trigonometry: 50.
Trinity College: 42, 49, 69, 117, 185.
Types (see: Theory of Types): ordinal: 9-10,
22-24; (addition & multiplication): 10.
University of Chicago: xx note I, 205.

University of Oxford: 166
University of Turin: 22.
United States of America: 61, 205.
Unity: 54, 67, 76.
Viscount Amberley (see: Russell, lohn).
Viscountness Amberley (see: Stanley, Katherine).
Vivanti, Giulio (1859-1949): 25.
von Helmholtz, Hermann (1821-1894): 59.
Ward, lames (1843-1925): 51, 52 note 1.

Webb, S & B: 157.
Weierstrass, Karl (1815-1897): 51, 71, 76.
Well-Ordering Theorem, The( see also: set
(well-ordered): xix, xix note 2, 12-13,
31, 34, 35, 38, 40, 131, 132, 134, 137,
140, 143-146, 147 note 4, 149, 188.
Whitehead, Alfred N. (1861-1947): xv, xxiii,
49,51,58, 64, 123, 123 note I, 129,
134, 158, 160, 166, 171, 189; correspondence with Russell: 185-187; criticism of
Russell: 186; multiplicative class: 190;
telegram to Russell: 186; Universal Algebra: 65,67,68 note I, 74, 85 note 3,
160.
Whitehead, Evelyn: 114, 158, 159.
Whole (& part): 74,75, 109, 171.
Index
Women's sufrage: 43.
Wood, Allan: xxii-xxiii, 42 note 3, 60 note 1.
Wright, R. T.: 159.
Young, Grace (1868-1944): 29, 36 note I,
51.
Young, William H. (1863-1942): 29.
Zeno: 86, 95, WI, 102.

Zermelo, Ernst (1871-1953): xx note I, 31,
33,40, 131, 132, 140, 143, 167, 188,
191, 194, 203; Konig paradox: 137-140;
paradox: 134; (priority over Russell):
136.
Zig-Zag Theory: 144.

Bertrand Russell and The Origins of The Set-Theoretic 'Paradoxes'. Alejandro R. Garciadiego

Uploaded by

Document Information

Original Title

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

Bertrand Russell and The Origins of The Set-Theoretic 'Paradoxes'. Alejandro R. Garciadiego

Uploaded by

Copyright:

Available Formats

Bertrand Russell In 1907

Library of Congress Cataloging in Publication Data

1. Logic, Symbolic and mathematical. 2. Set theory. 3. Paradox. 4. Russell, Bertrand,

Cover deSign: Albert Gomm

I've have put in so many

TO LUPITA, GABRIELA & ALEJANDRO

Prologue by Ivor Grattan-Guinness

Chapter II. A standard interpretation

Chapter III. The philosophical and mathematical background to The Principles

Chapter IV. Russell's discovery of the 'paradoxes'

The history of the paradoxes of set theory is well known. Cantor

and reinterpretations to the emergence of the 'semantic' paradoxes (as

The love of power is terribly strong

Russell's published works include more than sixty books, several

These three passions that governed his life-which extended for

1. Russell 1967a, 13.

Russell & the 'paradoxes'

bl, .....t- ,.~

f~~ " ~ ..,. ci

h.4'i- A.. ~.6c....

P"....~ ~ fII~c~ ~ .. /o~

Plate 2. Russell drafted this version of his 'contradiction' at a

inconsistencies of the theory of transfinite cardinal and ordinal numbers. I

Russell & the 'paradoxes'

When asked about his greatest contribution to philosophical

Other scholars who were interested in constructing a more solid

{I, 2, 3, ... , n, ... }

Russell & the paradoxes ,

2. See: Moore & Garciadiego 1981a.

the philosophy of Gottfried Wilhelm Leibniz (1646-1716). At this time,

sponding with the French philosopher Louis Couturat (1868-1914).

Russell & the 'paradoxes'

trates my argument of ho~ Russell created the 'paradox' of the greatest

1. Ramsey I926a, 353.

spondence. 1 Some are included because they reflect Russell's emotional

2. Professional historians may skip this section.

Russell & the 'paradoxes'

The disadvantages of relying on the use of autobiographical sources

2. Russell 1968a, 96. See, also: Ibid., 151.

Allan Wood, one of Russell's best known biographers, provided an

Other problems are related to Russell's very personal writing style.

2. Jourdain 1919., 454.

Russell & the 'paradoxes'

There were successive periods of intense thinking, each of which culminated in a

These sentences accurately describe Russell's mature methodology, the

1. Wood 1959a, 203.

2. Cf Blackwell 1984a, 275-277.

2. See: Garciadiego 1983a.

Russell & the 'paradoxes'

Numerous professionals facilitated my work through different

Russell & the 'paradoxes'

lish: Whitehead's missives to Russell, Hardy's letters to Russell and E.

The time will come when these

1.1 Introduction. - The main goal of this introductory chapter is

1.2 Cantor's transfinite numbers.- The story of how Cantor

Russell & the 'paradoxes'

He then continued to define what he meant by the power or cardinal

2. Cantor, 1915a, 89.

Immediately following this discussion, Cantor defined what he

Russell & the 'paradoxes'

then, by applying the definition of a sum of cardinal numbers, one obtains,

cardinal numbers, which, in magnitude, lie between two consecutive

Russell & the 'paradoxes'