You are on page 1of 19

bs_bs_banner

Ratio Juris. Vol. 27 No. 3 September 2014 (311–29)

Starting Points: Kantian Constructivism


Reassessed
CARLA BAGNOLI

Abstract. G. A. Cohen and J. Raz object that Constructivism is incoherent because


it crucially deploys unconstructed elements in the structure of justification. This
paper offers a response on behalf of constructivism, by reassessing the role of such
unconstructed elements. First, it argues that a shared conception of rational agency
works as a starting point for the justification, but it does not play a foundational
role. Second, it accounts for the unconstructed norms that constrains the activity
of construction as constitutive norms. Finally, on this basis, it draws a contrast
between constructivist and foundational methods of ethics, such as deontology and
teleology.

The prospects of political liberalism depend heavily on the viability of construc-


tivism. Constructivism promises an objectivist conception of rational justification
that avoids the well-known epistemological and ontological difficulties of moral
realism. Constructivists purport to avoid such difficulties by embracing two dif-
ferent strategies. On the one hand, some simply bracket metaethical issues and
defend constructivism as a first-level normative account of the principles of justice.
This strategy is best exemplified by John Rawls’ conception of “political construc-
tivism” (Rawls 1993).1 On the other hand, some constructivists conceive of con-
structivism as a distinct metaethical doctrine, positively opposed to standard forms
of realism and relativism. This is the Kantian constructivist position that Onora
O’Neill (1988, 1; 1999, 2003a) advocates in contrast to Rawls. Kantian constructiv-
ism is more ambitious than political constructivism in two important respects. First,
it responds to the need to take a stand in metaethics, and particularly to claim a
distinctive place in the debate about the ontological status of normative proposi-
tions. Second, it is not limited to political principles of justice or moral obligations,
but it concerns all sorts of normative propositions about reasons. Third, it has

1
As I make clear in Section 1, constructivism as a first-level normative account can be
motivated also with the metaethical conviction that constructivism fails as a general view of
normative propositions. I call this “moral constructivism,” since its supporters are especially
concerned with moral obligations. For reasons that will be apparent, I do not regard Rawls’
“political constructivism” as a metaethical position. In this sense, I disagree with Tampio
(2012).

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main
Street, Malden 02148, USA.
312 Carla Bagnoli

significant consequences for the issue of political legitimacy, which is at the core of
political liberalism. This is because it claims that the results of constructions have
unrestricted authority, rather than being authoritative only for “reasonable people”
already interested in justice.
The most threatening objections to political liberalism come from metaethics,
even though they are rarely addressed as such. Debates on political legitimacy and
debates about the prospects of constructivism as a distinct metaethical option have
only recently merged (see Street 2010; Bagnoli 2011a). Critics argue that construc-
tivism is not a suitable basis for liberalism because it cannot avoid arbitrariness
without relying on an unconstructed foundational basis (see Cohen 2003; Raz 1990,
15; 2003; Timmons 2003).2 In short, it is not a coherent, stable, and self-standing
metaethical doctrine. If these critics are right, the agenda of constructivist liberalism
is self-defeating and liberals should look elsewhere for resources to ground their
conception of political legitimacy. Indeed, some liberals undertake this path exactly
because they share this criticism and argue that liberalism should espouse realism
(Larmore 2008; Bagnoli 2009).
In this article, I endeavor to reassess the prospects of Kantian constructivism as
a metaethical basis of political liberalism. The scope of my argument is limited in
two ways. First, it does not engage in a defense of political liberalism. Second, the
argument is meant to establish the coherence of a thoroughly constructivist rational
justification. Rather than a complete defense of constructivism or an attack on
realism, this is an attempt to show that some common objections from metaethics
are insufficient to defeat constructivism as a general approach to rational
justification.
This is how I shall proceed. In Section 1, I identify the basic claims of Kantian
constructivism understood as a metaethical doctrine. The main body of the article
is devoted to discussing the nature and role of unconstructed elements that feature
in the constructivist account of rational justification. In Section 2, I consider the
basis of construction, which is a conception of rational agency, showing that its role
is not foundational. In sections 3–5, I focus on the status and scope of procedural
constraints, which are formal and constitutive rather than material. By clarifying
the nature of constraints, I draw a contrast between constructivism and founda-
tional methods of ethics, such as deontology and teleology. Finally, in Section 6, I
respond to the objection that Kantian constructivism necessarily presupposes the
value of humanity, by proposing a dialogical view of respect and mutual recog-
nition. I conclude that Kantian constructivism survives current objections from
metaethics and represents a viable basis for political liberalism.

1. Varieties of Kantian Constructivism


Kantian constructivism entered the debates in political philosophy with Rawls’
Dewey Lectures: Kantian Constructivism in Moral Theory, where he traced the roots of
his theory about justice as fairness back to Kant’s moral theory (Rawls 1980, 1993,
1999a, 1999b). Rawls’ interest in reviving Kant’s moral theory arises out of a specific

2
Larmore (2008) criticizes constructivism in general, aiming specifically at Korsgaard and
Rawls. Similarly, Budde (2009, 199) focuses on O’Neill’s account of justice but she points to
the incoherence of constructivism in general.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 313

diagnosis about the inadequacy of foundationalism to address ethical disagree-


ments that are distinctive of a pluralist society. Rawls is particularly interested in
the problem of coordination through just institutions given the fact of value
pluralism. To address this problem, he argues that we would profit from retrieving
Kant’s moral theory, which had been absent from political debates until then.
According to Rawls, Kant offers an objectivist justification that claims to be
authoritative for everyone, regardless of interests, positions, and specific values.
Kant’s constructivism applies to all sorts of normative propositions, which are
supposed to be necessary and a priori true cognitions. Kant scholars disagree about
whether constructivism accounts for the obligatory nature and authority of norma-
tive propositions, or for their content.3 I shall set this exegetical issue aside and focus
on Kant’s legacy. While placing themselves in the Kantian tradition, contemporary
constructivists take different positions concerning the nature of normative propo-
sitions.4 To begin with, Rawls defends constructivism as a “political” conception,
which withdraws from debates about ontological and metaphysical questions
(Rawls 1993, 100; 1999a, 395, 354). This position is not merely an expression of
agnosticism about the ontological status of reasons, nor is it motivated by
metaethical worries. Rather, the driving concern of political constructivism is the
issue of political legitimacy. In contrast to Kant and other sorts of Kantian
constructivists, Rawls is not concerned with the standard of practical rationality tout
court, but with the standards of political legitimacy in a pluralist democracy (see
Rawls 1993, 49, 116; 1999b, 473–96; 1999c, 421–48). He addresses the question of
political legitimacy as it arises for “reasonable” fellow citizens in a pluralist
democracy who are already interested in justice and accept the burdens and
standards of cooperation. According to Rawls, “our exercise of political power is
fully proper only when it is exercised in accordance with a constitution the essentials
of which all citizens as free and equal may reasonably be expected to endorse in the
light of principles and ideals acceptable to their common human reason” (Rawls
1993, 137, see also 139–40, 143, 217, 225–6; see Freeman 2007). Rawls’ reasons for
defending constructivism as a political doctrine are first of all political, even though
they have important consequences for metaethics. In his view, realist doctrines are
mistaken about the standards of political legitimacy in a pluralist democracy, and
fail to properly address the issue of ethical disagreement. They assume that the
standards of political legitimacy are the same as the theoretical standards of
epistemological inquiries, which are guided by truth. But the appeal to truth in a
pluralist context is divisive, leading to parochial conceptions of political authority.
By contrast, Rawls argues that the necessary and sufficient condition for a doctrine
to be admissible is not its truth, but its acceptability to all reasonable

3
Kant’s constructivism is thus a form of practical cognitivism, see Bagnoli 2013, 2011b. For
this reason I disagree with Budde that Rawls is a Kantian constructivist, for reasons also
advanced in O’Neill 2003a. Rawls proposes a “procedural interpretation of Kant’s conception
of autonomy and the categorical imperative,” precisely to avoid Kant’s “dubious metaphys-
ics” of transcendental idealism (Rawls 1971, 256; 2000, 239). In the course of this discussion,
I thus use the name Kantian constructivist only for O’Neill’s and Korsgaard’s theories.
4
It is beyond doubt that Kant’s constructivism is a form of cognitivism, since Kant thought
moral judgments to be necessary and a priori normative truths. I thus disagree with Ronzoni
and Valentini (2008, 419) that “ontological agnosticism” is the general “animating concern of
constructivism.”

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
314 Carla Bagnoli

citizens.5 Political liberalism is the view that this criterion of admissibility should
regulate the discussion about the basic structure of our society, and, more generally,
it should govern the coercive use of power.
Unlike Rawls’ political constructivism, this third variety of (moral) constructiv-
ism is motivated by skepticism about the prospects of Kantian constructivism as a
global account of normative propositions. For instance, Thomas Scanlon defends
moral constructivism as a first-level account of moral obligations (Scanlon 1998,
chap. 1; Hill 2008). He agrees with realists that “constructivism all the way down”
is doomed to failure because rational justification inevitably makes use of
unconstructed and substantive elements, such as the primitive notion of reason. For
this reason, moral constructivists are not “Kantian constructivists” in the sense that
is under scrutiny in this article (Scanlon 2003, 2012).
In contrast to these political and moral varieties of constructivism, Onora O’Neill
has forcefully argued for an ambitious form of Kantian constructivism, whose
scope is broad enough to include all sorts of normative propositions. She insists
that if constructivists like Rawls fail to provide an adequate account of political
legitimacy, it is only because of their reluctance to undertake a thoroughly con-
structivist account of justification (see O’Neill 1988, 2002, 2003a).6 Against the
mainstream view, O’Neill holds that Kant’s original constructivism provides a
better framework for moral and political justification than our contemporaries
(O’Neill 1989; 2002, ix–x). This is because Kant deploys only weak, uncontroversial,
merely abstract elements to count as “materials of construction” (O’Neill 2002, 7,
48, 64, 179, 210), including a plausible account of rational agents, marked by
specific competences, vulnerabilities, and mutual dependence. Rawls’ error is that
he departs from Kant’s abstract and sober list of materials and instead deploys
controversial idealizations of rational agency and personhood, which are inad-
equate and unrealistic (ibid., 64, 93, 179, 211). Clearly, O’Neill is well aware that this
claim is bound to attract immediate reactions, given that Kant’s ethics is widely
believed to rest on a cumbersome metaphysics, requiring a rather unrealistic sort
of psychology. In claiming allegiance to Kant, O’Neill distinguishes her construc-
tivist theory from Rawls’ under two main aspects: “In the first place it assumes only
an abstract, hence nonidealizing and banal account of agents and of conditions of
action. Secondly, it aims to articulate and to vindicate a conception of practical
reasoning without appeal either to unvindicated ideals or to unvindicated particu-
larities” (ibid., 48). In contrast to Rawls, she argues that the construction must start
from “the least determinate conceptions both of the rationality and of the mutual
independence of agents. A meager and indeterminate view of rationality might
credit agents only with the capacity to understand and follow some form of social
life, and with a commitment to seek some means to any ends (desired or otherwise)
to which they are committed” (O’Neill 1989, 9–10). Moreover, “a meager and
indeterminate view of the identity and of the mutual independence of agents can

5
Here is a neat formulation of this criterion: “No doctrine is admissible as a premise in any
stage of political justification unless it is acceptable to all reasonable citizens, and it need not
be acceptable to anyone else” (Estlund 1998, 254).
6
Interestingly, this was also Richard M. Hare’s charge in his review of A Theory of Justice. Hare
(1989, 145–74) argues that Rawls appeals all too often to intuitions that he never justifies. See
also Brink 1987.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 315

assume only that agents have capacities for varying sorts and degrees of depend-
ence and interdependence” (O’Neill 1988, 10).
My purpose here is not to assess the intramural dispute between Rawls and
O’Neill, but to consider whether a thoroughly constructivist justification is possible.
I will thus focus primarily on O’Neill’s position. The most distinctively “Kantian”
aspect of O’Neill’s theory is her account of practical reason and its authority. Kant’s
objection against foundationalist theories is that they are heteronomous, and
therefore dogmatic. It is important to underscore that the objections of heteronomy
and dogmatism run in parallel. The charge of heteronomy is directed against those
ethical theories, such as intuitionist rationalism, which derive moral obligations
from an external order of value. On this standard rationalist account, the function
of reason is merely recognitional, rather than productive or constructive. Its job is
to recognize truths that are such prior to and independently of reasoning about
them. Practical reasoning is not the agent’s own “doing,” in any significant sense.
Being reasonable means being responsive to reasons that are already there to be
grasped or discovered.
The Kantian objection is that a foundationalist account of justification takes such
moral truths for granted, and it is thus guilty of dogmatism. Dogmatism is not only
an objectionable ethical and epistemic attitude. It is also ineffective in the face of
widespread ethical disagreements. A dogmatic agent is, by definition, insensitive to
reasoning and challenges. But for this very reason his epistemic attitude is inad-
equate in contexts marked by pluralism. The dogmatic rationalist does not engage
with other evaluators holding (radically) different values. He simply dismisses
them as blind. But this arrogant attitude produces a strategy of insulation that is
ultimately self-defeating. Failure to interact with others amounts to a failure to
provide a credible and authoritative justification for one’s claims.7 Dogmatism
allows only for a parochial authority, which can be challenged from any other
ethical perspective. It thus ends up with no authority at all. Genuine authority is
unrestricted. This leads to the claim that genuine authority is warranted by the
requirement of universality: For something to count as a reason of justification, it
must be such that it could be endorsed as a law. The moral law is valid for and
binds all rational agents.8
Kant’s argument in support of this claim is controversial in many respects, but
my aim is to call attention to the fact that the requirement of universality is justified
as an answer to the issue of authority. For instance, O’Neill appropriates Kant’s
argument against dogmatic doctrines when she writes that all the conceptions of
practical reasoning except constructivism fail the “elementary standard of reason”
(O’Neill 2002, 51–2). In O’Neill’s reformulation, to count as reason a consideration
must be “followable” or “accessible” to others (ibid.). Constructivists hold that to
avoid the charge of arbitrariness and impracticality that begets dogmatic rational-
ism, ethical theory must agree with Kant about the universality of the standards of

7
As Korsgaard puts it, “the realist’s belief in the existence of normative entities is not based
on any discovery. It is based on his confidence that his beliefs and desires are normative”
(Korsgaard 1996, 48, see also 64–6, 40ff., 90).
8
See Kant (1911, 425–6, 4:433). Notice that for Kant the purpose of the moral law is for the
rational agent “to appraise himself and his actions from this point of view,” and this leads to
the idea of universal cooperation represented as the kingdom of ends (see Kant 1996, 433,
425–6).

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
316 Carla Bagnoli

rational validity (O’Neill 2003b, 26). Such standards are authoritative only if they
are universally shared. This is a rather strong claim, especially when we consider
its consequences for political legitimacy. In Bernard Williams’s sarcastic words, “the
moral law is more exigent than the law of an actual liberal republic because it
allows no emigration” (Williams 1985, 178). As we shall see in Section 5, it can be
escaped only if rational agency is escapable.

2. Objections from Metaethics


One of the most fundamental challenges to the constructivist solution to the
problem of legitimacy comes from metaethics. Critics argue that constructivists
tacitly (and dogmatically) rely on unconstructed elements which they never justify.
This objection takes different forms. For instance, Joseph Raz expresses this general
concern by considering the standards of correctness of a theory of justice:

To recommend one as a theory of justice for our societies is to recommend it as a just theory
of justice, that is, as a true, or reasonable, or valid theory of justice. If it is argued that what
makes it the theory of justice for us is that it is built on an overlapping consensus and
therefore secures stability and unity, then consensus-based stability and unity are the values
that a theory of justice, for our society, is assumed to depend on. Their achievement—that is,
the fact that endorsing the theory leads to their achievement—makes the theory true, sound,
valid, and so forth. This at least is what such a theory is committed to. There can be no justice
without truth. (Raz 1990, 15)

For Raz liberalism cannot be defended as a merely “political” doctrine, because “[the
rational justification] of moral and political values depends in part on the way they
can be integrated into a comprehensive view of human well-being” (Raz 1990, 23).9
From this realist standpoint, Raz then argues that liberalism should grant the
relevance of truth, and the liberal conception of justice should be proposed as part
and parcel of a particular moral view that concerns the fundamental nature of moral
value. In other words, the constructivist justification is always incomplete and its
completion inevitably invokes moral truths. To undertake the task of completion is
the path from “epistemic abstinence” to “epistemic indulgence” (Larmore 2008). As
it appears, to follow this path amounts to abandoning the distinctive aspiration of
political liberalism to be universally authoritative, unless one is able to point to a
shared comprehensive view of human well-being. The realist is thus implicitly
committed to denying a “reasonable pluralism,” even though he is not forced to
deny the brute fact of pluralism. It is questionable, then, whether this path is open
to someone committed to the defense of reasonable pluralism.
A second formulation of the objection attacks the logical structure of the
constructivist argument. For instance, Cohen (2003, 221) argues that fundamental
normative principles are not grounded on facts and thus they are not as fact-
sensitive as the constructivist account of rational justification requires. Cohen
considers his arguments against Rawls to be a model of his critical strategy against
constructivism in general (see also Raz 2003, cf. Timmons 2003). I take this to be a

9
Similarly, Besch (2004) argues that constructivism can defend its universalist commitments
only if it invokes a perfectionist view of the good. According to Besch, the universalist
strategy of justification is grounded on moral realism.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 317

metaethical statement, even though Cohen is right that his thesis about fact-
insensitivity is neutral with respect to the debate about moral ontology. The
fundamental issue concerns the status of the norms of reasoning, and more
generally the status and role of the starting points of rational justification.10
A third formulation of the objection questions the conception of practical ration-
ality that is at the core of constructivism. For instance, Charles Larmore remarks
that “the essential incompleteness of constructivism,” depends on “the failure to
properly appreciate the truth that reason is responsive to reason” (Larmore 2008,
85, see also 87–167). This diagnosis presupposes the realist view that the function
of practical reason is recognitional, rather than constructive or productive. Reason-
ing amounts to responding to reasons that are already there, prior to and inde-
pendently of our reasoning. As for Raz, for Larmore the constructivist justification
of political liberalism is never complete because it ultimately rests on a moral
principle of respect for persons. This is the pillar of moral thinking that lies beyond
justification.
Cohen’s, Larmore’s, and Raz’s statements are exemplary of a general skepticism
about the capacity of constructivism to stand out as a distinct and internally
coherent metaethical option (see Scanlon 2003, 2012; Shafer-Landau 2003, 42; Enoch
2009; cf. Bagnoli 2011a, 2013). In the following sections, I attempt to show that
constructivism can be defended from this sort of criticism by clarifying the role and
nature of the unconstructed elements of its model of rational justification.

3. The Basis and Materials of Construction


As Rawls writes, “not everything, then, is constructed; we must have some
material, as it were, from which to begin” (Rawls 1993, 104). Remarkably, no
constructivist theory claims to do without unconstructed elements or starting
points (Schwartz 1986). On the contrary, the identification of the correct starting
points is exactly what makes the difference between realist and constructivist
accounts of rational justification. The metaphor of “construction” is supposed to
capture the structure of rational justification in contrast to realist and anti-realist
views. Unlike the anti-realist metaphor of creations ex nihilo, the metaphor of
construction suggests that we need to start somewhere. The basic claim of Kantian
constructivists is that for the justification to be authoritative, the materials of
construction should be empirical facts, such as the considerations about the
circumstances of action. This claim distinguishes constructivism from realism,
which prefers the metaphors of discovery and perception. Reasoning amounts to

10
I agree with Valentini and Ronzoni’s conclusion that Cohen’s fact-insensitivity thesis does
not provide a successful refutation of constructivism, but I do not agree with their argument
(see Ronzoni and Valentini 2008). They argue that Cohen misses his target because he does
not engage with constructivism at the right level. In contrast, I think that Cohen raises the
right sort of concern about the logical status of the starting points of the constructivist
justification. Insofar as such starting points are unconstructed, the question arises whether
constructivism is coherent. A defense of constructivism from this objection requires an account
of the constitutive role of normative principles of construction. This strategy does not commit
constructivism to anti-realism or ontological agnosticism. Rather, purports to identify a third
view between anti-realism/relativism and realism. I thus disagree with Ronzoni and Valentini
that “justified fundamental normative principles are fact-sensitive follows from a commitment
to agnosticism about the existence of objective moral facts.” See also Section 5.

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
318 Carla Bagnoli

discovering moral truths or perceiving moral facts. Moral facts are absolute and
fixed starting points, prior to and independent of any construction. They lay down
the foundation of any further moral conclusion. The realist standards of correctness
are thus material.
By contrast, the standards of correctness for the constructivist justification are
formal. Kantian constructivists qualify the metaphor of construction in terms of a
procedure (Rawls 1993, 116, 239; Street 2010; Bagnoli 2011a). On what grounds is
the procedure selected? Constructivists hold that the procedure is correct not
because it tracks down some independent moral reality, but because it adequately
expresses a shared conception of rational agency. Critics object that the procedure
is either arbitrary or relies on a realist foundation. A convincing reply to the
objection should thus take into account both the basis of the construction and its
procedural constraints.
To be engaged in any sort of rational decision, agents must be free rather than
determined, impeded, or coerced, and they should take themselves to have the
same standing as any other agent. As Rawls emphasizes, this conception of rational
agency marked by freedom and equality works as the “basis of construction.” This
is an important qualification and it should serve as a warning against taking
constructivism to be a mere form of proceduralism, which mechanically transforms
factual inputs and normative outputs.11 The distinctive feature of Kantian construc-
tivism is that the form of justification, which can be represented as a procedure,
expresses the relevant conception of the constructors as rational agents. The
Kantian claim is that the selected procedure is correct not because it tracks down
some moral reality, but because it identifies the practical standpoint from which all
deliberation begins. To this extent, the procedure is respectful of the sort of rational
agents we are. But this is no concession to realism. To identify the practical
standpoint, or to specify the “basis of the construction” is not the same as laying
down foundations. The basis of construction is not analogous to a punctum
Archimedis and its role is not foundational. Rather, to identify a basis of justification
is to say that the issue of justification arises for practical subjects, capable of rational
assessment and sensitive to reason. Unlike the Archimedean point from which the
evaluator assesses the subject of its inquiry at one remove, the practical standpoint
simply identifies an activity, that is, the activity of evaluation. In contrast to realists,
constructivists do not envision the search of objectivity as a search for the
Archimedean point. By contrast, they advocate a conception of objectivity that
results from an ideal agreement among rational agents.
How does constructivism select the relevant conception of rational agency?
Rawls says that the relevant conception of rational agency used in Kantian
constructivism is elicited from common moral experience. It is important to explain
how constructivists avoid the objection that the basis of the construction is
ultimately grounded on arbitrary intuitions. First, the conception of rational agency
is abstract and minimal. Second, it largely coincides with our pre-philosophical
representation of ourselves. The hypothesis that we act under the representation of

11
Korsgaard (2003) emphasizes the practical significance of constructivism by focusing on the
deliberative or first person perspective. For a critique of the proceduralist interpretation, and
an alternative definition in terms of practical standpoint, see Street 2010, 366ff.; cf. Bagnoli
2011a.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 319

ourselves as free and equal is an inescapable mode of self-representation: It


constitutes the basis and the condition of possibility of deliberation. Third, it is a
practical attitude that we take toward ourselves and others. It is neither a hypoth-
esis that belongs to descriptive metaphysics, nor a substantive moral claim. It
neither commits constructivists to take a stand in the metaphysical debate about
free agency, nor does it recommend that people should be treated equally. Rather,
it is a claim about the sort of self-representation that is entailed in exchanging
reasons. If we did not think of ourselves as capable of engaging in and having an
impact on the world, we would not ask how best to act. Aside from this sort of self-
representation, the issue of rational deliberation does not arise. It is the minimal
claim that the deliberative or normative question specifically arises for self-
reflective agents capable of exchanging reasons. This is a relatively unproblematic
claim that Kantians share with other philosophical perspectives.12 As it appears, the
role that the conception of reflective agency plays in the construction is not
foundational. The basis does not identify a moral domain or a special sector of
reality from which moral conclusions should be derived. Rather, it accounts for the
practical rationale of the activity of valuing. Constructivists disagree about what
follows from this characterization of rational agency.13 Clearly, it is open to critics
to present objections against any characterization of rational agency. However,
these will be substantive objections against the specific conceptions of rational
agency. They do not prove the general point that the presence of unconstructed
elements undermines the constructivist nature of rational justification. This general
point remains ungrounded.

4. Status and Scope of Procedural Constraints


In the previous section, I began by stating the difference between constructivism
and realism in terms of standards of correctness. While realists invoke material
standards, constructivists invoke formal standards. For Kantians, there is a formal
requirement that that the justification should take a universal form. Its function is
constitutive. That is, it tells us what counts as a properly formed reason of
justification. This methodological norm is amenable to different formulations, such
as the forms of Kant’s categorical imperative, Korsgaard’s test of reflective endorse-
ment, or O’Neill’s principle of followability. Even though there are important
differences among these formulations, for the sake of the argument I will talk
generically of methodological norms. What matters to the present discussion is that
such methodological norms work as procedural constraints.
In so far as it is formal, a methodological norm is not constructed, but simply
“laid out” (Rawls 2000, 212–4), and thus it stands in need of support. For instance,
Kant argues for the categorical imperative as the only method that grants the
autonomy and authority of reason. As Rawls explains, the categorical imperative
can be formulated procedurally, even though it is misleading to think of it as a mere

12
Cf. Street 2010. To be sure, there are important differences among Kantian and non-Kantian
models of reflective agency. My claim is merely that different kinds of deliberative theories
start with this minimal hypothesis.
13
One relevant source of disagreement concerns what the practical standpoint entails; see
Street 2010; cf. also Bagnoli 2011a.

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
320 Carla Bagnoli

decision-making procedure. More accurately, it is a procedural representation of the


requirements of practical reason. The procedural representation is a useful heuristic
device (Rawls 1971, 21; 1993, 26). It has the advantage of avoiding the complica-
tions of Kant’s transcendental idealism, while pointing out the key Kantian claim,
which is that the objective status of reasons importantly depends on the nature of
rational agents. It is a general feature of Kantian constructivism that it takes the
basic methodological norm to express the relevant conception of the constructors as
free and equal rational agents. While formal, such a norm is sufficient to signifi-
cantly constrain our deliberation.
Interestingly, constructivists are not impressed by the objection of empty formal-
ism that threatens Kantian ethics. This is partly because they neither seek nor
promise a complete normative system. Their most fundamental concern is not
normative determinacy, but authority. That is, they are concerned with finding
answers to practical questions that all relevant others would find binding or
authoritative. This is the reason why constructivism is propounded as a promising
way of addressing the issue of political legitimacy as a deliberative question. The
constructivist starts with qualifying the deliberative contexts marked by value
pluralism. We inhabit a world populated by agents like us in fundamental respects,
holding different values, interdependent and mutually vulnerable. Our audience is
not restricted to the “close circle” of loved ones and kindred spirits, so we need to
produce justifications that all relevant others could understand and follow (O’Neill
2002, 57–8; 2003b, 24–5). Interestingly, this is a principle of intelligibility common
to both the theoretical and practical domains. O’Neill’s requirement of followability
works as a deliberative constraint in that it selects considerations that qualify as
reasons. That is, for something to be called a reason it must be such that all relevant
others could not only entertain the thought of, but also consider it as a real option
for them, whether or not they would eventually choose to adopt it themselves as
a reason (O’Neill 2002, 59).
Since the requirement of followability is also a condition for intelligibility, we
should block a possible source of misunderstanding. In the face of unfollowable
claims, rational agents would not react as if a lion spoke to them but they could not
understand it.14 To judge that a claim is not followable is to say that it does not
withstand proper scrutiny: It is unintelligible as a reason, and therefore not an option
for us as rational agents who are in the business of exchanging reasons. Once the
stringency of the requirement is so qualified, an important source of reservations
against constructivism is dismissed.15
A second source of perplexity about the potentiality of constructivism concerns
the scope of the methodological norm. Critics doubt that that the requirement of
universality is sufficient to secure the sort of agreement that constructivists seek. To
address this objection, it is important to carefully review the structure of construc-
tivist arguments. Its critics often assume that the argument rests on the assumption
that a subject S could not follow p if p derives from assumptions a1–an that S
does not endorse. In this case, the principle of followability would make the

14
Both O’Neill and Korsgaard use this Wittgensteinian analogy, but I think it is misleading
and unnecessarily strong to put the issue this way (cf. Korsgaard 1996, 138–9).
15
Budde (2009, 209–10) attacks O’Neill on the basis of this misunderstanding about the scope
of the intelligibility requirement.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 321

constructivist account of intelligibility depend on actually (and often idiosyncrati-


cally) held beliefs.16 On this reconstruction, the universalist test of the non-
intelligibility of p for S assumes that it is correct for S to judge the followability of
p, by inferring p from S’s set of beliefs ß. If a1–an do not belong to ß, S would not
be in the position to derive p; he thus could not follow p. To follow p, S should be
able to infer or derive p from ß. This explanation assumes that for any subject S to
rationally justify p amounts to inferring p from some set of held assumptions ß.
This reconstruction is misleading, though. The constructivist test is meant to
establish whether p qualifies as a justified ground for a reason. It says that p is not
a justified ground for a reason if it could not (possibly) be accepted, regardless of what
S’s ß contains. The claim is modal. It excludes cases where p is either incoherent
or grounded on some partial assumptions a1–an that not all practical subjects could
endorse. Interestingly, however, the requirement rules out as not followable also
cases where S does actually endorse p because and insofar as p is grounded on S’s
ß. In fact, the requirement excludes not only cases of arbitrariness, but also cases
of special pleading. Providing a rational justification for p does not amount to
deriving p from further assumptions, whether present or not present in S’s ß. The
constructivist claim is that any justification that grounds p on particular assump-
tions a1–an fails to meet the elementary standard of reason, whether or not the
assumptions appear to be S’s own assumptions. What drives constructivists is not
a craving for absolute and determined answers on which all rational agents must
converge. Rather, it is the preoccupation to exclude arbitrary considerations.
Immorality is excluded as a fundamental form of arbitrariness.
The lesson to draw from this discussion is important and it concerns the methods
of ethics. The methodological norm is not a criterion to identify the practical
inferences that a subject S could correctly draw given his deliberative set ß, and
those that S could not correctly draw given his deliberative set ß. Critics often
mistakenly assume that the constructivist account is a form of deduction. But the
constructivist account of justification differs sharply from the deductive form of
practical reasoning that is privileged by foundationalist theorists and often asso-
ciated with Kantian deontology. In fact, constructivists decisively oppose the
deontological interpretation of Kantian ethics as profoundly mistaken (see Herman
1993, 208–43; Richardson 1995).17 The agenda of constructivism is far more radical
than many of its critics assume. It purports to be a genuine alternative to the
classical forms of practical reasoning.

5. Coordination as a Problem, Cooperation as a Solution


In the previous section I argued that some critics mistakenly treat constructivism as
a form of deductivism. In this section I focus on a criticism that mistakenly reduces
constructivism to a form of teleology. According to Budde (2009), O’Neill’s argu-
ment for cooperation exhibits all the features of end-orientation. To be successful,
it must be shown that cooperation is in itself worthwhile, but the constructivist

16
“That people may find this account of reasoning incomprehensible or unconvincing does
not logically entail that it could not be followed” (ibid., 209).
17
Rawls’ Lectures on Kant’s moral philosophy were conceived as the attempt to correct this
mistake (see Rawls 2000).

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
322 Carla Bagnoli

never offers such a demonstration. Budde claims that O’Neill’s failure to provide
a thoroughly constructivist justification of the principles of justice “indicates an
inherent contradiction within constructivism” (ibid., 199). As it stands, construc-
tivism is incomplete, and the necessary completion reveals that it rests on unjus-
tified moral grounds.
Kantian constructivism agrees with the basic tenets of the rationalist tradition
that reason is a successful instrument for coordination. Indeed, it also shares the
rationalist claim that morality is a fundamentally cooperative enterprise. Its dis-
tinctive claim is that coordination is something we have to produce, by endorsing
principles of action that others could share as reasons. For instance, O’Neill takes
the requirement of universality or followability to secure “the necessary conditions
for any coordination, however minimal, by those among whom the reasoning is to
count” (O’Neill 2002, 60). To identify the requirement of universality as a necessary
condition for basic coordination commits to the claim that it is also “the most
fundamental principle of practical reasoning” (ibid.). This requirement constrains
any basic coordinative enterprise, such as action and communication.
Budde remarks that coordination is understood to be cooperation among free and
equal agents. Not just any coordination device will do, but only those cooperative
activities that are grounded on respect and mutual recognition of autonomous
agency. For Budde this shows that cooperation represents a substantive moral end.
The authority of O’Neill’s principles of justice and virtues is thus conditional and
parochial. It is conditional upon the endorsement of the value of autonomous
agency, and it is restricted to the specific communities where this value is endorsed
or enforced. O’Neill’s argument can be rationally rejected by those who do not
share the value of autonomous agency, and thus it does not meet its own criterion
of followability. The constructivist argument is self-defeating.
Budde is right that coordination is not part of the circumstances of action, the
facts of the matter. However, this does not prove yet that coordination serves as the
moral foundation of the constructivist reasoning. In contrast to teleological justi-
fication, constructivists do not conceive of coordination as an end to achieve, but
as a problem to solve (Korsgaard 2003). The difference is subtle but important, and
to adequately assess the prospects of constructivism, it is crucial to take this claim
seriously. Practical reasoning is a resource for addressing practical problems. The
problem of coordination is particularly challenging in contexts marked by plural-
ism, but it is not generated by pluralism. The demand that we act in concert is
pervasive in social worlds, where agents are mutually dependent and reciprocally
vulnerable. In all such contexts the issue of coordination importantly arises. The
peculiar challenges presented by value pluralism is that in contexts marked by
value pluralism, rational agents do not share the same conception of the good, they
do not have the same goals, and even more dramatically, they do not each have a
unified account of value to pursue. Value-pluralism is a significant complication not
only in planning shared or collective actions, but also and more basically in
planning individual action. Like Hobbesian rationalists, Kantian constructivists
hold that this problem is to be solved by rational deliberation because they think
of reason as capable of providing arguments of unrestricted authority. They
contend that reason provides a solution to the problem of coordination. In contrast
to other sorts of rationalists, however, Kantians hold that the solution is the
particular sort of coordination marked by equal respect and mutual recognition that

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 323

is called cooperation.18 Neither coordination nor cooperation is understood as “the


end.” Coordination is the problem to which cooperation is the solution. The
structure of the constructivist argument is thus not end-oriented but problem-
oriented.
The question remains, however, as to whether the constructivist account of
cooperation as a solution to the problem of coordination is convincing. The popular
view is that the constructivist defense of cooperation is not thoroughly construc-
tivist, as it rests on some unconstructed evaluative assumptions, e.g. the value or
dignity of humanity (see Wood 1992; Watkins and Fitzpatrick 2002; Fitzpatrick
2013; Stern 2013; cf. Bagnoli 2013; Sensen 2013). To reply to this objection it is
important to examine the role that the claims about equal respect and recognition
play in the constructivist argument for cooperation.
In Section 2 I argued that the hypothesis about rational agency works as the basis
of construction in the sense that it identifies the practical standpoints from which
the practices of exchanging reasons make sense. It is not a piece of descriptive
metaphysics, but a practical stand. On this minimal construal, rational agency
amounts to the capacity to be guided by values and norms. The relevant conception
of normative guidance includes the capacity to conceive of a norm and act under
its representation. Normative guidance is not limited to acting according to a norm,
as when we follow instructions or orders. Arguably, one can follow and execute
orders without being guided by them. More significantly, rational agency requires
the capacity to act on the basis of some norms, that is, driven by the subjective
understanding of the content of such norms.19
The critic may question whether the commitment to rational agency is one that
holds universally for all. The argument provided by Kantian constructivists is that
commitment to rational agency is constitutive of our being human. It is uncon-
ditional and universal insofar as it is not escapable. In action theory, this claim
is called constitutivism, and is the focus of a growing debate.20 The constructivist
sides with the constitutivist, arguing that one cannot opt out of rational agency.
We can choose to suspend ordinary activities such as taking a walk or reading a
newspaper, but we cannot stop engaging in any form of agency all at once.
During the suspension of ordinary activities some form of rational agency is
nonetheless present. This is because even when we refrain from performing
action, there is a basic sense in which we are still active. The Kantian argument
is that being reflective counts as a form of activity.21 Since our minds are by
constitution self-reflective, some sort of activity is present even when we are not

18
This is a significant disagreement between Kantians and Hobbesians, such as Gauthier
1986, Copp 2007, but I have to set this debate aside.
19
The claim is not that such a capacity should be effectively exercised, so it allows for cases
where rational agents fail to act on their best judgment.
20
Critics object that rational agency is optional and conditional as any other ordinary activity
in which we engage; see, e.g., Enoch 2006. For a convincing reply endorsed by other
constitutivists, see Ferrero 2010.
21
For some prominent accounts of rational agency, reflection is the highest form of activity.
Aristotle thinks that reflection is the highest form mental activity, and qualifies reflexivity as
the godlike characteristics of humans (Aristotle 2000, 1178a9-1178b8). On this reading,
contemplation and action are two distinct modes of activity, to which different sorts of
excellences and styles of the good life correspond, Aristotle (2000, 1177b15-1177b25). By
contrast, Kantians recognize the primacy of action.

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
324 Carla Bagnoli

performing an action. In this basic sense, rational agency is not optional, and to
this extent it is inescapable. It is a constitutive feature of our human condition
insofar as it depends on the self-reflective structure of the mind.
This hypothesis about reflexivity and rational agency, however cashed out in
detail, is not challenged by the fact of pluralism or the claim that values are
incommensurable. On the contrary, the fact of value pluralism gains relevance
exactly because of the hypothesis of rational agency. Values matter because we take
ourselves to be capable of being guided by them. We are worried by values we do
not share because we recognize others to be capable of being guided by ideals we
find dangerous, repugnant, or simply not worthwhile. We sanction morally inde-
cent practices because we recognize others capable of engaging in such practices
responsibly. But who are “we”? Kantians hold that membership of the moral
community is open to anyone capable of exchanging reasons, hence capable of
mutual recognition. This is because they take the practice of exchanging reasons to
be informed by the principle of recognition of mutual standing.
Is this claim about recognition of equal standing such that it undermines the
constructivist nature of the rational justification that constructivists provide? To
answer in the positive, one needs to show that the principle of equal respect and
recognition is necessarily founded or derived from the value of humanity. In the
next section, I will argue for a dialogical construal of this key practical attitude,
which does not assume realism.

6. Equal Respect and Mutual Recognition: The Dialogical Account


Liberal theories accord an important place to respect for persons, but there is a
divisive scholarly disagreement about the notions of “respect” and “recognition”
and their role.22 The question for us is whether respect must play a foundational
role in the justification of liberalism. Critics object that constructivism is ambiguous
in this regard and avoids the important question. Charles Larmore thus presses the
case: “We need to make clear why it is that the validity of coercive principles
should depend upon reasonable agreement” (Rawls 1971, 586). Larmore holds that
“political liberalism, conceived as a strictly political doctrine, rests on this moral
foundation of respect” (Larmore 2008, 149). Larmore (ibid., 139ff.) criticizes Rawls
(and Habermas) for losing sight of the moral values that inspire the liberal project.
To rebut this objection that constructivism is realism in disguise, it is sufficient to
show that respect and mutual recognition play a central role in the constructivist
justification, but not a foundational role. These notions can be construed without
assuming a realist account of value such as the absolute value (or dignity) of
humanity.
In the constructivist model of justification, the qualifiers in “equal respect” and
“mutual recognition” are meant to point to a relation of reciprocity among agents.
The context of discussion is the social world, populated by interdependent and
mutually vulnerable persons, where coordination is a problem. How exactly is
coordination a problem? Let us start with a simple example. Each of the inhabitants
of the world claim goods, and goods are scarce. By claiming their supply of goods

22
For Rawls (1971, 586) this disagreement reveals that respect is not “a suitable basis” for
political liberalism.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 325

in such a situation, each inhabitant makes claims on others that their demand be
satisfied. What are the grounds of such demands? It is important to distinguish
between two levels at which this question arises. First, we should ask how and on
what grounds others are entitled to make these demands. A second, separate issue
is how and on what grounds one agrees to satisfy the requests of others. The first
is an issue that concerns the normative status of persons to make demands. The
second concerns the normative schemes of distributions of goods.
It is easy to equivocate the role that respect plays in these two related but
separate levels of justification. At the first level, the notion of equal respect and
mutual recognition is a practical attitude that demands recognition of equal
standing. Nothing more derives from it, except the requirement that all claims be
heard. At the second level, on the other hand, respect serves as the substantive
ground for a just distribution. This second normative role of respect should be
bracketed when we are assessing the prospects of Kantian constructivism. The
requirement of followability, as well as any other formulation of the universality
requirement, works as a deliberative constraint but it does not directly prescribe
any determined scheme of distribution. The requirement of followability deploys
the notion of respect as a generic recognition of status. Under this capacity, respect
requires that we relate to each other as “self-originating sources of valid claims”
(Rawls 1980, 546).
The crucial question is whether the recognition of equal status is an act that
constructivism can successfully explain without resting the argument on the
absolute value of humanity. According to critics, it cannot. The prospects of such
an argument to be authoritative depend on its completion by an additional
argument proving the value of humanity to be the true one. Interestingly, this is a
position that several Kantians share (cf. Audi 2004; Larmore 2008). These Kantians
side with realists against constructivists because they hold that rational justification
begins with the recognition of a moral fact, namely, the value of humanity.
Accordingly, they take respect to be the practical attitude we owe to others because
it responds to the recognition of a substantive value property. It is the attitude fit to
a specific sort of substantive value. On the basis of this construal of respect, we are
required to respect others because and insofar as others possess a given value
property (e.g., dignity or autonomy of the will). The question arises as to how to
show that others embody some such value.
On the other hand, the constructivist is not under pressure to provide any proof
of the value of humanity. This is because she can take respect as a constitutive
mode of valuing persons as such. I call this account of respect dialogical.23 Respect
is the practical attitude of holding each other accountable, rather than the fitting
response to a value property. It is a normative relation that we entertain with
ourselves and with others, rather than a form of epistemic awareness. This is
simply to say that respect and mutual recognition are part of the basis of
construction, in the sense outlined in Section 2.
Respecting others as sources of valid claims means respecting them as constraints
on our deliberation. This means that respect is the source of procedural constraints.

23
The dialogical account of respect is similar but not identical with Darwall’s account of the
second-person standpoint (Darwall 2006). Darwall takes his proposal to be a variation of the
fitting attitude theory (cf. Bagnoli 2007).

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
326 Carla Bagnoli

However, the substantive implications of respect are undetermined and indirect.


Recognizing others as having equal standing allows us to enter dialogical practices
of exchanging reasons, but it does not force us to actually share or endorse the
content of others’ reasons. Nor does it bind us to specific ways to treat them. The
requirement of universality or followability does not demand homogeneity, nor is
the dialogue directed to produce some sort of convergence on a specific set of
assumptions. The dialogue creates a public arena where moral differences are
recognized as based on reasons that we cannot simply dismiss. We can challenge
such reasons as reasons, that is, dispute their validity, weight, and importance, but
we cannot ignore them just because they are not ours. To respect others means to
grant them equal standing and thus take part in a dialogue with them as our peers.
The dialogical interpretation presents some important advantages over its com-
petitors. First, it has the merit of addressing directly the issue of authority, which
seems to elude the realist interpretation of respect. Realism holds that there are
certain moral properties of facts that make persons worthy of respect. However, it
is problematic for realism to show that it is the possession of this particular moral
property that makes persons authoritative sources of valid claims. Correspondingly,
realism is at a loss to explain how it is that mere awareness of another’s existence
makes evaluators sensitive to their status and responsive to the authority of others
(Nagel 1970).24
Secondly, the dialogical interpretation avoids the ontological and epistemological
issues that arise for realism. For instance, how do we know whether and to what
extent persons possess autonomy of the will, which is the metaphysical basis of the
dignity of humanity? And what is the range of this property? The dialogical view
takes dignity and respect to be a nominal and relational property, namely, a property
of entertaining a normative relation of mutual recognition with others. The relation
with others is neither prior to nor supervenient on any other basic properties. There
are no substantive properties that make anyone worthy of respect or entitled to be
respected. The relation of mutual recognition is exactly what constitutes equal
respect. The epistemology associated with the dialogical view of respect is itself
dialogical. It requires nothing more than open discussion and reciprocal scrutiny.
Third, the dialogical view offers some significant advantages over the view that
the ontological ground of respect for persons is provided by the social practices of
recognition. Social accounts of recognition agree that respect is a relational notion,
but they take it to be determined by actual members of actual communities. Such
views are driven by the conviction that recognition should be cast in terms of actual
practices that involve concrete individuals. They object to the Kantian view of
respect that it is too formalistic and abstract to do justice to individuals, and thus
at risk of ignoring morally relevant differences. In contrast to the Kantian models,
such theories insist that the relevant dialogue takes place in an actual arena. These
proposals adopt material criteria of correctness, and thus they renounce the idea
that rational justification exhibits unconstrained and unconditional authority. To
this extent, they are unfit to furnish the basis of political liberalism. When combined
with a realist view of moral values, liberalism becomes a comprehensive moral

24
I cannot deal with the large meta-ethical debate about the capacity of realism to account
for normativity. My point here is simply that realism faces difficulties in this domain that
constructivism avoids.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 327

doctrine faced with the objection of dogmatism. Furthermore, to link the agenda of
liberalism to that of realism is a dangerous move, considering the serious difficul-
ties that realist metaethical theories encounter. Perhaps these difficulties are not
insurmountable, but realism appears to have no metaethical advantage over
constructivism.
An important challenge to political liberalism is that it depends on an unstable
and doubtful metaethics, namely, constructivism. The objection is that constructiv-
ism is an internally incoherent metaethical doctrine, which ultimately rests on a
realist foundation and is at a loss to establish the universal authority of its
conclusions. What drives this sort of objection is the conviction that there is no
logical space between realism and relativism. My reply is that Kantian construc-
tivism importantly deploys unconstructed elements, but such elements do not play
a foundational role, and thus do not undermine the constructivist structure of its
justification. I thus claim that Kantian constructivism survives the objections from
metaethics. Clearly, this argument in itself is not meant to provide a complete
defense of political liberalism. Nonetheless, I hope to have shown that it is a more
plausible basis for political liberalism than many of its critics have found it; and this
I believe is an achievement of sorts.25

University of Modena and Reggio Emilia


Largo Sant’Eufemia 19
41121 Modena, Italy
E-mail: carla.bagnoli@unimore.it

References
Aristotle. 2000. Nicomachean Ethics. Ed. S. Broadie. Oxford: Oxford University
Press.
Audi, R. 2004. The Good in the Right. Princeton: Princeton University Press.
Bagnoli, C. 2007. Respect and Membership in the Moral Community. Ethical Theory
and Moral Practice 10: 113–28.
Bagnoli, C. 2009. Review of Charles Larmore “The Autonomy of Morality.” The
Philosophical Review 118: 536–40.
Bagnoli, C. 2011a. Constructivism in Meta-Ethics. Stanford Encyclopedia of Philosophy,
Winter Archive: 1–51.
Bagnoli, C. 2011b. Emotions and the Categorical Authority of Morality. In C.
Bagnoli Ed. Morality and the Emotions, Oxford: Oxford University Press, 62–81.
Bagnoli, C. 2013. Constructivism about Practical Knowledge. In Constructivism in
Ethics. Ed. C. Bagnoli, 153–82. Cambridge: Cambridge University Press.
Besch, T. M. 2004. On Practical Constructivism and Reasonableness. Doctoral Thesis.
University of Oxford.
Brink, D. 1987. Rawlsian Constructivism in Moral Theory. Canadian Journal of
Philosophy 17: 71–90.
Budde, K. 2009. Constructivism all the Way Down—Can O’Neill Succeed where
Rawls Failed? Contemporary Political Theory 8: 199–223.
Cohen, G. A. 2003. Facts and Principles. Philosophy and Public Affairs 31: 211–45.

25
I would like to thank Ian Carter, Michele Bocchiola, Elijah Millgram, Valeria Ottonelli, and
Robert Schwartz for comments on previous drafts.

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.
328 Carla Bagnoli

Copp, D. 2007. Morality in a Natural World. Cambridge: Cambridge University Press.


Darwall, S. 2006. The Second Person Standpoint. Cambridge, MA: Harvard University
Press.
Enoch, D. 2006. Agency, Shmagency: Why Normativity Won’t Come from What is
Constitutive of Agency. Philosophical Review 115: 169–98.
Enoch, D. 2009. Can there Be a Global, Interesting, Coherent Constructivism about
Practical Reason? Philosophical Explorations 12: 319–39.
Estlund, D. 1998. The Insularity of the Reasonable: Why Political Liberalism Must
Admit the Truth. Ethics 108: 252–75.
Ferrero, L. 2010. Constitutivism and the Inescapability of Agency. Oxford Studies in
Metaethics 4: 303–33.
Fitzpatrick, W. 2013. How not to be an Ethical Constructivist: A Critique of
Korsgaard’s Neo-Kantian Constitutivism. In Constructivism in Ethics. Ed. C.
Bagnoli, 41–62. Cambridge: Cambridge University Press.
Freeman, S. 2007. The Burdens of Public Justification: Constructivism,
Contractualism, and Publicity. Politics, Philosophy and Economics 6: 5–43.
Gauthier, D. 1986. Morals By Agreement. Oxford: Oxford University Press.
Hare, R. M. 1989. Rawls’ Theory of Justice. In R. M. Hare, Essays in Ethical Theory,
145–74. Oxford: Clarendon Press.
Herman, B. 1993. Leaving Deontology Behind. In B. Herman, The Practice of Moral
Judgment, 208–43. Cambridge, MA: Harvard University Press.
Hill, T. 2008. Moral Construction as a Task: Sources and Limits. Social Philosophy and
Policy 25(1): 214–36.
Kant, I. 1911. Grundlegung zur Metaphysik der Sitten, in Kants gesammelte Schriften.
Preussische Akademie der Wissenschaften. vol. IV. Berlin: Reimer.
Kant, I. 1996. Groundwork of the Metaphysic of Morals, in Practical Philosophy. Ed. A.
Wood. Cambridge: Cambridge University Press.
Korsgaard, C. M. 1996. The Sources of Normativity. Ed. O. O’Neill. Cambridge:
Cambridge University Press.
Korsgaard, C. M. 2003. Realism and Constructivism in Twentieth-Century Moral
Philosophy. The Journal of Philosophical Research, APA Centennial Supplement,
Philosophy in America at the End of the Century 1: 99–122.
Larmore, C. 2008. The Autonomy of Morality. Cambridge: Cambridge University
Press.
Nagel, T. 1970. The Possibility of Altruism. Princeton, NJ: Princeton University Press.
O’Neill, O. 1988. Constructivisms in Ethics, Presidential Address. Proceedings of the
Aristotelian Society 89: 1–22.
O’Neill, O. 1989. Constructions of Reason: Exploration of Kant’s Practical Philosophy.
Cambridge: Cambridge University Press.
O’Neill, O. 1999. Kantian Constructivisms. In Rationalität, Realismus, Revision. Ed. J.
Nida-Rümelin, 3–16. Berlin, New York: De Gruyter.
O’Neill, O. 2002. Towards Justice and Virtue. Cambridge: Cambridge University
Press.
O’Neill, O. 2003a. Constructivism in Rawls and Kant. In The Cambridge Companion
to Rawls. Ed. S. Freeman, 347–67. Cambridge: Cambridge University Press.
O’Neill, O. 2003b. Bounds of Justice. Cambridge: Cambridge University Press.
Rawls, J. 1971. A Theory of Justice. Cambridge, MA: Harvard University Press.
Rawls, J. 1980. Kantian Constructivism in Moral Theory. Journal of Philosophy 77:
515–72.
Rawls, J. 1993. Political Liberalism. New York: Columbia University Press.
Rawls, J. 1999a. Justice as Fairness: Political not Metaphysical. In J. Rawls, Collected
Papers. Ed. S. Freeman, 388–414. Cambridge, MA: Harvard University Press.

© 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd. Ratio Juris, Vol. 27, No. 3
Kantian Constructivism Reassessed 329

Rawls, J. 1999b. Themes in Kant’s Moral Philosophy. In J. Rawls, Collected Papers.


Ed. S. Freeman, 497–528. Cambridge, MA: Harvard University Press.
Rawls, J. 1999c. The Idea of Overlapping Consensus. In J. Rawls, Collected Papers.
Ed. S. Freeman, 421–48. Cambridge, MA: Harvard University Press.
Rawls, J. 2000. Lectures on the History of Moral Philosophy. Ed. B. Herman and C. M.
Korsgaard. Cambridge, MA: Harvard University Press.
Raz, J. 1990. Facing Diversity: The Case of Epistemic Abstinence. Philosophy and
Public Affairs 19: 3–46.
Raz, J. 2003. Numbers, with and without Contractualism. Ratio 16: 346–67.
Richardson, H. 1995. Beyond Good and Right: Toward a Constructive Ethical
Pragmatism. Philosophy and Public Affairs 24: 108–41.
Ronzoni, M., and L. Valentini. 2008. On the Meta-ethical Status of Constructivism.
Politics, Philosophy & Economics 7: 403–22.
Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard
University Press.
Scanlon, T. M. 2003. Metaphysics and Morals. Proceedings and Addresses of the
American Philosophical Association 77: 7–22.
Scanlon, T. M. 2012. The Appeal and Limits of Constructivism. In Lenman, J. and
Shemmer, Y. Constructivism in Practical Philosophy. Oxford: Oxford University
Press.
Schwartz, R. 1986. I’m Going to Make You a Star. Midwest Studies in Philosophy 11:
427–39.
Sensen, O. 2013. Kant’s Constructivism. In Constructivism in Ethics. Ed. C. Bagnoli,
63–81. Cambridge: Cambridge University Press.
Shafer-Landau, R. 2003. Moral Realism. Oxford: Clarendon.
Stern, R. 2013. Moral Skepticism, Constructivism, and the Value of Humanity, In
Constructivism in Ethics. Ed. C. Bagnoli, 21–40. Cambridge: Cambridge Univer-
sity Press.
Street, S. 2010. What is Constructivism in Ethics and Metaethics? Philosophy Compass
5: 363–84.
Tampio, N. 2012. A Defense of Political Constructivism. Contemporary Political
Theory 11(3): 305–23.
Timmons, M. 2003. The Limits of Moral Constructivism. Ratio 16: 391–423.
Watkins, E., and W. Fitzpatrick. 2002. O’Neill and Korsgaard on the Construction
of Normativity. The Journal of Value Inquiry 36: 349–67.
Williams, B. A. O. 1985. Ethics and the Limits of Philosophy. Cambridge: Cambridge
University Press.
Wood, A. 1992. Review of Constructions of Reasons: Explorations of Kant’s
Practical Philosophy, by Onora O’Neill. Philosophical Review 100: 647–50.

Ratio Juris, Vol. 27, No. 3 © 2014 The Author. Ratio Juris © 2014 John Wiley & Sons Ltd.

You might also like