Chapter 1 Introduction - Utrecht University Repository

Document technical information

Format pdf
Size 2.6 MB
First found May 22, 2018

Document content analysis

Category Also themed
Language
English
Type
not defined
Concepts
no text concepts found

Persons

Henri Matisse
Henri Matisse

wikipedia, lookup

Ingrid Robeyns
Ingrid Robeyns

wikipedia, lookup

Organizations

Places

Transcript

Quaestiones Infinitae
Publications of the Department of Philosophy and Religious Studies
Utrecht University
Volume 99
Copyright © 2017 by Sem de Maagt
All rights reserved
ISBN 978-94-6103-059-7
Cover design: Fleur Jongepier & Marlaine Verhelst
Printed by Wöhrmann B.V.
Constructing Morality
Transcendental Arguments in Ethics
De Constructie van Moraal
Transcendentale Argumenten in de Ethiek
(met een samenvatting in het Nederlands)
Proefschrift
ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector
magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor
promoties in het openbaar te verdedigen op vrijdag 27 januari 2017 des middags te 2.30 uur
door
Sem de Maagt
geboren op 12 juli 1987 te Tilburg
Promotoren:
Prof.dr. I.A.M. Robeyns
Prof.dr. M. Düwell
Copromotor:
Dr. R.J.G. Claassen
This work is part of the research programme ‘What Can the Humanities Contribute to Our
Practical Self-Understanding?’ with project number 317-20-010, which is financed by the
Netherlands Organisation for Scientific Research (NWO).
Contents
Acknowledgements
ix
Chapter 1 Introduction: Constructing Moral Objectivity
13
1. Introduction
13
2. Moral Objectivity and Moral Relativism
15
3. Kantian Constructivism about Morality
16
3.1 ‘Constructivism’ in Kantian Constructivism
16
3.2 ‘Kantian’ in Kantian Constructivism (1): Objectivity
17
3.3 ‘Kantian’ in Kantian Constructivism (2): Transcendental Argumentation
18
4. The Goal of the Thesis
21
5. Argumentative Strategy: A Potential Dilemma for Kantian Constructivism
25
5.1 Categoricity
26
5.2 Universality
26
5.3 Substance
27
PART 1: METHODS
31
Chapter 2 Reflective Equilibrium
33
1. Introduction
33
2. The Method of Reflective Equilibrium
35
3. Reflective Equilibrium and Moral Objectivity
41
4. Rescuing Objectivity
45
4.1 Considered Judgements
45
4.2 Scanlon on Considered Judgements
49
4.3 General Reflective Equilibrium
52
5. Sacrificing Objectivity
5.1 The Only Game in Town?
54
54
5.2 What is Wrong with Moral Relativism?
6. Conclusion
Chapter 3 Transcendental Arguments
57
62
65
1. Introduction
65
2. What is a Transcendental Argument?
67
3. Transcendental Arguments in Theoretical Philosophy
74
4. Transcendental Arguments in Ethics
78
4.1 Transcendental Arguments and Moral Realism
78
4.2 Transcendental Arguments and Kantian Constructivism
86
5. Conclusion
PART 2: CATEGORICITY
Chapter 4 General Meta-Ethical Objections to Kantian Constructivism
90
93
95
1. Introduction
95
2. Korsgaard’s Argument for the Value of Humanity
98
3. The Escapability Objection
101
4. The Shmagency Objection
107
4.1 Enoch’s Objection to Kantian Constructivism
107
4.2. Why Only Moral Realists Believe in Shmagents
112
5. The No-Normativity Objection
117
5.1 Street’s No-Normativity Objection to Kantian Constructivism
117
5.2 Categorical Instrumental Reasons
119
6. The Contingency Objection
125
7. Conclusion
131
PART 3: UNIVERSALITY
Chapter 5 Transcendental Arguments from the Second Person
133
135
1. Introduction
135
2. Transcendental Arguments from the Second Person
137
3. Two Objections to Transcendental Arguments from the First Person
141
3.1 The Logical Objection
141
3.2 The Wrong Kind of Reason Objection
144
4. Evaluating Transcendental Arguments from the Second Person
146
4.1 The Argument from Argumentation
146
4.2 The Argument from Communication
153
4.3 The Argument from the Publicity of Reasons
161
4.4 The Argument from the Second Person Standpoint
168
5. What is Wrong with Transcendental Arguments from the Second Person?
178
6. Conclusion
180
Chapter 6 Transcendental Arguments from the First Person I:
The Universal Value of Humanity
183
1. Introduction
183
2. Korsgaard’s Argument from the Sufficiency of Agency
185
3. The Argument for the (Universal) Value of Humanity
189
3.1. Which Humanity? What Value? Which Argument?
189
3.2. The Regress of Identities Argument
194
3.3. The Source of Reasons Argument
199
4. Conclusion
210
Chapter 7 Transcendental Arguments from the First Person II:
The Principle of Generic Consistency
211
1. Introduction
211
2. Gewirth’s Argument from Purposive Agency
213
3. Part 1: Rights to the Generic Features of Agency
219
4. Part 2: Universal Rights to the Generic Features of Agency
225
5. The Logical Objection Revisited
233
6. The Wrong Kind of Reason Objection Revisited
236
7. Conclusion
239
PART 4: SUBSTANCE
241
Chapter 8 The Emptiness Objection and the Appeal to Anthropology
243
1. Introduction
243
2. The Structure of Justification of Kantian Constructivism
245
3. The Recurring Emptiness Objection to Kantian Ethics
250
4. The Emptiness Dilemma
255
4.1. First Horn: No Moral Substance
255
4.2. Second Horn: Unvindicated Premises or Material from Outside
257
5. Kantian Constructivism and Anthropology
260
5.1. The Appeal to Anthropology
261
5.2. Anthropology as Material from Inside
264
6. Conclusion
Chapter 9 The Emptiness Objection and Kantian Constructivism
268
269
1. Introduction
269
2. Which Conception of Agency?
270
3. Onora O’Neill’s Justification of Substantive Morality
275
3.1 The Principle of Followability
275
3.2 The Rejection of a Principle of Severe Injury
278
3.3 Injury, Agency and the Emptiness Objection
282
4. Alan Gewirth’s Generic Features of Agency
288
4.1 The Application of the Principle of Generic Consistency
289
4.2 Identifying the Generic Features of Agency
294
4.3 Additive Goods and Dispositional Freedom
295
4.4 Specifying Agency 1: Successful Agency
300
4.5 Specifying Agency 2: Prospective Agency
302
4.6 Phenomenological Inescapability and Specifications of Agency
308
4.7 The Generic Features of Prospective Agency
313
5. The Emptiness Objection Revisited
315
6. Conclusion
317
Chapter 10 General Conclusions
319
1. Summary
319
2. Confronting my Interlocutors
321
Bibliography
327
Samenvatting
343
Curriculum Vitae
349
Acknowledgements
Writing this thesis would not have been possible without the help and support of many
people.
I would like to begin by thanking my supervisors Ingrid Robeyns, Marcus Düwell and
Rutger Claassen. I am very grateful to Ingrid for taking me on as a PhD student, for the
detailed comments on the chapters of my thesis, and for the continuing support of my
project. Although I am pretty sure that Ingrid would have never expected the original PhD
project to result in a thesis on Kantian approaches to the foundations of morality, she has
always encouraged and supported me to find my own way in philosophy – even though my
way in philosophy turned out to diverge quite significantly from hers. Ingrid has also given
me the basic confidence and an enormous amount of advice that is necessary to develop as
a philosopher and as an academic in general.
Marcus became my official supervisor during the final phase of the project, although he
always played an important role in the background. Marcus was the second reader of my
research master thesis (when I was still a ‘hesitant Rawlsian’) and the principal investigator
of the Horizon research project of which this PhD project was a part. Marcus encouraged
me to look beyond what is fashionable in mainstream ethics and to keep the big picture in
mind. He taught me to detect, in Marcus’s technical terms, philosophical ‘bullshit’.
Marcus’s expertise on transcendental argumentation was indispensible for this project.
This thesis would not have been what it is now without the many comments,
suggestions and challenging questions put forward by Rutger. And maybe just as
importantly, writing this thesis would have been much less fun without Rutger. He has
continually challenged me to improve this thesis, both in terms of specific arguments and
in terms of its general structure. He always did this in an extremely good-humoured and
constructive way. He encouraged me when I needed it most. Rutger’s intense supervision of
the thesis went far beyond any reasonable expectation, and I consider myself lucky to have
been supervised by him.
ACKNOWLEDGEMENTS
x
I would like to thank the Manchester Centre for Political Theory and the Political Theory
Group of the London School of Economics. I am thankful to these two research groups for
hosting me as a visiting PhD student, and for the possibility to present my work. I am
grateful to all the members of these research groups for their comments and suggestions.
Special thanks are due to my two ‘supervisors abroad’, Miriam Ronzoni (Manchester)
and Katrin Flikschuh (LSE), for their extremely warm welcomes and for the time they took
to discuss my work. During my stay in Manchester in 2014, the thesis was finally starting to
take shape, and Miriam commented on the earliest versions of some of the chapters in this
thesis. During my time in London in 2015, Katrin commented on several chapters of this
thesis and we had several stimulating discussions about the project as a whole. These
discussions are amongst the most inspiring and fun ones during my time as a PhD
researcher, and her influence goes far beyond the pages of this book.
I would also like to thank Onora O’Neill. Despite her extremely busy schedule, Onora
made the time to discuss my work with her on several occasions while I was in London.
Although I have been relatively critical of Onora’s work in this thesis, I hope that it is at the
same time evident that I have only been critical, and could only have been critical, because
there is so much of value in her work. Her work is one of my main sources of inspiration in
philosophy. It was therefore an honour to be able to work with her.
This PhD project was part of a NWO-funded research project on the question ‘What Can
the Humanities Contribute to Our Practical Self-Understanding?’. I would like to thank all
the members of this research group for the discussions we had over the years and for the
feedback I received from them on several occasions. I would specifically like to thank Wout
Cornelissen, Caroline Harnacke, Annemarie Kalis, Annemarie van Stee and Naomi van
Steenbergen.
A special word of gratitude goes out to my ‘paranymphs’, Caroline Harnacke and Jurriën
Hamer, and my ‘unofficial,’ third paranymph Esther Keymolen. Caroline, Jurrien and
Esther are not just three of my favourite colleagues but, more importantly, good friends as
well. They supported me academically and otherwise when I needed it most, and they
reminded me of the fact that there is more to life than philosophy.
I am much indebted to my colleagues at the Department of Philosophy and Religious
Studies at Utrecht University, and the members of the Practical Philosophy Colloquium in
ACKNOWLEDGEMENTS
xi
particular. Special thanks to Joel Anderson, Deryck Beyleveld, Gerhard Bos, Morten
Byskov, Dascha Düring, Henk van Gils, Marie Göbel, Julia Hermann, Wouter Kalf, Michael
Klenk, Jos Philips, Stephen Riley, Peter Sperber and Joeri Witteveen. I would also like to
thank, more generally, my fellow PhD students and postdocs at the infamous attic. The attic
is a perfect environment to write a PhD thesis and the daily lunches and coffees where a
perfect distraction from work. I also want to express my sincere gratitude to the support
staff, specifically, Biene Meijerman and Suzanne van Vliet.
I would also like to thank my former colleagues at the Faculty of Philosophy of the Erasmus
University Rotterdam. I started writing this thesis in Rotterdam (where Ingrid was based at
that moment – she moved to Utrecht almost halfway through my PhD project and I was
able to transfer with her). Special thanks to the members of the Practical Philosophy
Research Group and the members of the ‘Justice Meetings’ and specifically to Constanze
Binder, Patrick Delaere, Conrad Heilmann, Gijs van Oenen and Maureen Sie for their
comments on my work in the early stages of my PhD project.
I could not have written this thesis without the support of my family and my family-in-law
(to be). I know that my parents are proud of me; I can only hope that my brother reads
more pages of my thesis than of Jürgen Habermas’s The Theory of Communicative Action
which functions as one of his bookshelf fillers; and that my sister will one day sing me a
song about transcendental arguments.
Finally, I would like to thank Fleur. Saying anything that would come even close to
describing the role that Fleur played in writing this thesis would require more pages than
the thesis itself.
Chapter 1
Introduction: Constructing Moral
Objectivity
1. Introduction
Should I buy regular coffee beans or should I choose the fair trade option? What are my
obligations towards refugees? To what extent is socio-economic inequality justifiable? Do I
have any obligations towards future generations?
We constantly have to come up with answers to these kinds of practical questions, i.e.
questions about how we ought to act. We do this implicitly when, for instance, we go to the
supermarket for groceries. We do this explicitly when, for instance, we discuss the daily
newspaper, make up our minds about the upcoming elections or make career choices.
Reflecting on how we ought to act is part and parcel of our everyday life.
There are different ways to evaluate actions: one could ask whether an action is
prudent, polite or legal, but one could also ask whether an action is moral. These standards
do not necessarily overlap. It might be prudent for me to lie to you about the state of the
coffee machine that I am trying to sell you, but that does not mean that it is the morally
right thing to do. In some societies it is illegal to engage in homosexual relationships, but
that does not mean that these relationships are immoral. The focus of this thesis is
specifically on the moral evaluation of action, and it tries to answer one of the most
fundamental questions of ethics: what could possibly justify a moral claim, such as a
concrete moral judgement (e.g. “I ought not to lie about the poor quality of the coffee
machine I am trying to sell you”), a moral norm (e.g. “lying is immoral”) or a moral
CHAPTER 1
14
principle (e.g. the categorical imperative).1
If I claim that there is a coffee machine on my kitchen worktop, we have a rather good
idea of what makes this claim correct and how to justify this claim. The claim that there is a
coffee machine on my kitchen worktop is made correct by there actually being a coffee
machine on my worktop. I can know this by seeing that there is a coffee machine on my
worktop. The source of the validity of empirical claims is thus the existence of empirical
facts such as coffee machines. The normal way through which we come to know these facts,
it seems, is by perception.
But what is it that makes moral claims correct? Are there moral facts that are analogous
to empirical facts, such as the existence of coffee machines? Or should the nature of
morality be understood in a different way? And how can we come to know whether a moral
claim is justified? Can moral requirements be perceived or should moral requirements be
justified differently? The answers to these kinds of questions obviously depend on how one
defines the moral domain (cf. Dorsey 2016). For instance, if one thinks that moral claims
are nothing more than the expression of a personal opinion or feeling, the justification of a
moral claim will be very different from when one thinks that a moral claim has a certain
objective validity.
The goal of this thesis is to explore the question of whether moral objectivity can be
vindicated, that is, whether there are any objective moral principles, norms or judgements.
More specifically, I explore and defend a Kantian constructivist affirmative answer to this
question, according to which moral claims ultimately follow from one’s self-understanding
as an agent. 2 This will need some spelling out. Therefore, in the remainder of this
introduction, I will set myself the task of explaining, first, why vindicating moral objectivity
should be our goal in the first place, rather than accepting some sort of moral relativism;
second, clarifying what would make the current project a version of Kantian constructivism
and explaining the relevance of this project by situating Kantian constructivism in the
(meta-)ethical landscape; and finally, providing an outline of this thesis.
1
As I explain below, for reasons of space the focus of this thesis is primarily on principles and norms.
Let me stress from the beginning that in this thesis I sidestep the important, exegetical question of how Kantian
constructivism exactly relates to Kant’s own views. Although Kantian constructivism obviously takes its cue from
Kant’s view on ethics, the aim of the thesis is first and foremost systematic, exploring the plausibility of a
constructivist justification of morality, not an exegetical one. References to Kant are limited to points at which it is
relevant to understand current positions or discussions.
2
INTRODUCTION
15
2. Moral Objectivity and Moral Relativism
Why should one be interested in moral objectivity in the first place? Why should one not
understand moral claims merely as expressions of more or less contingent preferences,
desires or social norms? Could it not be that the objectivity of morality is, to borrow
Immanuel Kant’s famous phrase, nothing more than just a “chimerical idea” (Kant 1998,
51;445)? This is the position taken by various strands of moral relativism, including
postmodernism and post-structuralism, but also by meta-ethical theories such as
Humeanism and expressivism. Richard Rorty, for instance, is famous for having said, “I do
not think there are any plain moral facts out there in the world, nor any truths independent
of language, nor any neutral ground on which to stand and argue that either torture or
kindness are preferable to the other” (Rorty 1989, 173).
Objectivists and relativists about morality can agree that many standards of action,
such as rules of etiquette, are only valid relative to the specific context in which they are
accepted. What they disagree about is whether there are any practical standards which are
objectively valid, whatever objectivity precisely amounts to in the context of ethics (I will
come back to this below). A relativist denies that there are any such standards, whereas an
objectivist thinks that these standards exist. This of course still leaves open what the range
of (objective) moral standards is.
Moral relativism denies that there is such a thing as objectivity in ethics, as is, for
instance, illustrated by the quote from Rorty, and therefore crucially hinges on the failure of
accounts of moral objectivity.3 Understood as such, moral relativism is first and foremost a
negative position. What is crucial in this context is thus that the best or only argument for
moral relativism is an argument against moral objectivity. This does not mean that the
burden of proof is on the moral relativist to show that accounts of moral objectivity fail.
Instead, the point is that even moral relativists should be interested in the question of
whether there is a plausible account of the nature and justification of objective moral
claims. That is, they should be interested in investigating the strongest possible account of
moral objectivity before they conclude that moral requirements are only valid relative to a
contingent standard and that there is no universally valid “ground on which to stand and
argue that either torture or kindness are preferable to the other” (Rorty 1989, 173).
Also, some versions of moral relativism are likely to have important and potentially
3
I elaborate on this point in chapter 2.
CHAPTER 1
16
troubling practical implications.4 If moral relativism is true, we might, for instance, have to
conclude that slavery is only wrong from our perspective, that Hitler was right from his
perspective (also known as the “Argumentum ad Nazium” (Hocutt 2000, 297)) or that
human rights are nothing more than a contingent, modern Western construction. Of
course, it might well be that we ultimately have to accept these troubling conclusions,
because there is no convincing account of moral objectivity. But before we accept them we
should be sure that there really is no plausible account of moral objectivity.
This thesis thus has as its provisional starting point the idea that one of the
distinguishing features of morality, as opposed to other practical domains, is that moral
standards are objective. By this I mean that moral standards hold categorically, i.e. that they
are valid independent of one’s contingent beliefs, preferences or desires, and that they hold
universally, i.e. that their scope includes all human beings. This definition of moral
objectivity in terms of categoricity and universality is meant to deliberately leave open the
metaphysical question of whether a moral claim is valid in this way because there is some
kind of moral fact to which the moral claim corresponds (analogous to empirical facts) or
whether moral principles might be objective even if there are no such moral facts. The
former view would be an ontological or realist conception of moral objectivity, and the
latter view would be a practical or constructivist conception of objectivity.
3. Kantian Constructivism about Morality
Kantian constructivism can be used to refer to a wide range of positions, so let me explain
here what kind of position I have in mind.
3.1 ‘Constructivism’ in Kantian Constructivism
A moral realist thinks of moral facts as analogous to my coffee machine or, in other words,
a moral realist thinks that moral facts refer to so-called mind-independent moral facts in
order to justify a moral claim. Moral realists have different ideas about what constitutes a
moral ‘fact’, but what they share is the idea that something is the morally right thing to do if
4
I say ‘some versions’ because arguably recent versions of moral relativism are not committed to these kinds of
conclusions. I will come back to this in chapter 2.
INTRODUCTION
17
it corresponds to a fact of the matter that somehow exists independent of us.5 Moral realists
thus can be said to take a so-called impersonal or third-personal approach to morality. The
Kantian constructivism that I aim to defend is, by contrast, a constructivist view. A
constructivist claims that the validity of moral claims is somehow dependent on us, i.e. the
question of what makes moral claims true or correct is in an important sense minddependent. According to constructivists, something is the morally right thing to do if it
follows from our self-understanding.
This notion of ‘self-understanding’ can be understood in different ways, as will become
clear in the thesis, but on the most general level it refers to the idea that the starting point
for developing a moral theory is an internal, first- or second-person perspective, and not a
third-person perspective (pace moral realism). Constructivism locates morality in the
practical commitments that follow, for instance, from understanding oneself as an agent or
from engaging in certain kinds of interaction, such as communication or argumentation.
3.2 ‘Kantian’ in Kantian Constructivism (1): Objectivity
The proposed account is Kantian in two senses. First, it is Kantian insofar as it holds that
although the validity of moral claims is mind-dependent, morality is nevertheless objective.
This can be contrasted with, for instance, Humean constructivism, according to which “the
substantive content of an agent’s normative reasons is a function of his or her particular,
contingently given, evaluative starting points” (Street 2012, p. 41). Humean constructivism
starts from the contingent, substantive evaluations of an individual and subsequently claims
that the validity of normative claims is relative to these substantive starting points. Kantian
constructivism, on the other hand, starts from an inescapable aspect of our selfunderstanding. This means that it starts from the formal structure of agency and from
interaction as such, not from the particular, substantive purposes or reasons of an agent or
5
When I mention moral realism, I mainly refer to non-naturalist moral realism, i.e. the position that there are
moral facts which cannot be reduced to natural facts. Recent proponents of non-naturalist moral realism include
Schafer-Landau (2003) and Enoch (2011b). Some authors claim that constructivism is not, or at least not in any
interesting sense, distinguishable from naturalist realism, (see e.g. Copp 2013). I do not want to take a stance on
the classificatory question of whether Kantian constructivism is truly an alternative to various forms of naturalist
moral realism. Note, however, that Kantian constructivism is not just an alternative to (non-naturalist) moral
realism and discussions about Kantian constructivism should therefore not be reduced to the question of the
extent to which it offers an alternative to various forms of moral realism. I prefer to understand Kantian
constructivism as a comprehensive approach to moral justification which might be reconstructed as having certain
metaphysical and epistemological elements.
CHAPTER 1
18
from particular forms of interaction. Subsequently, Kantian constructivism claims that
certain normative claims follow from agency or interaction as such. The resulting norms
are thus taken to be categorically valid insofar as agency or interaction is inescapable (the
second feature of objectivity, the universality of these claims, will be discussed below).
3.3 ‘Kantian’ in Kantian Constructivism (2): Transcendental Argumentation
Second, Kantian constructivism is Kantian because it uses a so-called transcendental
argument in order to justify moral claims.6 Transcendental argumentation is the method
through which Kantian constructivism tries to argue from a formal starting point to a
categorical and universal principle of morality (and subsequently to substantive norms of
action). It is difficult to give a concise definition of what makes an argument a
transcendental argument. A substantial part of the thesis will be concerned with exactly this
question. For a start, a transcendental argument can formally be understood as an
argument that tries to show that commitment Y is a necessary condition of the possibility
of X – where, given that X is somehow inescapable for person A, A must necessarily accept
Y. More specifically, a transcendental argument works by asking the question of what is
necessarily presupposed by understanding oneself as, for instance, a believer (in theoretical
philosophy) or an agent (in ethics). A transcendental argument is thus a self-reflexive
argument, and the point of the argument is to show that anyone who understands herself in
a certain way is necessarily committed to whatever are the necessary conditions of the
possibility of this self-understanding.
Different transcendental arguments in ethics take a different ‘X’ as their (alleged)
inescapable starting point. An important difference is that some try to argue from the firstperson perspective, whereas others try to argue from the second-person perspective or to
defend a hybrid view combining first-personal and second-personal elements. Alan
Gewirth and, on one possible reading of her work, Christine Korsgaard have attempted to
justify objective moral principles in terms of the necessary preconditions of agency, more
specifically in terms of the self-understanding of someone who understands herself as
acting purposively (acting so as to achieve a purpose) and voluntarily (chosen in an
unforced way). This is a transcendental argument from the first-person perspective (referred
6
There have also been attempts to use a transcendental argument to justify a moral realist conclusion (Illies 2003;
Enoch 2011b). In chapter 3, I will argue that these attempts fail and that transcendental arguments can only lead to
constructivist conclusions.
INTRODUCTION
19
to from now on as a transcendental argument from the first person), because it tries to
justify a moral principle on the basis of the self-understanding of an individual agent. Of
course, much hinges on how exactly agency is defined, and this will be one of the things I
focus on in the evaluation of this kind of argument.
Karl-Otto Apel, Jürgen Habermas, Onora O’Neill and Stephen Darwall have all, in
slightly different ways, attempted to justify objective moral principles on the basis of the
necessary preconditions of communication, argumentation or reasoning or on what
Darwall calls the ‘second-person’ standpoint. These authors have put forward
transcendental arguments from the second-person perspective (referred to from now on as
transcendental arguments from the second person), because they take as their starting point
certain forms of interaction, i.e. the relation between persons as opposed to individual selfreflection. 7 Again, much hinges on how these different kinds of interactions are
understood, and this will also be one of the points I focus on in the discussion about these
authors.
As will become clear later in this thesis, Christine Korsgaard, on another possible
reading of her work, can be understood as putting forward a hybrid transcendental
argument which combines elements from both transcendental arguments from the first
person and transcendental arguments from the second person.
One way to get a better understanding of transcendental argumentation is by
contrasting transcendental arguments with other methods of moral justification, such as
the method of reflective equilibrium and intuitionism. 8 The method of reflective
equilibrium was first introduced in ethics by John Rawls, and its main idea is that
“justification is a matter of the mutual support of many considerations, of everything fitting
together into one coherent view” (Rawls 1999b, 507).9 According to the method of reflective
7
I use the label second person in a rather liberal way to refer both to theories which state that there is an
irreducible second-person perspective (Darwall) and to theories which locate moral normativity in the relation
between people without making further claims about social ontology (e.g. O’Neill). Of course, there are differences
that are just as important between different arguments from the second person as the differences between
arguments from the first person and arguments from the second person (I come back to this in chapter 5). But I
still think that this is a useful, although admittedly rough, way to distinguish between different transcendental
arguments in ethics.
8
The point here is not to provide an exhaustive overview of different methods of moral justification. I focus on
reflective equilibrium and intuitionism because these methods are arguably the most dominant methods of moral
justification.
9
Although it is true that Rawls mainly discusses the method of reflective equilibrium in the context of normative
political theory, he introduced reflective equilibrium as a general methodology in ethics as an alternative to
intuitionism (see also chapter 2).
CHAPTER 1
20
equilibrium, when we are confronted with a moral question, we have to take into account
all the considerations that are relevant to the issue at hand, and consequently we should try
to make all these considerations coherent by working ‘back and forth’ between different
judgements, beliefs and theories. Reflective equilibrium is thus a form of coherentism in the
sense that there is no moral claim that has a privileged epistemic status independent of its
coherence with other relevant considerations.
Intuitionism is, very roughly, the view that fundamental moral claims can be justified
non-inferentially, meaning that no further reasons can be or have to be given for these
fundamental moral claims. Instead, these claims are considered to be somehow self-evident.
Intuitionism is thus a form of epistemological foundationalism. An intuitionist might, for
instance, claim that the moral judgement ‘killing for fun is immoral’ is simply self-evidently
true and that it cannot, and need not, be further justified. On the intuitionist view, a moral
claim is thus either itself self-evident or, if it is not, it can be justified by reference to a moral
claim that is self-evident.
Transcendental arguments can neither be reduced to the coherentism of reflective
equilibrium nor to foundationalism, or at least in the way these approaches are standardly
understood in moral epistemology. Transcendental arguments are not foundationalist,
because, unlike intuitionism, they do not rely on a self-evident, i.e. non-inferentially
justified, moral claim. Instead, the starting point is an aspect of our self-understanding that
is independent of the contingent content of our specific beliefs, desires or preferences. A
transcendental argument takes as its starting point not the content of a specific moral belief,
but a claim about our self-understanding, i.e. the internal perspective of someone having
any belief, desire or reason whatsoever. Pace intuitionism, this starting point is thus neither
substantive nor moral.
Although transcendental arguments might be understood as a form of coherentism, it
is important to note that they diverge significantly from the specific type of coherentism of
reflective equilibrium. Reflective equilibrium focuses on the coherence of our contingently
held belief set, including, for instance, judgements, principles and background theories. A
transcendental argument, on the other hand, focuses on the principles that one cannot deny
on pain of contradicting that one is an agent or on pain of contradicting that one is engaged
in certain forms of interaction. A transcendental argument thus focuses on the coherency
between our self-understanding and the particular reasons or beliefs of an individual. The
specific criterion of coherence in transcendental arguments is the absence of selfcontradiction, and not coherence simpliciter.
INTRODUCTION
21
In short, I use the label Kantian constructivism to refer to any position that (1) aims to
justify moral objectivity (2) without relying on the existence of moral facts through (3) an
appeal to transcendental argumentation. I think that the success of Kantian constructivism
ultimately hinges on the plausibility of the specific transcendental arguments it offers. After
all, the main challenge for Kantian constructivism is to show if and how certain objective
moral claims follow from an inescapable aspect of one’s self-understanding. Transcendental
argumentation is therefore the main focus of the thesis.
4. The Goal of the Thesis
The main goal of this thesis is to defend a Kantian constructivist account of moral
objectivity. Having introduced the distinction between realism and constructivism, Kantian
and Humean constructivism, and transcendental argumentation as an alternative to other
methods of moral justification, I am now in a position to be more concrete about the goal
of this thesis. I will do so by describing five more specific goals.
The first goal is to show that it is possible to provide an account of moral objectivity
without relying on the existence of moral facts. Providing an account of moral objectivity
without relying on metaphysically suspect moral facts has been called the “The Holy Grail”
of secular ethics (Street 2016, 165). Although there is a growing literature on Kantian
constructivism, this literature has been largely critical of the Kantian constructivist project
(recent critics of Kantian constructivism include, for instance, Enoch 2006; Enoch 2011a;
Street 2012; Tiffany 2012). The first goal of this thesis is to engage with this critical
literature and to argue that, at least on one interpretation, Kantian constructivism can
justify objective morality.
Let me stress that with respect to moral realism, the goal of this thesis is not so much to
criticize moral realism, but to constructively show that there is an alternative approach to
moral objectivity. Philosophers interested in the metaphysics of morality typically believe
that the source of moral objectivity should be understood along moral realist lines.10 David
Enoch, for instance, claims that “like many other realists (I suspect), I pretheoretically feel
10
I do not want to suggest that all or most meta-ethicists are moral realists. Instead, my point is that typically those
who believe in moral objectivity are moral realists. Of course, there are many anti-realists, but this position is
typically associated with forms of subjectivism or relativism. The standard view in meta-ethics is that one is either
a moral realist and believes in moral objectivity or one is an anti-realist and is therefore committed to some form
of subjectivism or relativism.
CHAPTER 1
22
that nothing short of a fairly strong metaethical realism will vindicate our taking morality—
or perhaps normativity more generally—seriously” (Enoch 2010, 111). Interestingly, even
(some of) those who are critical of the Kantian constructivist project seem to agree that the
attractions of a Kantian constructivist conception of objectivity are significant. Street, for
instance, writes that “were a Kantian constructivist argument to succeed, metaethical
constructivism could boast to have captured and vindicated the intuitions of robust moral
objectivity that otherwise might tempt many theorists to go realist rather than
constructivist” (Street 2012, 42). I take a similar stance towards intuitionism. That is, I do
not directly argue against moral intuitionism, but I assume that intuitionism has certain
problems and subsequently I analyse whether there is an alternative method to justify
objective moral claims.11
The dialectical position with respect to the method of reflective equilibrium is different.
The second goal of this thesis is to criticize the dominance of the method of reflective
equilibrium and to assess the plausibility of an alternative, transcendental approach to
moral justification in ethics. Philosophers interested in moral epistemology or moral
justification typically believe that (objective) moral claims should be justified through the
method of reflective equilibrium. In many areas of ethics, reflective equilibrium is simply
considered to be “the only game in town” (Kelly and McGrath 2010, 326). 12 Thomas
Scanlon even claimed recently that reflective equilibrium is “the only defensible method:
apparent alternatives to it are illusory” (Scanlon 2003, 149). In addition, the general
assumption among proponents of reflective equilibrium is that it can safeguard moral
objectivity without relying on self-evident premises. 13 I argue, however, that reflective
equilibrium cannot safeguard moral objectivity (or at least if it wants to remain true to its
fundamental methodological commitments). I argue that this implies that proponents of
11
Cf. Rawls, who concludes his discussion of Kantian constructivism by saying: “I should stress ... that for all I
have said it is still open to the rational intuitionist to reply that I have not shown that rational intuitionism is false
or that it is not a possible basis for the necessary agreement in our judgments of justice. It has been my intention to
describe constructivism by contrast and not to defend it, much less to argue that rational intuitionism is mistaken.
In any case, Kantian constructivism, as I would state it, aims to establish only that the rational intuitionist notion
of objectivity is unnecessary for objectivity” (Rawls 1980, 570). In recent years, there has been a resurgence of
interest in intuitionist views (see e.g. Shafer-Landau 2003; Dancy 2004; Audi 2005; Enoch 2011b). Unfortunately, I
cannot discuss these recent defences of intuitionism in this thesis.
12
Of course, there are many exceptions to this rule. One prominent exception in normative political theory is G.A.
Cohen, who adopts an intuitionist methodology (2008, 4). See also Peter Singer’s intuitionist approach to (applied)
ethics (Singer 1993; de Lazari-Radek and Singer 2014). For transcendental argumentation in applied ethics see
Beyleveld and Brownsword (2001) and Düwell (2013). For an overview of virtue ethical approaches to normative
ethics and applied ethics see Walker and Ivanhoe (2007).
13
Or at least, this is what I argue in chapter 2.
INTRODUCTION
23
reflective equilibrium either have to take alternative methods of moral justification, and
more specifically transcendental arguments, more seriously than they have done so far, or
accept that the method of reflective equilibrium implies moral relativism. The relevance of
this project is to question the current dominance of the method of reflective equilibrium in
normative and applied ethics.
Let me briefly say something about the relation between the first two goals of this
thesis, the goals of providing an alternative to moral realism (and intuitionism) and of
providing an alternative to the method of reflective equilibrium. Traditionally, moral
realists have adopted some form of intuitionism as their moral epistemology. The method
of reflective equilibrium is introduced in ethics as an alternative to intuitionism. The
relation between reflective equilibrium and moral realism is, however, complex. On the one
hand, moral realists increasingly rely on the method of reflective equilibrium as their
preferred moral epistemology (Holmgren 1987; Brink 1989; Scanlon 2014; for a critical
discussion of the relation between reflective equilibrium and moral realism see Kelly and
McGrath 2010). These reflective-equilibrium–moral realists believe that the coherence of all
relevant considerations is an indication of a realist, moral truth. Others, however, believe
that the outcome of the process of reaching a reflective equilibrium is objective without
claiming that the outcome corresponds to mind-independent moral facts, i.e. that there is
room for reflective equilibrium without moral realism. John Rawls, for instance, is a
reflective-equilibrium constructivist (but not a Kantian constructivist on my definition of
the term), according to whom “moral objectivity is to be understood in terms of a suitably
constructed social point of view that all can accept. Apart from the procedure of
constructing principles of justice, there are no moral facts” (Rawls 1980, 519). Moral
realism is thus compatible with different methods of justification or different accounts of
moral epistemology. Reflective equilibrium is compatible with different views about the
source of moral objectivity or moral metaphysics.
However, the orthodoxy in ethics is that when it comes to the source of moral
objectivity, the standard view assumes that we need to rely on moral realism, leaving open
which moral epistemology to accept (intuitionism or reflective equilibrium). When it
comes to the method of justification of objective moral claims, the orthodox view is to turn
to the method of reflective equilibrium, leaving open the question of which moral
metaphysics should go along with reflective equilibrium. Kantian constructivism provides
an alternative to this orthodox view.
Let me now turn to the third goal of the thesis. Although several transcendental
CHAPTER 1
24
arguments have recently been put forward in ethics (Gewirth 1978; Apel 1980; Habermas
1990a; O’Neill 1989b; Korsgaard 1996; Illies 2003; Darwall 2006; Enoch 2011b),
transcendental arguments are nevertheless a fringe phenomenon in both meta-ethics and
normative ethics.14 There is not much literature that discusses the respective strengths and
weaknesses of different transcendental arguments that have been put forward in ethics
and/or which critically compares the method of reflective equilibrium and transcendental
argumentation (except for Illies 2003).15 This, then, is the third goal of this thesis.
The final two goals are mainly meant to offer contributions internal to the literature on
transcendental arguments. The fourth is to criticize transcendental arguments from the
second person and to defend a transcendental argument from the first person. Most
transcendental arguments that have been put forward in ethics take as a starting point the
second-person perspective. I argue that these arguments cannot justify a categorical and
universal principle of interpersonal morality. Instead, against the grain of the literature on
transcendental arguments in ethics, I defend a transcendental argument from the firstperson perspective that has been put forward by Alan Gewirth (1978). Contrary to, for
instance, the work of Korsgaard, O’Neill and Darwall, Gewirth’s transcendental argument is
rarely discussed in mainstream (meta-)ethics.
Fifth, and finally, Kantian constructivism first and foremost aims to justify a categorical
and universal supreme principle of morality through a transcendental argument, and
14
Not all of these authors self-identify as a Kantian constructivist, though Habermas, Korsgaard and O’Neill do. In
addition, I think that Apel and Gewirth are best understood as Kantian constructivists. Things are different in the
cases of Darwall, Enoch and Illies. Illies and Enoch explicitly claim that they use a transcendental argument for
moral realist conclusions. Darwall, on the other hand, claims that his argument is compatible with both Kantian
constructivism and realism (2006, 293). In chapter 3, I argue that a transcendental argument cannot lead to realist
conclusions by discussing David Enoch’s transcendental argument for moral realism. For this reason, I think that
Enoch’s and Illies’ transcendental arguments fail. In addition, I therefore think that Darwall’s theory about moral
metaphysics cannot remain neutral and that it should be interpreted as a form of Kantian constructivism.
15
There is a certain tension between the fact that some of the most prominent names in ethics – Korsgaard,
O’Neill and Darwall – have put forward a transcendental argument for their fundamental moral principles, with
the claim that transcendental arguments are a fringe phenomenon in ethics. I think the reason for this is that
although these authors make use of transcendental arguments, they do so rather hesitantly. O’Neill, for instance,
never explicitly claims that she puts forward a transcendental argument, at least as far as I am aware, even though I
think her argument for the categorical imperative (see in particular O’Neill 1986) clearly has a transcendental
structure (see chapter 5). The same applies to Darwall (at least in The Second-Person Standpoint (2006)). In The
Sources of Normativity, Korsgaard mentions only once that her argument is a transcendental argument (Korsgaard
1996, 123); the term is absent in Self-Constitution, although her idea of ‘constitutive standards’ (Korsgaard 2009,
28–29) could, on some interpretations of the idea of constitutivism, be understood as another way of saying that
she makes a transcendental argument (I briefly discuss the relation between constitutivism and transcendental
arguments in chapter 3). Things are different in the case of Apel, Gewirth and Habermas. Although they are (or
‘were’ in the case of Habermas) loud and proud defenders of a transcendental argument, their transcendental
arguments are rarely discussed in mainstream ethics.
INTRODUCTION
25
subsequently applies this principle to specific contexts of action in order to justify both
norms of action and concrete judgements about actions. The fifth and final contribution of
this thesis is to critically analyse the substantive implications of Kantian constructivism, i.e.
the extent to which Kantian constructivism can justify any concrete and substantive moral
permissions, prohibitions, requirements or constraints, and to discuss the relation between
the justification of a supreme principle of morality and its application.
5. Argumentative Strategy: A Potential Dilemma for Kantian
Constructivism
The argumentative strategy adopted in this thesis can be understood as an argument by
elimination. I begin, in chapter 2, by criticizing what I take to be the most dominant
approach to moral justification in ethics, the method of reflective equilibrium. In the
remainder of the thesis, I evaluate to what extent existing transcendental arguments can
succeed where reflective equilibrium fails, which, taken together, is meant as a defence of a
Kantian constructivist justification of moral objectivity.
In chapter 3, I begin by examining the general strategy of transcendental
argumentation, which can be found inside and outside ethics. I think that it is fair to say
that Kantian constructivism, and transcendental argumentation more generally, is typically
met with widespread scepticism. One reason for scepticism is the largely critical reception
of transcendental arguments in theoretical philosophy.16 The thought seems to be that if
transcendental arguments fail in theoretical philosophy, they will certainly be of no use in
ethics. In this chapter, I take issue with this line of reasoning by showing that
transcendental arguments in ethics, because of their normative subject matter, escape the
problems that transcendental arguments in theoretical philosophy are said to suffer from, at
least when they are not used to justify moral realist conclusions.
The remainder of the thesis is concerned with what I’ll refer to as the potential ‘general
dilemma for Kantian constructivists’. For even if it is true that transcendental arguments in
ethics do not (necessarily) have the problems of transcendental arguments in theoretical
philosophy, important sources of scepticism still remain. These objections, I suggest, can be
(re)formulated in terms of a potential dilemma: either the starting point of a transcendental
16
The recent debate on transcendental arguments in theoretical philosophy took off with Stroud (1968). For an
overview of transcendental argument in theoretical philosophy see Stern (1999; 2000; 2013a).
CHAPTER 1
26
argument is inescapable, but in that case no interesting (moral) conclusions can follow
from the starting point, or Kantian constructivism might succeed in justifying interesting
(moral) conclusions, but in that case the starting point turns out to be merely optional (not
inescapable) and hence we are no longer dealing with a genuine transcendental argument.
But what are ‘interesting (moral) conclusions’? My suggestion is to distinguish between
three sets of potentially interesting conclusions: categorical normative claims (chapter 4),
categorical and universal normative claims, i.e. claims about interpersonal morality
(chapters 5 to 7), and substantive categorical and universal normative claims (chapters 8
and 9). I use this tripartite distinction to structure the discussion of transcendental
arguments, and I evaluate existing transcendental arguments in light of these three
instantiations of the potential general dilemma for Kantian constructivism.
5.1 Categoricity
Some of those who are sceptical about transcendental arguments think that a
transcendental argument cannot lead to any categorical normative conclusions, for instance
because any starting of a transcendental argument will be escapable (Enoch 2006; Enoch
2011a; Silverstein 2015) or because, even if some starting point, such as agency, is
inescapable it cannot lead to any normative conclusion (Street 2012). I think this is the kind
of scepticism that can be found in recent debates in meta-ethics and that needs to be
answered before we can even start thinking about interpersonal or substantive morality. In
chapter 4, I argue that a transcendental argument from the first person is not susceptible to
these general meta-ethical objections to the Kantian constructivist project, but that this
leaves open the possibility that there might be specific starting points, e.g. specific
conceptions of agency, that are escapable, or that there are specific transcendental
commitments which can be rejected without contradicting one’s own self-understanding.
5.2 Universality
Even if one allows that there are categorical reasons that can be justified through a
transcendental argument, one might question the plausibility of this argument leading to
any interpersonal moral principles, i.e. categorical and universal principles having to do
with the question of how we ought to treat others. Because morality is typically (although
not necessarily exclusively) understood in terms of the regulation of interpersonal conduct,
INTRODUCTION
27
this is a crucial desideratum for a successful transcendental argument in ethics.17 In this
context, the distinction between what I have called transcendental arguments from the first
person and transcendental arguments from the second person is of particular relevance.
In chapter 5, I turn to so-called transcendental arguments from the second person and
argue that they are all susceptible to one or both horns of the general dilemma for Kantian
constructivism. Either participation in some form of interaction is inescapable, but in that
case it is unclear why this kind of interaction has any universal principles of interpersonal
morality among its necessary conditions of possibility, or interaction has a universal
principle of interpersonal morality among its necessary conditions of possibility – but in
that case interaction seems to be optional and the moral principles are therefore not
categorically valid.
In chapter 6, I reconstruct several first person interpretations of Christine
Korsgaard’s argument for the value of humanity, and I argue that none of these
interpretations succeeds in justifying a categorical and universal principle of interpersonal
morality. Again, the reason for this is that either the starting point of her argument is
escapable or the starting point is inescapable, in which case no normative principle of
interpersonal morality follows from it.
In chapter 7, I discuss and defend Alan Gewirth’s transcendental argument from
the first person for a supreme principle of interpersonal morality, the principle of generic
consistency. I describe the main features of his argument and argue why Gewirth succeeds
where Korsgaard fails and why I think that the most common criticisms of Gewirth fail.
5.3 Substance
Even if one thinks that a categorical and universal moral principle could be justified
through a transcendental argument, one might question whether this principle can lead to
any concrete and substantive moral permissions, prohibitions, requirements or constraints.
That is, one might wonder whether it is possible to derive any substantive conclusions from
my self-understanding as an agent or from the necessary preconditions of interaction
which form the starting point of the transcendental argument. In the final part of the thesis,
I address the question of substantive morality. I limit my focus to the way in which Kantian
17
On some accounts of morality, it could also include duties towards oneself; it is therefore important not to
equate morality with interpersonal morality. Note, however, that the authors discussed in this thesis all try to
justify interpersonal moral duties.
CHAPTER 1
28
constructivism aims to justify certain substantive norms of action and I do not discuss the
question of judgement, i.e. the question of how to justify concrete judgements about
specific actions (versus norms about certain action types). The reason for this is not that
this question is unimportant but because I simply cannot discuss the complex question of
judgement within the scope of this thesis.
In chapter 8, I introduce the emptiness objection to Kantian ethics in general and
Kantian constructivism in particular, and I discuss the general strategy that Kantian
constructivists use to argue for substantive norms of action on the basis of a supreme
principle of morality. The strategy is to apply the supreme principle of morality through a
reliance on both general and contingent anthropological facts. I argue that I agree with
Kantian constructivists that reliance on these facts is both necessary and legitimate in the
application of a supreme principle of morality.
In chapter 9, I discuss the question of whether the reliance on anthropology is not just
necessary but also sufficient in order to justify substantive norms of action. I try to answer
this question by providing an in-depth analysis of the application of a supreme principle of
morality by both Onora O’Neill and Alan Gewirth. I argue that the question of whether or
not Kantian constructivism can justify a wide range of norms of action ultimately depends
on its specification of the abstract conception of agency and that this has not received
sufficient attention from Kantian constructivists. As long as the specification of the abstract
conception of agency is not made explicit and defended, critics are right to complain that it
is unclear how substantive norms of action can be justified on the basis of a supreme
principle of morality. However, at the same time, critics are only justified in their
scepticism about the substantive potential of Kantian constructivism once it is concluded
that there is no specific conception of agency that is inescapable. This counts neither as a
defence of the substantive ambition of Kantian constructivism nor as a criticism of this
approach. Instead it shows where the action should be.
Let me end this introduction by making explicit that this thesis operates somewhere in
between meta-ethics and normative ethics. The thesis operates in meta-ethics insofar as it
explores the plausibility of Kantian constructivism, and more specifically transcendental
argumentation as a method of moral justification, and insofar as it engages with objections
raised to both Kantian constructivism in general and transcendental argumentation
specifically in recent debates in meta-ethics. This thesis operates in normative ethics insofar
as it discusses the possibility of justifying a supreme principle of interpersonal morality
INTRODUCTION
29
through a transcendental argument and the question of how substantive norms of action
can be justified on the basis of this supreme principle of morality. The general idea behind
this focus is that an account of the source of moral objectivity is incomplete without an idea
of which principles and norms can be justified as (objective) moral claims. But this should
also be considered the other way around: we can only assess the plausibility of principles
and norms that are put forward as moral claims once we know what justifies moral claims
in the first place. To summarise, the success of meta-ethics (as least insofar as it provides an
account of moral justification) depends at least in part on normative ethics, and the success
of normative ethics depends in part on meta-ethics.18 In this sense, this thesis goes against a
dominant trend in ethics that conceives of meta-ethics and normative ethics as
independent domains.19
Because my focus in this thesis is neither purely meta-ethical nor purely normative
ethical but rather on the relation between these two domains, this has implications for the
selection of authors and positions that are discussed in this thesis. For instance, within
meta-ethics I focus exclusively on cognitive theories, i.e. theories that claim that moral
statements can be true or false or at least right or wrong. When it comes to methods of
moral justification, I focus mainly on the method of reflective equilibrium because this
method is by far the most dominant method of moral justification in both normative ethics
and applied ethics and it is therefore particularly relevant in the context of discussing the
possibility of a justification of objective moral claims. Finally, in normative ethics my focus
is mainly on questions in normative political theory, and theories of justice in particular.
The reason for this is that many people discussed in this thesis, e.g. O’Neill and Gewirth,
18
I think the most interesting work in ethics is located at this intersection. Similar projects, although in different
ways and also to different degrees, are, I think, pursued by, for instance, 1) Rawls in A Theory of Justice (1999b),
insofar as it makes a contribution to normative ethics and also reflects quite extensively on the question of moral
justification (although the main focus is clearly on normative (political) theory), 2) Onora O’Neill throughout
most of her work (see in particular O’Neill 1989b; O’Neill 1996) and 3) Alan Gewirth (1978) in Reason and
Morality. Of course, my point is not to compare this thesis to these monumental works in ethics, but instead to use
them as illustrations of the kind of project I am interested in.
19
A dominant assumption of meta-ethics is that meta-ethics is neutral when it comes to questions of normative
and applied ethics, i.e. that meta-ethical theories are compatible with a great variety of different normative and
applied ethical positions (Sayre-McCord 2014). Sometimes this neutrality is even used as a criterion to distinguish
meta-ethical positions from normative ethical positions. On this view, a meta-ethical theory is defined in terms of
its neutrality (for an influential example see Dworkin 1996; Dworkin 2011). Similar assumptions can be found in
normative ethics. John Rawls, for instance, famously claims that normative and also applied ethics can be pursued
independently from meta-ethics. This assumption can be traced back to Rawls (1974). See also Rawls (2005). Of
course, I am not the first or the only person to question these assumptions (see e.g. Gewirth 1960; Korsgaard
2008a; McPherson 2012), but it is still fair to say that these assumptions are widespread and often considered to be
so self-evident that they are typically unreflectively endorsed.
CHAPTER 1
30
have written mainly on political morality, and the context of political morality is therefore a
good test case for these theories.
This thesis can thus be understood as operating at the intersection of meta-ethics and
normative ethics. My hypothesis is that a proper understanding of what justifies moral
claims requires the adoption of a unified approach towards ethics. This, however, is not
something that I will be arguing for explicitly in the chapters to come, but which, I hope,
the thesis as a whole can be taken to demonstrate.
Part 1: Methods
Chapter 2
Reflective Equilibrium
1. Introduction
The central aim of this thesis is to analyse the plausibility of a Kantian constructivist
justification of moral objectivity, which aims to justify objective moral claims through a
transcendental argument. 1 Before assessing the plausibility of different transcendental
arguments that have been put forward in ethics, it is, however, important to first make clear
why I think that alternative methods of moral justification are problematic. As already
noted, I will limit myself to discussing the method of reflective equilibrium. The reason for
this is that reflective equilibrium is the most dominant method of moral justification in
both normative and applied ethics.2 In addition, reflective equilibrium is widely accepted by
both moral realists and moral constructivists as their preferred moral epistemology.3 If it
turns out that the method of reflective equilibrium should be rejected, there is a good
reason to take a closer look at an alternative, transcendental argument for moral claims.4
One of the main appeals of reflective equilibrium as a moral methodology lies in its
1
A shorter and earlier version of this chapter has been published as De Maagt (2016).
The method of reflective equilibrium has its origin in Nelson Goodman’s (1955) defence of the rules of inductive
inference. John Rawls (1999b) introduced reflective equilibrium in ethics as an alternative to rational intuitionism.
For an overview of the historical predecessors of reflective equilibrium in ethics see Brink (2014). For an overview
of the use of reflective equilibrium in applied ethics see Arras (2007).
3
For references see section 2 of this chapter.
4
Another alternative would be to reconsider intuitionism, to which the method of reflective equilibrium is put
forward as an alternative. As I have already noted in chapter 1, I do not directly discuss intuitionism in this thesis
but simply assume that there are important problems with intuitionism, which at least warrants an investigation
into an alternative, transcendental approach to moral justification. In addition, as I will show below, proponents of
reflective equilibrium typically reject intuitionism, and transcendental argumentation might therefore be
particularly interesting for proponents of reflective equilibrium, because transcendental arguments do not have
the same epistemological and metaphysical problems as intuitionism.
2
CHAPTER 2
34
ambition to provide an account of moral objectivity without having to bear the
controversial epistemological burdens of alternative, ‘foundationalist’ methods of moral
justification, such as rational intuitionism, which claim that certain moral beliefs can be
known non-inferentially (see e.g. Daniels 1979b; Daniels 1980; Rawls 1999b; Walden 2013;
Scanlon 2014).
However, ever since the introduction of reflective equilibrium in ethics, it has been
argued that reflective equilibrium cannot have it both ways. Either one accepts reflective
equilibrium as a distinctive, non-foundationalist method of moral justification, in which
case one has to give up the objectivist aspirations of ethics or one may want to preserve
moral objectivity, in which case one has to be prepared to commit oneself to the kind of
foundationalism proponents of reflective equilibrium reject. That is, the objection is that
reflective equilibrium is either committed to moral relativism or turns out to be a form of
intuitionism in disguise (Hare 1973; Singer 1974; Singer 2005; Brandt 1979).
Defenders of reflective equilibrium typically respond to these kinds of objections in two
ways. First, they try to show that reflective equilibrium does in fact have the resources to
safeguard moral objectivity without relying on foundationalist premises (Daniels 1979b;
Holmgren 1987; Brink 1989; Scanlon 2014).5 Second, they argue that even if reflective
equilibrium ultimately leads to moral relativism, this constitutes no objection to the
methodology simply because no reasonable alternative to reflective equilibrium exists as a
method of moral justification in ethics (Scanlon 2003, 149; Rawls 2003, 31; DePaul 2007,
618; Walden 2013, 254; Floyd forthcoming, 12).
In this chapter, I take issue with both responses. First, I critically evaluate the most
recent defences of reflective equilibrium and argue that they fail to show that reflective
equilibrium can lead to objective moral claims, at least insofar as reflective equilibrium is to
remain a distinctively non-foundationalist methodology. I focus mainly on Thomas M.
Scanlon’s (2014) defence of reflective equilibrium in his latest book, Being Realistic about
Reasons. Second, I argue that once intuitionism has been discarded, reflective equilibrium is
not the only game in town and that we should not just settle for moral relativism. The aim
of the chapter is to argue that given their own ambition of vindicating moral objectivity,
combined with their rejection of intuitionism, proponents of reflective equilibrium have
reason to take alternative methods of moral justification, and more specifically
5
An important exception to this rule is Street’s (2006, 110) use of reflective equilibrium as part of her ‘Humean
constructivism’, which explicitly rejects moral objectivity (see also Street 2012). However, in recent writings even
Street (2016) is reluctant to accept moral relativism.
REFLECTIVE EQUILIBRIUM
35
transcendental arguments, more seriously than they have done so far.
The structure of the chapter is as follows. In section 2, I briefly introduce the method of
reflective equilibrium. In section 3, I provide some prima facie reasons to question the
ability of reflective equilibrium to deliver the kind of moral objectivity it aspires to deliver.
In what follows, I discuss recent attempts by proponents of reflective equilibrium to face
these challenges, and conclude that reflective equilibrium cannot live up to its objectivist
promise, at least as long as one wants to retain reflective equilibrium as a distinctive moral
methodology (section 4). More specifically, I criticize both the reliance on considered
judgements in the method of reflective equilibrium (4.1 and 4.2) and the move towards a
general or intersubjective reflective equilibrium (4.3). In the final section of the chapter
(section 5), I argue that reflective equilibrium is not the only possible method of moral
justification in ethics once intuitionism has been discarded (5.1) and that we should not
settle too easily for moral relativism (5.2).
2. The Method of Reflective Equilibrium
According to the method of reflective equilibrium, moral justification does not depend on
an ultimate moral foundation but on the coherence between all moral and non-moral
considerations that are relevant to a certain issue at hand. These relevant considerations
typically include a set of ‘initial judgements’,6 a set of moral principles and a set of relevant
background theories. 7 In this section, I briefly discuss the main features of reflective
equilibrium.
Let me start by briefly illustrating how the method of reflective equilibrium can be put
to work to justify moral claims by discussing Rawls’ usage of reflective equilibrium as a
method of justification of principles of justice. Rawls starts with a set of initial judgements
about justice at different levels of abstraction, which he takes as ‘provisional fixed points’ in
6
I deliberately leave out the qualification ‘considered’. I will come back to the strategy to limit the relevant
judgements to ‘considered judgements’ in section 4.
7
It is common to make a distinction between ‘narrow reflective equilibrium’, which only includes judgements and
principles, and ‘wide reflective equilibrium’, which also includes background theories (Daniels 1979b). In this
chapter, I focus exclusively on wide reflective equilibrium because this seems to be the strongest and most widely
adopted interpretation of reflective equilibrium – to the best of my knowledge there is no one who explicitly
adopts narrow reflective equilibrium as a method of justification (in A Theory of Justice, Rawls does not explicitly
include background theories, but in later writings he makes it clear that the relevant equilibrium is indeed a wide
equilibrium and not just a narrow equilibrium (see e.g. Rawls 1974, 8)).
CHAPTER 2
36
moral theorizing (Rawls 1999b, 18).8 For Rawls, these provisional fixed points include, for
instance, the initial judgement that “religious intolerance and racial discrimination are
unjust” (1999b, 17) and that “no one deserves his greater natural capacity nor merits a more
favourable starting place in society” (1999b, 87; see also 1999b, 274).
The next step is to formulate moral principles that can account for these initial
judgements. Rawls defends two principles of justice (1999b, 266). The first principle grants
equal rights to liberties to each person. The second principle states that socio-economic
inequalities should be arranged in such a way that they are to the benefit of the worst-off
members of society (the difference principle) and that offices and positions should be open
to everyone under conditions of fair equality of opportunity. In order to formulate these
principles and in order to evaluate which principles account best for our initial judgements,
Rawls introduces the original position as an ‘expository device’ (1999b, 19). The conditions
that characterize the original position – i.e. the information available to the parties, the
characteristics of the parties and the formal conditions that apply to principles of justice –
are to be settled by analysing which description of the original position generates principles
that cohere with our initial judgements.
Besides initial judgements and moral principles, a third element in the method of
reflective equilibrium are so-called background theories. Relevant background theories
include, for example, already existing moral theories, such as utilitarianism and
Kantianism, meta-ethical theories, psychological and sociological theories about
personhood and society, and/or theories about personal identity (Rawls 1974; Daniels
1979a; Daniels 1979b). The idea is that background theories should be included because the
search for reflective equilibrium should include all possibly relevant considerations.
Importantly, background theories should have a certain degree of independence from both
initial judgements and principles, because otherwise including them would not add
anything to the justificatory process.
For instance, Rawls’ reliance on, what he calls “the natural facts about men in society”
(1999b, 137) is an important part of the argument for the difference principle.9 Rawls
argues that inequalities which benefit the worst-off members of society are justified
8
The method of reflective equilibrium does not prescribe a chronological order in which different elements should
be introduced. One could start with specific judgements, principles or relevant background theories. What follows
is just one example of how one could use the method of reflective equilibrium; nothing hinges on the specific
order in which elements are introduced.
9
For a further discussion and reconstruction of the role of facts about human nature in Rawls’ theory, see de
Maagt (2014).
REFLECTIVE EQUILIBRIUM
37
because, among other reasons, in socio-economic institutions in which no inequalities
whatsoever are allowed, the ‘talented’ would not be as productive as they would be in the
context of institutions which allow for inequality-generating incentives. In addition, Rawls
claims that the economic competition that is allowed for by the difference principle will not
lead to excessive inequalities (1999b, 137). In these ways, Rawls’ conception of justice,
‘justice as fairness’, does not just depend on the coherence between initial judgements and
principles but also on the alleged fact that human beings tend to be demotivated in the
context of strict egalitarian institutions and on the idea that the difference principle will not
lead to too much inequality.
The formulation of principles that cohere with initial judgements in light of relevant
background theories is not a linear process. In cases of conflict between initial judgements,
principles and background theories, the method might require the revision of principles,
initial judgements or background theories. According to the method of reflective
equilibrium, there are, in principle, no elements in moral theorizing, including judgements
and convictions, that are immune to revision.10 Rawls therefore claims that we should
proceed “by going back and forth, sometimes altering the conditions of the contractual
circumstances, at others withdrawing our judgments and conforming them to principle”
(1999b, 18 my emphasis). This process of going back and forth will probably not lead to a
reflective equilibrium that is fixed once and for all. Instead, the process of mutual
adjustment “continues indefinitely” (Rawls 1999b, 43).
Rawls introduces reflective equilibrium in the context of normative political theory.
Reflective equilibrium is used to justify the principles of justice that should regulate the
basic structure of society. Rawls (1999b, 95–96), however, explicitly stresses that he
introduces reflective equilibrium as a general methodology in ethics, and not as a method
that is limited to normative political theory or to the task of justifying principles of justice.
Unsurprisingly, reflective equilibrium is now used in various areas in normative and
applied ethics, and it is also adopted as a general moral epistemology.11
The widespread usage of the method of reflective equilibrium in a plurality of different
moral domains means that, at least to a certain extent, the methodology has no strictly
defined characteristics. Instead, it might be more appropriate to understand ‘the’ method of
reflective equilibrium as an umbrella term which covers a wide range of more specific
10
Pace Peter Singer (1974; 2005), reflective equilibrium is therefore not intuitionism in disguise. See also Brun
(2014).
11
See the introduction to this chapter for references.
CHAPTER 2
38
methodologies. Nevertheless, I think it is possible to detect a common core of the
methodology. 12 The distinctive claim of reflective equilibrium as a general moral
methodology is that moral justification does not depend on an ultimate moral foundation
but on the coherence between all moral and non-moral considerations that are relevant to
the issue at hand (such as initial judgements, principles and background theories). 13
Reflective equilibrium thus offers a ‘non-foundationalist’ or ‘coherentist’ account of moral
justification. Reflective equilibrium is non-foundationalist because it claims that there is no
moral claim that has a special justificatory standing independent of engaging in the process
of pursuing reflective equilibrium. 14 Reflective equilibrium is a form of coherentism
because, according to this methodology, justification, as Rawls famously writes, “is a matter
of the mutual support of many considerations, of everything fitting together into one
coherent view” (1999b, 507).
As a method of moral justification, reflective equilibrium is, at least to some extent,
neutral with respect to the metaphysics of morality.15 This is illustrated by the fact that it is
an influential method of moral justification for both moral realists and moral
constructivists. David O. Brink, for instance, combines reflective equilibrium as a method
of moral justification with his naturalist moral realism. He claims that “coherence with,
among other things, considered moral beliefs provides evidence of objective moral truth”
12
One reason for this is that most people who adopt reflective equilibrium refer back to Rawls’ description of
reflective equilibrium in A Theory of Justice and/or Norman Daniels’ (1979b) reconstruction and specification of
the method (in September 2016, Daniels’ article has more than 850 citations).
13
This obviously raises the question of what counts as the appropriate type of coherence and the criteria that
should guide the process of reaching reflective equilibrium. Rawls does not go beyond the metaphorical idea that
we should go ‘back and forth’ to describe the criteria that should guide the process of going back and forth
between different (moral) beliefs. Others have tried to provide a list of epistemic desiderata that should guide this
process, including consistency, systematicity, generality and simplicity (Kappel 2006; Campbell 2014).
14
Some authors claim that reflective equilibrium is compatible with – or even most properly understood as – a
form of modest foundationalism according to which certain beliefs, i.e. considered judgements, have a privileged
epistemic status independent of their coherence with other beliefs even though these ‘foundational’ beliefs are not
immune to revision (DePaul 1986; Holmgren 1987; Ebertz 1993; McMahan 2000). In section 4.1, I will argue that a
modest foundationalist interpretation of reflective equilibrium collapses into a form of intuitionism and that we
should therefore reject a modest foundationalist interpretation of reflective equilibrium and understand reflective
equilibrium as a distinctively coherentist methodology (see also Brink 2014; Brun 2014).
15
The idea that reflective equilibrium is neutral with respect to the metaphysics of morality should be qualified in
at least two ways. First, although it might be true that while reflective equilibrium is compatible with both moral
realism and moral constructivism, it might not be compatible with all possible metaphysical positions. For
instance, it seems that for some forms of expressivism, the question of moral justification as it appears in this
chapter wouldn’t arise in the first place. In addition, it might be possible that there are specific challenges and
problems for reflective equilibrium depending on whether one is a moral realist or a constructivist. For instance,
the realist would have to explain why reflective equilibrium has any contact with the metaphysical reality of moral
truths, whereas a constructivist would have to explain why we are morally committed to whatever follows from
reflective equilibrium. For a criticism of reflective equilibrium realism, see e.g. Altehenger et al. (2015).
REFLECTIVE EQUILIBRIUM
39
(1989, 143; for similar views see Holmgren 1987; Ebertz 1993; Scanlon 2012; Scanlon 2014).
Brink thus embraces reflective equilibrium as a moral epistemology and at the same time
holds that moral facts are ‘mind-independent’, so that there is a conceptual gap between
moral justification and moral truth. This implies that a moral belief might be ultimately
false, even though the belief is justified in the light of all possibly relevant considerations we
have in favour of the acceptance of the belief.
Rawls, on the other hand, defends a constructivist interpretation of reflective
equilibrium, claiming that his “constructivism holds that moral objectivity is to be
understood in terms of a suitably constructed social point of view that all can accept. Apart
from the procedure of constructing principles of justice, there are no moral facts” (1980,
519). In Rawls’ constructivist interpretation of reflective equilibrium, truth thus collapses
into justification: moral truth is simply whatever we arrive at through a suitable procedure
of constructing justice.
In order to understand the attraction of reflective equilibrium to both realists and
constructivists, it is crucial to note that reflective equilibrium is introduced in ethics, first
and foremost, as an alternative to various forms of epistemological intuitionism. The
introduction of reflective equilibrium as an alternative to intuitionism is most clearly
articulated in Rawls’ A Theory of Justice, in which he explicitly rejects the idea that his
principles of justice “are necessary truths or derivable from such truths. A conception of
justice cannot be deduced from self-evident premises or conditions on principles” (1999b,
19; for similar contrasts between reflective equilibrium and intuitionism see Rawls 1980,
559; Brink 1989, 8–9; Daniels 1979b, 264–67; Scanlon 2014, 69–72). Instead, Rawls claims
that the aim of his theory is to set up “an Archimedean point for assessing the social system
without invoking a priori considerations” (Rawls 1999b, 231). The introduction of reflective
equilibrium as a method of moral justification is thus meant to disprove the idea that “there
is no alternative but to judge institutions in the light of an ideal conception of the person
arrived at on perfectionist or on a priori grounds” (Rawls 1999b, 230). This goal of
searching for an alternative moral epistemology to intuitionism is not limited to moral
constructivists such as Rawls, but, as I have already noted above, also concerns a substantial
number of moral realists who aim to disentangle the metaphysics of moral realism from the
epistemology of intuitionism.
Although proponents of reflective equilibrium reject the foundationalism of
intuitionism, they share its objectivist aspirations. The ambition to combine reflective
equilibrium with an objective ethics is most clearly expressed by Rawls when he claims that
CHAPTER 2
40
[the] principles [of justice] are objective. They are the principles that we would
want everyone (including ourselves) to follow were we to take up together the
appropriate general point of view … We do not look at the social order from our
situation but take up a point of view that everyone can adopt on an equal footing.
In this sense we look at our society and our place in it objectively: we share a
common standpoint along with others and do not make our judgments from a
personal slant. Thus our moral principles and convictions are objective to the
extent that they have been arrived at and tested by assuming this general
standpoint and by assessing the arguments for them by the restrictions expressed
by the conception of the original position (1999b, 453; for similar claims about the
objectivist aspirations of reflective equilibrium see Daniels 1979b, 273–81; Brink
2014, 677f26; Scanlon 2014, 93–94).16
Rawls thus suggests that the principles of justice are objective in the sense that they should
be accepted by any individual independent of his or her personal preferences or desires.
If reflective equilibrium succeeds in fulfilling this ambition, the method seems to be
able to combine the best of both worlds, because it would be able to deliver objective moral
claims while steering clear of the epistemological controversies surrounding intuitionist
accounts of moral justification. In addition, reflective equilibrium constructivism could
provide an alternative to moral realism. However, in what follows, I will argue that
proponents of reflective equilibrium cannot have it both ways, i.e. they cannot justify moral
objectivity while defending reflective equilibrium as a distinctive coherentist moral
methodology.
16
Although they often do not explicitly claim that their theories are objective (a notable exception is Caney (2006,
25–57)), I think that most normative and applied ethicists who use the method of reflective equilibrium at least
implicitly think that their theories have an objective validity. Although this is an empirical claim that I cannot
substantiate, there are some observations to support my claim. First, there are, so far as I am aware, almost no
philosophers working in normative and applied ethics who explicitly call themselves a moral relativist. Second,
normative and applied ethicists typically refer back to both Rawls and Daniels when describing their methodology.
Both Rawls and Daniels explicitly claim that reflective equilibrium can vindicate moral objectivity, so until
ethicists explicitly reject moral objectivity it seems safe to assume that they do believe that moral objectivity can be
vindicated through the method of reflective equilibrium.
REFLECTIVE EQUILIBRIUM
41
3. Reflective Equilibrium and Moral Objectivity
In order to evaluate whether or not the method of reflective equilibrium is indeed able to
generate objective moral claims, we should first be more precise about what exactly the idea
of moral objectivity refers to. In the introduction to this thesis, I have already distinguished
realist from constructivist conceptions of objectivity. In this section, I say a bit more about
this distinction and more specifically on how it relates to reflective equilibrium and its
ambition to vindicate objectivity. The point here is not so much to argue for a specific
conception of moral objectivity, but to get a better understanding of the idea of moral
objectivity that proponents of reflective equilibrium themselves accept (and subsequently to
criticize reflective equilibrium on its own terms).
For a start, we can understand the concept of objectivity as referring to the idea that the
truth or correctness of a certain claim is in some way independent of us. This intuitive
understanding of objective in terms of independence is, however, open to different
interpretations. I think that it is useful to distinguish a minimal from a maximal
interpretation of objectivity.
On a minimal interpretation of objectivity, a moral claim is objective when the
correctness of this claim is independent of whether or not someone actually accepts this
particular moral claim. That is, on this understanding of objectivity, the fact that one
accepts a particular moral claim – e.g. the judgement that redistributive taxation is a
violation of property rights – is not what justifies the moral claim. Minimal objectivity thus
allows for the possibility that one can be wrong about a particular moral claim, because it
separates the acceptance of a particular claim from its being justified.17
Obviously, reflective equilibrium can account for this minimal type of objectivity. In
reflective equilibrium, a moral claim is not simply justified when an individual actually
accepts a specific claim but only when this particular claim coheres, in the appropriate way,
with all possibly relevant considerations relating to the issue at hand. One’s judgement that
redistributive taxation is a violation of property rights might, for instance, be undermined if
it is shown to be inconsistent with the moral norms one accepts about the justification of
private property and/or the relation between the idea of self-ownership and democratic
decision-making. In reflective equilibrium, a particular moral claim might thus be
17
Note that this minimal interpretation of objectivity is compatible with most forms of moral relativism. In what
follows, I therefore focus on a more maximal interpretation of objectivity. I give reasons for this below.
CHAPTER 2
42
disproven when it is shown to be inconsistent with other moral claims the individual holds,
assuming that it is more rational to give up the particular claim under consideration rather
than the background beliefs with which the claim is shown to be inconsistent.18
However, although the possibility that one might be wrong about a particular claim
captures an important feature of moral objectivity, moral claims are typically thought to be
objective in a more robust sense than is demanded by minimal objectivity. In the
introduction, I have defined moral objectivity in terms of categoricity, the idea that moral
claims are those claims that should be accepted independent of the contingent beliefs or
desires one happens to have, whether or not these beliefs and desires are internally
coherent, and universality, i.e. the idea that the scope of moral claims include all human
beings (e.g. respect for all human beings). On a maximal interpretation of objectivity, the
correctness of a moral claim is thus independent of someone’s contingently held belief set.
There are different ways to understand this maximal interpretation of objectivity
(referred to from now on as objectivity). Moral realists think that only the existence of
moral facts can account for the objectivity of moral claims (ontological conception of moral
objectivity), whereas a constructivist claims that reference to moral facts is not necessary in
order to justify moral objectivity (practical conception of moral objectivity). Rawls’
intersubjective conception of objectivity – the ‘common standpoint’ we share with others
which I quoted above – is an example of this kind of practical objectivity.
Despite the important differences between these different conceptions of objectivity,
they have in common the idea that objectivity demands that an individual could be wrong
about a moral claim even though the claim coheres with the other relevant moral and nonmoral beliefs he or she happens to hold. And even though realists and constructivists
ultimately have different conceptions of moral objectivity – mind-independent moral truth
vs. a practical standpoint every rational agent should adopt – there are several reasons to be
sceptical about the idea that reflective equilibrium can deliver objectivity, however one
understands it.
First, note that in order to generate objective moral claims, we should have reasons to
believe that there will be (sufficient) convergence of a set of moral claims in reflective
equilibrium (Daniels, 1979, 273-274. This kind of convergence seems to be essential to a
method of moral justification because otherwise there might be as many possibly
18
Inconsistency is a very minimal interpretation of coherence, and some authors argue that reflective equilibrium
needs a more substantive idea of coherence in order to be plausible (see e.g. Arras 2007).
REFLECTIVE EQUILIBRIUM
43
conflicting, justified moral claims as there are people engaging in the method of reflective
equilibrium. Sufficient intersubjective agreement is the most important indicator of moral
objectivity in the method of reflective equilibrium, because without this kind of agreement
there might be a plurality of incompatible reflective equilibria for which reflective
equilibrium offers no resources to favour one of them over the other equilibria.
But why should we believe that convergence on a reflective equilibrium is likely to
happen? Not only is there persistent disagreement about initial judgements in specific
cases, but this is also the case with respect to background theories. For instance, there is a
plurality of incompatible metaphysical theories about personhood that have their own
normative implications (for relevant discussions see Parfit 1973; Rawls 1974; Daniels 1979a;
Scheffler 1982; Korsgaard 1989). In addition, there is disagreement about which
background theories are morally relevant in the first place (Cohen 2003; Estlund 2011; de
Maagt 2014). It is unclear how bringing different beliefs into coherence could increase our
confidence in the possibility of intersubjective agreement on morality. It is therefore not
surprising that the method of reflective equilibrium is used to support a wide range of
normative (political) theories and positions, including utilitarianism (Brink 1989); Rawlsian
political theory (Rawls 1999b); left-libertarianism (Vallentyne, Steiner, and Otsuka 2005);
cosmopolitanism (Caney 2006); and nationalism (Miller 2012). It is important to note that
the problem here is not merely that there might be a plurality of possibly conflicting
reflective equilibria, but more importantly that the methodology offers no resources to
adjudicate between different equilibria, which means no more and no less than that it is
normatively arbitrary which equilibrium one ends up accepting.
Second, there is a deeper concern about reflective equilibrium as a coherentist method
of justification in relation to its claim to moral objectivity. Even if we assume that
convergence of certain moral claims is to be expected, we might wonder why the very fact
of coherence provides us with a reason to accept some moral conception, i.e. whether this is
a reason to believe that it tracks independent moral truth or that it expresses a practical
standpoint that every individual should adopt. This challenge is known as the ‘garbage in,
garbage out’ objection (Jones 2005, 75). Richard Brandt formulates this challenge as
follows:
There is a problem here which is quite similar to that which faces the traditional
coherence theory of justification of belief: that the theory claims that a more
coherent system of beliefs is better justified than a less coherent one, but there is
CHAPTER 2
44
no reason to think that this claim is true unless some of the beliefs are initially
credible – for some reason other than coherence, say, because they state facts of
observation (1979, 20).
Imagine that someone starts from an initial judgement that entails moral disapproval of
homosexual relationships and consequently formulates moral principles that account for
this alleged immorality (e.g. the principle that only sexual relationships between men and
women are morally acceptable). Checking this principle against relevant background
theories might show that the principle in fact coheres with certain religious or evolutionary
background theories.19 Although we obviously think that this initial judgement and the
subsequent choices made in reflecting on this judgement that ultimately lead to a reflective
equilibrium are unreasonable, it is unclear whether reflective equilibrium has the
theoretical resources to rule out these kinds of equilibria, which are obviously unreasonable
from our point of view (and in the next section I will argue that reflective equilibrium does
not have these resources).
The question is thus why we should believe that coherent homophobia is better
justified than incoherent homophobia and, in fact, why coherent homophobia is justified at
all. It seems that the method of reflective equilibrium cannot provide the resources to
exclude these moral prejudices being brought into coherence in reflective equilibrium.
Thus, as long as the input into the method of reflective equilibrium is not initially credible,
the equilibrium arrived at after due reflection “may be no more than a reshuffling of moral
prejudices” (Brandt 1979, 22). This seems to threaten the idea that reflective equilibrium is
an appropriate method to generate objective moral claims.
These two objections provide prima facie reasons to doubt reflective equilibrium’s
ability to generate objective moral claims. In the next section, I will discuss recent attempts
by proponents of reflective equilibrium to respond to these kinds of objections and argue
that they have not succeeded in formulating a satisfactory answer – at least not one that is
able to guarantee reflective equilibrium’s objectivist aspirations while at the same time
remaining a distinctive, coherentist moral methodology.
19
Although the use of background theories might also uncover that one’s disproval of homosexual relationships to
some extent expresses a (contingent) cultural idea about sexual relationships, this does not have to lead to a
revision of one’s initial judgement (are there initial moral judgements which do not, in one way or another,
express certain cultural ideas?).
REFLECTIVE EQUILIBRIUM
45
4. Rescuing Objectivity
Proponents of reflective equilibrium typically point out that their moral methodology does
have the theoretical resources to exclude unreasonable equilibria and to make convergence
on a reflective equilibrium more likely than one might expect, i.e. they claim that reflective
equilibrium can vindicate moral objectivity after all. In this section, I discuss two
prominent strategies pursued by proponents of reflective equilibrium in an attempt to
rescue moral objectivity: first, I discuss the role of considered initial judgements (referred to
from now on as considered judgements) in reflective equilibrium (4.1), and, more
specifically, Scanlon’s recent defence of the use of considered judgements in reflective
equilibrium (4.2). Second, I discuss the idea of a general or intersubjective reflective
equilibrium (4.3).
4.1 Considered Judgements
Whereas the challenges mentioned in the previous section presuppose that any input is
allowed into the method, most (if not all) versions of reflective equilibrium limit the set of
judgements that are allowed into the process of reaching a reflective equilibrium. Rawls
pursues this first strategy when he limits the relevant set of initial judgements to what he
calls considered judgements. According to Rawls (1999b),
considered judgments … enter as those judgments in which our moral capacities
are most likely to be displayed without distortion … For example, we can discard
those judgments made with hesitation, or in which we have little confidence.
Similarly, those given when we are upset or frightened, or when we stand to gain
one way or the other can be left aside. All these judgments are likely to be
erroneous or to be influenced by an excessive attention to our own interests
(1999b, 42).
Rawls suggests that only those judgements that are made under conditions that are
favourable for making correct judgements should be allowed into the method of reflective
equilibrium. Rawls mentions two types of constraints on judgements that should guarantee
that judgements have some initial credibility. First, we should be sufficiently certain about a
judgement. This constraint excludes judgements in which we have little confidence.
CHAPTER 2
46
Second, Rawls mentions several psychological states – distress, fear, etc. – that are
unfavourable in making reasonable judgements. Thus, only those judgements we make
when we are ‘in the right state of mind’ are allowed to enter reflective equilibrium.
The idea here is thus that reflective equilibrium is able to guarantee objective
objectivity because the starting points should have at least some initial credibility. In order
to prevent the outcome of reflective equilibrium being unreasonable or merely a set of
moral prejudices, the set of initial judgements that is allowed in reflective equilibrium is
limited. In addition, by limiting the set of initial judgements that is admitted into the
process of reaching equilibrium, it might also become more likely that there is convergence
of a set of principles, because a significant amount of moral disagreement is ruled out from
the very start.
However, the use of considered judgements raises two problems. First, one might
wonder whether limiting the input to considered judgements is necessary and sufficient for
making reasonable judgements. That is, judgements made under unfavourable conditions
do not necessarily seem to be unreasonable, and judgements made under favourable
conditions do not necessarily seem to be reasonable judgements (cf. Kelly and McGrath
2010, 349). It seems unlikely, for instance, that all individuals who agreed on the moral
permissibility of slavery were not ‘in the right state of mind’. In addition, it might be
possible that despite being uncertain or upset one can make a reasonable judgement –
social and political activism against, say, slavery or child soldiers often has its roots in the
fact that people are upset. And it might well be that in times when homosexuality was
commonly regarded as a sin, someone who judged the opposite was in fact hesitant about
this judgement. Surely we do not want to exclude these kinds of progressive moral
judgements from our moral methodology simply because people were upset or uncertain
about their judgements.
In a recent article, Kelly and McGrath (2010) therefore propose replacing Rawls’
procedural notion of ‘consideredness’ with a more substantive characterization of which
initial judgements are reasonable. Thus, according to them, we should first justify a set of
reasonable moral judgements before we can use them as input into reflective equilibrium.
In a similar way, Brink proposes including not just moral judgements that have been
formed “under conditions of general cognitive reliability but also those that have been
formed on the basis of an impartial and imaginative consideration of the interests of the
relevant parties” (1989, 132 my emphasis). It seems that by relying on a more substantive
characterization of reasonable initial judgements, reflective equilibrium can indeed respond
REFLECTIVE EQUILIBRIUM
47
to the objections mentioned above, because only those moral claims are justified that
cohere with other moral beliefs that have a certain epistemic authority independent of
reflective equilibrium.
However, limiting the admissible input to either reasonable or impartial judgements
has significant consequences for reflective equilibrium as a distinctive method of moral
justification, because arguably “the most interesting part of the story concerns not the
pursuit of equilibrium itself, but rather what makes it the case that certain starting points
are more reasonable than others, and how we manage to recognize or grasp such facts”
(Kelly and McGrath 2010, 353–54). This brings us to the second problem with the
introduction of considered or reasonable judgements.
Even if one does not accept the conclusion that one needs more than merely procedural
criteria in order to generate a reasonable equilibrium, any qualification of the starting
points of reflective equilibrium – e.g. by referring to ‘considered’ or ‘reasonable’
judgements – that is substantive enough to guarantee objectivity needs a justification that
cannot be provided by reflective equilibrium itself. With the introduction of limits on the
input into reflective equilibrium, the method of reflective equilibrium thus becomes
parasitic on some other method of justification, such as intuitionism.
Although this need for justification is most obvious in variants of the methodology
which depend on a substantive characterization of the admissible input, Rawls’ procedural
notion of consideredness is in similar need of justification. For one thing, Rawls’ criteria for
considered judgements simply exclude egoism as a possible normative position. 20 It
therefore seems that his notion of consideredness implies certain substantive normative
commitments that should be justified instead of remaining implicit in his moral
methodology.
Saving reflective equilibrium by qualifying the starting points thus undermines
reflective equilibrium as a distinctive coherentist moral methodology, because it is no
longer the coherence of all possibly relevant considerations but the reasonableness or
20
Rawls (1999b) seems to believe that the specific conditions he introduces are fairly uncontroversial. He claims,
for instance, that “the criteria that identify these judgments are not arbitrary. They are, in fact, similar to those that
single out considered judgments of any kind. And once we regard the sense of justice as a mental capacity, as
involving the exercise of thought, the relevant judgments are those given under conditions favorable for
deliberation and judgment in general” (42). However, either the criteria are in fact uncontroversial, but in that case
they are insufficient to do any interesting normative work, or the criteria are substantial enough to guarantee
objectivity – but in that case they are in need of justification.
CHAPTER 2
48
consideredness of the input that bears most of the justificatory weight.21 Or at least, it
makes it hard to see to what extent reflective equilibrium is still an alternative to other
moral methodologies, because the idea that given certain (foundational) starting points our
belief set should be coherent seems to be compatible with any other method of moral
justification. Introducing limits on the input into reflective equilibrium thus comes at the
cost of undermining one of the most important features of the method of reflective
equilibrium: the idea that there are no morally valid claims independent of the procedure of
reaching a reflective equilibrium that can determine its reasonableness.
Some proponents of reflective equilibrium, however, argue that the reliance on an
epistemically privileged set of moral claims, e.g. considered judgements, does not
undermine reflective equilibrium as a distinctive method of moral justification. Reflective
equilibrium, so they argue, is best understood not as a coherentist methodology but as a
‘modest foundationalist’ method of moral justification (DePaul 1986; Holmgren 1987;
Ebertz 1993; McMahan 2000). According to this interpretation of reflective equilibrium,
certain judgements have a privileged epistemic status independent of their coherence with
other beliefs. This foundationalism is, however, ‘modest’ because these foundational beliefs
are not immune to revision. For instance, if, on due reflection, one discovered that one’s
considered judgements have been formed under conditions that are likely to import moral
biases into one’s belief set, this would constitute a reason to revise the set of considered
judgements.
Although a modest foundationalist interpretation of reflective equilibrium might be
able to deliver objective moral beliefs, it comes at the cost of reintroducing intuitionism
through the back door. Whereas reflective equilibrium was originally introduced in ethics
as an alternative to intuitionism, a modest foundationalist interpretation of reflective
equilibrium can be best understood as a form of intuitionism, because it accepts the
metaphysical and epistemological commitments of intuitionism and only adds the
qualification that our intuitions should be revisable. 22 In other words, in the modest
foundationalist interpretation, the most interesting part of the justificatory story is not the
critical reflection on all relevant considerations but the question of why certain beliefs have
a special epistemic status.
21
I agree with Kelly and McGrath when they conclude that “of course, once that move is made, one might very
well wonder whether the picture of inquiry that emerges still deserves the name ‘the method of reflective
equilibrium’” (2010, 353).
22
A qualification that recent versions of intuitionism would be happy to accept as well (see e.g. Audi 2005, 29).
REFLECTIVE EQUILIBRIUM
49
Of course, it might turn out that a modest foundationalist interpretation of reflective
equilibrium is to be preferred over a coherentist interpretation, but in that case the
proponents of reflective equilibrium will have to engage in the kind of epistemological and
metaphysical questions they typically want to avoid. Thus, insofar as we want to understand
reflective equilibrium as a truly distinctive methodology and as an alternative to
intuitionism, we should focus on reflective equilibrium as a coherentist moral
methodology.23 And insofar as it wants to remain a coherentist methodology it cannot
deliver objective moral beliefs. Or so I have argued.
4.2 Scanlon on Considered Judgements
It seems that the method of reflective equilibrium can only guarantee objectivity by
undermining the basic methodological commitments that set it apart from foundationalist
alternatives such as intuitionism. But this conclusion is reached too quickly. Recently,
Scanlon (2014) has responded to the criticism that the use of considered judgements makes
reflective equilibrium parasitic on another moral methodology, and he argues that the
inclusion of considered judgement is in fact not in tension with reflective equilibrium once
we properly understand the way considered judgements are selected.
Scanlon agrees that we need to make a substantive evaluation about which judgements
qualify as considered judgements in order for a reflective equilibrium to have any
justificatory credentials. According to Scanlon, mere coherence is not enough to confer
justificatory status on a set of beliefs, and there is thus no reason to accept homophobia
simply by virtue of the fact that it could be part of a coherent belief set. Instead, Scanlon
claims that “the justificatory force of the fact that we have arrived at certain judgments in
reflective equilibrium depends on the substantive merits of the judgments we make along
the way, in beginning with certain considered judgments and in modifying these judgments
and others as we progress” (2014, 82). Thus, he claims that “in deciding whether to count a
belief among our considered judgments, the question we ask is whether it is something that
it is reasonable to believe” (2014, 81). In order to determine whether or not something is
reasonable to believe, we have to establish whether “it seems reasonable to believe this”
(2014, 81).
It might appear as if Scanlon, by relying on the reasonableness of judgements,
23
An additional reason to reject a modest foundationalist interpretation of reflective equilibrium is that this
modest foundationalism is clearly in conflict with constructivist interpretations of reflective equilibrium.
CHAPTER 2
50
reintroduces intuitionism through the back door. This suspicion is further strengthened by
his claim that “it seems that we can discover normative truths and mathematical truths
simply by thinking about these subjects in the right way” (2014, 80). Drawing an analogy
between normative truths and mathematical truths suggests that certain normative truths
can be known a priori, through thinking about them in ‘the right way’. In addition, he
claims that considered judgements are judgements “that seem clearly to be correct” (2014,
77) and “that seem to me to be clearly true when I am thinking about the matter under good
conditions for arriving at judgments of the kind in question” (2014, 82). The question of
whether or not a judgement is a considered judgement then depends “on the truth of claims
about which things are or are not reasons for action” (Scanlon 2012, 236). These kinds of
claims look very similar to the kinds of claims that intuitionists make about the way we
arrive at true moral judgements.24
Scanlon, however, makes it clear that reflective equilibrium “does not privilege
judgments of any particular type – those about particular cases, for example – as having
special justificatory standing” (2014, 77). More specifically, he claims that the
reasonableness of judgements should not be determined independent of engaging in
reflective equilibrium, but that this question is itself part of the process of reaching
reflective equilibrium: “[it] is not something separate from the method of reflective
equilibrium but, as I have emphasized, a crucial part of carrying out that method” (2014,
84–85). In other words, the question of whether or not certain judgements are considered
judgements is just another question that should be taken into consideration when engaging
in reflective equilibrium. It is thus a mistake to think that reflective equilibrium’s reliance
on considered judgements in any way undermines reflective equilibrium as a distinctive
and autonomous moral methodology, at least once we properly understand the way in
which considered judgements are to be selected.
Scanlon’s expansion of the scope of reflective equilibrium, to include questions about
the reasonability of judgements, thus does not seem to be susceptible to the objections
raised in the previous section, because it explicitly rejects the idea that we need a
determinate set of considered judgements independent of engaging in reflective
equilibrium. This proposal does, however, have some problems of its own.
First, it is unclear how considered judgements can lead to objective moral beliefs if the
24
There is, for instance, a striking similarity between Scanlon’s terminology and George Bealer’s (1996) theory of
intuitions as a priori, intellectual seemings.
REFLECTIVE EQUILIBRIUM
51
question of what counts as a considered judgement is included in the search for reflective
equilibrium. The reason for this is that there seems to be significant (normative)
disagreement about the conditions that are conducive to making good moral judgements.
For instance, for certain crude forms of utilitarianism the conditions under which good
judgements about morality are made are conditions of detached calculative rationality. For
virtue ethicists, on the other hand, good judgements can only be made under conditions in
which one is thoroughly involved in the practices about which a normative claim is being
made. For a contractualist like Scanlon, the appropriate conditions are conditions of
impartiality. It is therefore unclear how adding the question of what counts as a considered
judgement to the elements that should be taken into consideration can make it more likely
that there is convergence on a reflective equilibrium, or how reflective equilibria that are
obviously unreasonable from our specific normative perspective can be ruled out. Of
course, there might be less disagreement about what counts as the conditions under which
we make good judgements compared to disagreement concerning first-order normative
claims, but it seems unlikely that this kind of reduction of normative disagreement on the
level of what counts as a considered judgement is sufficient to safeguard maximal
objectivity.
Second, Scanlon’s defence of reflective equilibrium is an instance of a more common
expansionist tendency among defenders of reflective equilibrium to include just about any
kind of disagreement – be it normative, meta-ethical or otherwise – in their methodology.25
When someone asks the proponent of reflective equilibrium to give an account of when
judgements are reliable or reasonable, how one has to make choices when a conflict
between different beliefs arises, or which background theories are relevant when reflecting
on the truth of a moral belief, often the reply is that these questions should themselves be
settled by using the method of reflective equilibrium.
This kind of expansionism is problematic for several reasons. One problem is that by
including just about any possible disagreement related to the justification of our moral
beliefs in its methodology, reflective equilibrium runs the risk of becoming vacuous as a
method of moral justification, because ultimately reflective equilibrium will simply be
reduced to reasoning about ethics in general. That is, if any kind of disagreement is
included in the search for reflective equilibrium, it is not evident that it can still function as
25
Other contexts in which this happens is the context of possible biases that influence our normative judgements
(see e.g. Tersman 2008; Brink 2014, 55–56) and context of the question of how to delineate the moral domain in
the first place (Dorsey 2016, 772).
CHAPTER 2
52
a method of moral justification.
A second problem is that this kind of expansionism undermines the method’s
practicability (cf. Arras 2007, 55-56). One of the reasons for the widespread use of reflective
equilibrium, especially in the context of normative and applied ethics, is that it is a
practicable methodology. Reflective equilibrium is not just a meta-ethical position; it can be
employed to make substantive normative judgements and to apply these judgements to
actual practical problems. However, given the almost infinite number of considerations that
should be included in reaching a reflective equilibrium if we follow the expansionist
strategy, it seems that the method will never be able to reach a normative judgement. In
other words, the more considerations one has to include the less practicable the method
becomes.26
4.3 General Reflective Equilibrium
Until this point I have assumed that reaching a reflective equilibrium is essentially the
private accomplishment of one individual. That is, I have assumed that a moral claim is
justified when it is the outcome of a process of private reflection on all considerations
relevant to the issue at hand. Only in a second step, when we raise the question of whether
our own equilibrium converges with the equilibria of other persons, do other individuals
become important. This private or individualist reading of reflective equilibrium is
suggested by Rawls’ remark that “[he] shall not even ask whether the principles that
characterize one person’s considered judgments are the same as those that characterize
another’s” (1999b, 44).27 This individualist reading of reflective equilibrium gives rise to the
worries that the equilibria of different individuals might not converge and that a coherent
set of beliefs might still be no more than a reshuffling of moral prejudices.
Although the individualist interpretation of reflective equilibrium seems to be the
dominant interpretation of reflective equilibrium, one could also imagine an intersubjective
interpretation of reflective equilibrium. According to this intersubjective conception of
reflective equilibrium, a moral claim would only be justified if it is the outcome of a suitable
26
Arras therefore claims that when reflective equilibrium is actually employed by bioethicists, it will look more like
a narrow equilibrium than a wide equilibrium (2007, 56). This seems highly problematic, especially in light of all
the biases that are known to influence our moral judgements.
27
At the same time, Rawls continuously refers to “our considered judgments” (1999b, 20 my emphasis)
throughout A Theory of Justice. However, this reference to our considered judgements seems to be mainly
rhetorical.
REFLECTIVE EQUILIBRIUM
53
intersubjective process of reflection in which one not only asks whether a person can
coherently hold a certain claim but also whether others could coherently accept the same
claim as well. On the intersubjective interpretation, a moral claim cannot be justified by
private reflection but only by engaging in public reflection on all considerations relevant to
the issue at hand with all relevant others. This is the interpretation of reflective equilibrium
that is defended in Rawls’ later work. For instance, in A Reply to Habermas, Rawls claims
that “the overall criterion of the reasonable is general and wide reflective equilibrium”
(1995, 141). In a footnote to this remark, Rawls explains the idea of a general reflective
equilibrium in terms of the idea that “the same conception is affirmed in everyone’s
considered judgements” (1995, 141f16). And he therefore concludes that “reflective
equilibrium is fully intersubjective: that is, each citizen has taken into account the reasoning
and arguments of every other citizen” (1995, 141f16).
It seems that the idea of a general reflective equilibrium could answer both the
scepticism about convergence and the ‘garbage in, garbage out objection’. Reflective
equilibrium is more than just the reshuffling of moral prejudices, because we should
critically reflect on our prejudices by engaging in a public deliberation about our (moral)
claims. In addition, the scepticism about convergence and the corresponding worry of
relativism lose much of their force because there are no justified equilibria independent of
engaging in the activity of public reflection.
However, I think that the idea of a general reflective equilibrium faces a similar
problem to the reliance on considered judgements. The problem is that any idea of general
reflective equilibrium that succeeds in safeguarding moral objectivity will undermine
reflective equilibrium as a distinctive, coherentist moral methodology.
In order to illustrate this claim, it might be helpful to distinguish two ways in which the
intersubjective reflective equilibrium could be further developed. According to one possible
version of an intersubjective reflective equilibrium, moral justification would be constituted
by the actual acceptance of a set of moral beliefs through a process of reflection by all
relevantly affected persons. That is, a moral claim would be justified if, through a process of
public reflection, everyone actually agrees on the same set of coherent (moral) beliefs.
However, if actual acceptance is what matters, then the methodology is susceptible to a
variant of the ‘garbage in, garbage out objection’ which questions the importance of actual
agreement for moral justification. That is, one might wonder whether the fact that people
agree on moral issues is a reliable indicator of moral objectivity or reasonableness. In
addition, the scepticism about convergence objection reappears at a different point in
CHAPTER 2
54
theorizing because it remains unclear why we should believe that it is to be expected that we
will all agree on an intersubjective reflective equilibrium.
A more plausible version of reflective equilibrium therefore does not rely on the actual
acceptance of an intersubjective reflective equilibrium, but on its acceptability. By providing
an idealized account of the conditions under which public reflection and deliberation is
reasonable, both reasonableness and convergence could be guaranteed. However, although
in this way an intersubjective reflective equilibrium might not be susceptible to the
objections, its reliance on acceptability undermines reflective equilibrium as a distinctive
method of moral justification. Analogous to the problems associated with introducing the
notion of considered judgement, emphasizing acceptability and reasonable deliberation
implies that the most interesting part of the methodology is not the pursuit of a reflective
equilibrium but the justification of the criteria for reasonable deliberation, i.e. the
conditions under which a reflective equilibrium would be acceptable. In other words,
emphasizing the intersubjective nature of reflective equilibrium makes the methodology
parasitic on another method of moral justification.
Of course, this does not mean that the idea of acceptability should be rejected or that
the idea of an intersubjective reflective equilibrium should be dismissed. My point here is
only that this interpretation of reflective equilibrium can only get off the ground by
presupposing another, more fundamental, theory of moral justification, i.e. a method that
justifies a conception of acceptability. Intersubjective reflective equilibrium in terms of
acceptability is thus at best an incomplete method of moral justification.
The problem with the intersubjective interpretation of reflective equilibrium is thus
that it either does not succeed in vindicating objectivity (if the focus is on actual
acceptance) or presupposes another method of justification (if the focus is on acceptability).
5. Sacrificing Objectivity
5.1 The Only Game in Town?
I have argued that the method of reflective equilibrium cannot deliver objectivity, at least
while remaining a distinctive methodology. One might wonder whether the fact that a
method of moral justification, e.g. reflective equilibrium, cannot deliver certain outcomes,
e.g. moral objectivity, is by itself a reason to reject this methodology. I do not think that this
follows, because it would assume that the truth of some moral claims (e.g. the existence of
REFLECTIVE EQUILIBRIUM
55
certain objective moral claims) is an argument for the appropriate method of moral
justification. This is putting the cart before the horse. Here, I am therefore in agreement
with Scanlon, who claims that “the fact that the method of reflective equilibrium could lead
to a result that called into question the objectivity of our moral beliefs is not an objection to
that method” (2003, 153).
However, one implication of the current discussion is that we should be more aware of
the limited normative potential of the methodology: moral objectivity might be beyond
reach and the outcome of its procedure might be no more than a reshuffling of moral
prejudices. It is important to note that this does not mean that ‘anything goes’ when
adopting the method of reflective equilibrium, because, depending on the exact way one
spells out the requirements of coherence, moral claims can be shown to be unacceptable in
light of one’s larger belief set. The method of reflective equilibrium thus does allow for the
idea that someone can be mistaken about a moral claim and offers at least some resources
for the critical revision of one’s claims, even though it cannot guarantee that one’s revised
claims are any more reasonable than one’s initial moral claims
However, as I have argued above, on a widely accepted view of morality (and one that
is shared by most proponents of reflective equilibrium), moral claims are those beliefs one
should accept independent of the contingent beliefs or desires one happens to have
(whether or not these beliefs and desires are internally coherent). It is this type of
(maximal) objectivity that reflective equilibrium cannot account for.28 This suggests that
proponents of reflective equilibrium cannot have it both ways: reflective equilibrium either
leads to moral relativism or it turns out to be a form of intuitionism in disguise.
Ultimately, proponents of reflective equilibrium suggest that we should adopt reflective
equilibrium, even if this means giving up on moral objectivity. For instance, in a recent
paper, Michael DePaul claims that there is simply no reasonable alternative to the method
of reflective equilibrium, claiming that
it is natural to want certainty or something else that is more secure than what
seems true upon reflection. But we are not destined to have such things, at least
not for very much of what we believe. The best we can do is think things through
and trust the conclusions we reach. The defender of reflective equilibrium calls it
28
As I have noted above, there are different ways to flesh out the idea of maximal objectivity (depending on
whether or not one is a moral realist), but this should not concern us here.
CHAPTER 2
56
like it is and says what is good enough. No more hankering after what cannot be
(2007, 618).
In a similar way, Kenneth Walden recently claimed that “it is the only conception of
normativity that we are left with once we discover that a kind of absolutism is
impracticable. It is the answer we are left with when we find that we cannot divine the
future or use some inner sense to tell right from wrong” (2013, 254). Scanlon states that “it
is the only defensible method: apparent alternatives to it are illusory” (2003, 149). Jonathan
Floyd claims that “the ultimate defence of Rawls’ method ... is that unless we can construct
an alternative, together with a convincing argument regarding its superiority, we should
just ‘keep calm and carry on’”(forthcoming, 12). Finally, Rawls stresses that “at any given
time, we cannot do better than that [i.e. use reflective equilibrium]” (2003, 31).
These philosophers all argue for the view that ultimately reflective equilibrium is ‘the
only game in town’: it is the only available method that is practically useful. I think that
there are several problems with the only game in town argument. First, I think that as a
matter of fact not many proponents of reflective equilibrium are willing to commit
themselves to moral relativism. Although Scanlon, for instance, claims that the implication
of relativism would not be an argument against the method of reflective equilibrium, he
continues to defend reflective equilibrium as a method that can justify objective moral
claims. One might think that the later work of Rawls would be an example of reflectiveequilibrium relativism. In his later work, starting roughly with the publication of ‘Justice as
Fairness: Political not Metaphysical’, Rawls does indeed relativize the validity of his theory
of justice to “existing democratic societies” (Rawls 1985, 225). 29 However, even after his socalled political turn, Rawls claims that there are at least some objective moral principles,
even if justice as fairness is only valid relative to already existing democratic societies. In
The Law of Peoples, Rawls, for instance, defends a (minimal) set of human rights which are
supposed to be “binding on all people and societies” (2000, 80) (i.e. universal) and which
are valid “whether or not they are supported locally” (i.e. categorical) (2000, 80).30 Of
course, this does not show that one cannot be a reflective-equilibrium relativist, but that
29
I do nor want to take a stance on the exegetical question of the extent to which there is (dis)continuity between
the early Rawls of A Theory of Justice and the publication of (the papers leading up to the publication of) Political
Liberalism. For a good overview of different positions regarding the (dis)continuity in Rawls’ work, see Laden
(2003).
30
And he explicitly claims that these human rights can be justified in reflective equilibrium (Rawls 2000, 86,
86f32).
REFLECTIVE EQUILIBRIUM
57
accepting the relativist implications of reflective equilibrium will have important
implications for both the self-understanding of proponents of reflective equilibrium and for
the first-order moral theories they defend (consider, for instance, Rawls’ reliance on human
rights). It might therefore be wise to be careful to conclude that there is no alternative to
reflective equilibrium (assuming that I am right that reflective equilibrium leads to
relativism).
Second, and more importantly, I think that proponents of reflective equilibrium have
simply failed to show that reflective equilibrium is indeed the only game in town. Insofar as
proponents of reflective equilibrium discuss alternative methodologies, it is mostly limited
to criticizing intuitionism. 31 The implicit assumption of proponents of reflective
equilibrium is thus that we have to choose between foundationalism, more specifically
intuitionism, or the coherentism of reflective equilibrium. Given the known problems with
intuitionism, we should adopt reflective equilibrium, even if this means giving up on moral
objectivity. Or so the argument seems to go. This argument, however, overlooks the fact
that there are alternative methods of moral justification.32 One often-overlooked alternative
is transcendental argumentation. In contrast to intuitionism, transcendental arguments are
seldom discussed by proponents of reflective equilibrium.33 Before biting the bullet of
relativism, we should therefore first investigate whether this alternative, transcendental
methodology can be made to work. This is the project that I take up in the rest of this thesis.
5.2 What is Wrong with Moral Relativism?
Before discussing transcendental arguments, however, we might wonder why we should not
just settle for moral relativism, i.e. the view that the correctness of moral judgements
31
For references see section 2.
Another problem with the argument is that it does not provide a positive reason to adopt reflective equilibrium.
That is, it is unclear why we should adopt reflective equilibrium and not, say, global scepticism about morality,
which denies that there are any justified moral beliefs at all.
33
There are a few exceptions. Rawls (1999a) briefly discusses the alleged failure of Kant’s transcendental deduction
of the categorical imperative. More recently, Scanlon claims that he is “not convinced by any argument I have seen
for the claim that we must see requirements as binding on us insofar as we see ourselves as acting at all” (2012,
238). However, both fail to systematically engage with recent attempts to provide transcendental arguments in
ethics, nor do they discuss the respective strengths and weaknesses of transcendental arguments and the method of
reflective equilibrium. In all fairness, the same could be said about proponents of transcendental arguments with
respect to the method of reflective equilibrium. Although proponents of transcendental arguments reject reflective
equilibrium, they typically only discuss reflective equilibrium in passing (see e.g. Illies 2003, 13; O’Neill 1996, 47–
48).
32
CHAPTER 2
58
somehow depends on a contingent standard, such as one’s motivational set, the specific
socio-historical context in which one lives or one’s conceptual scheme.34 Why even bother
trying to vindicate moral objectivity? What is so wrong with moral relativism?
I have already briefly discussed in the introduction some reasons for taking moral
objectivity as a provisional starting point. Now that I have argued, very roughly, that
proponents of reflective equilibrium have to make a choice between moral relativism and
transcendental argumentation, it might be necessary to say a bit more about the reasons for
not giving up on moral objectivity too easily, especially in light of the fact that in recent
years moral relativism has been gaining popularity in mainstream ethics (see e.g. Street
2009; Velleman 2013; Rovane 2013).
The aim of authors such as Street, Velleman and Rovane is to show that moral
relativism is a more respectable position than is often assumed and that some of the more
common objections to moral relativism are based on a caricature of the moral relativist
position. For instance, moral relativists deny that accepting moral relativism would lead to
problematic cases of so-called faultless disagreement, the idea that when two communities
disagree about a moral issue, say the (im)morality of Nazism, they might both be right from
their respective perspective (see in particular Velleman 2013, 2).35
I cannot do justice to these nuanced defences of moral relativism here. The point I
want to make, however, holds independent of the details of these accounts. The point is that
however nuanced a position moral relativism can be, it should still only be accepted once
we are absolutely sure that moral objectivity is beyond reach. The reason for this is that, as
these defenders of moral relativism themselves explicitly acknowledge, the validity of their
theories of moral relativism crucially hinges on a rejection of accounts of moral objectivity
(or so I argue below). Moral relativism, in other words, is first and foremost a negative
theory, or a ‘second-best’ theory.36 So to give an answer to the subtitle of this subsection, I
do not think that there is necessarily anything wrong with moral relativism, but I think that
relativism should only be accepted once we conclude that accounts of moral objectivity fail.
34
See e.g. Gilbert Harman’s definition of relativism: “there is no single true morality. There are many different
moral frameworks, none of which is more correct than the others” (Harman and Thomson 1996, 5).
35
There are also defenders of relativism as faultless disagreement. Consider, for instance, Gilbert Harman, who
defines moral relativism as the view which holds that “conflicting judgments can be equally correct or equally
justified” (1983, 308).
36
Note that moral relativism must also include a positive thesis in order to distinguish moral relativism from, for
instance, moral scepticism. A moral sceptic denies that there are any justified moral judgements. A moral
relativist, on the other hand, claims that moral judgements can be justified, even though there is no objective
standard. I briefly come back to this below.
REFLECTIVE EQUILIBRIUM
59
Let me illustrate this point by very briefly discussing two recent accounts of relativism, the
accounts put forward by Sharon Street and David Velleman.37 I will not try to evaluate their
positions. Instead, I will briefly describe the main features of their relativist position, and
subsequently I will show how both positions ultimately hinge on a rejection of moral
objectivity.
Sharon Street defends a form of moral relativism according to which “the substantive
content of an agent’s normative reasons is a function of his or her particular, contingently
given, evaluative starting points” (2012, 41). Street claims that all normative judgements,
including moral judgements, are only correct relative to the contingent evaluative starting
points of an individual. As I already noted in the introduction, Street calls her view
Humean constructivism because she holds that moral judgements are mind-dependent
(constructivism) and denies that there are any universal and categorical judgements (this is
what makes her constructivism Humean and not Kantian).
Street, however, claims that her Humean constructivism does not necessarily have to
lead to problematic cases of faultless disagreement. She argues, for instance, that even
though on her account moral standards are relative to our contingent evaluations, human
beings typically share certain contingent evaluations, i.e. there are contingent yet universal
evaluations (Street 2009). In addition, she claims that what is important is not simply what
an individual actually accepts as a moral judgement, but what she should coherently accept
as a moral judgement (given her evaluations). So although faultless disagreement on
Street’s account is in principle possible, Street claims that in practice it will be rare because
of the evaluations we contingently happen to share.
Velleman (2013) defends a different kind of moral relativism. For Street it might still be
possible that two individuals can disagree about a moral issue and that no one is at fault,
because they are both right from their own contingent evaluative starting point (even
though Street thinks this does not happen very often in practice). Velleman, however,
explicitly rejects the possibility of “faultless disagreement” (2013, 2). Instead, he defends
moral relativism by referring to the fact that different communities have different
conceptual schemes, which, according to Velleman, implies that they might simply be
talking at cross purposes rather than disagreeing with each other (which requires that they
talk about the same subject matter). Velleman claims that “disagreement about morality is
disagreement about what may or may not be done, and so it requires agreement about what
37
As I already noted above, in her recent work Street (2016) is more reluctant to accept relativism.
CHAPTER 2
60
is doable. For communities with different domains of doables, the question what may or
may not be done is therefore mute” (2013, 25). Velleman uses ‘doables’ not to refer to what
is feasible within a certain community but to the kinds of actions that are conceivable in the
first place. One of the main examples that Velleman puts forward of a non-universal doable
is the case of Russian-speaking communities that have a form of untruthfulness called
‘vranyo’, which does not have an equivalent in the English-speaking world. It cannot, for
instance, be reduced either to a lie or to a joke (Velleman 2013, 73–74). Velleman claims
that, in cases that invoke a concept which is not universally shared, moral disagreement is
impossible because there is no neutral, shared-act description: this kind of untruthfulness is
conceivable as an action in certain communities but not in others.38 At the same time, this
is a form of relativism because one can only (il)legitimately perform this kind of
untruthfulness in a specific community and not in others.
Again, my point is not to critically evaluate the details of these positions. But one might
wonder, for instance, whether Street is right in thinking that human beings contingently
share sufficient evaluations so that cases of faultless disagreement are rare. I think there is
less convergence of, say, human rights or justice than Street seems to suggest.39 Also, one
might wonder how successful Velleman’s reliance on the idea of different conceptual
schemes is: although Velleman might be right that in certain cases we are simply talking at
cross purposes with other communities (e.g. in the example of untruthfulness), this does
not seem to apply to all, or even most, moral questions. For instance, when thinking about
‘justice’ or ‘harm’ it seems more likely that different communities will have a substantive
disagreement about what constitutes justice or harm, rather than talking at cross purposes.
When X and Y disagree about whether or not homosexual acts constitute a criminal
offence, X and Y are not talking at cross purposes because they don’t share the concept of
38
A similar account of moral relativism is defended by Carol Rovane, who defends what she calls
‘multimundialism,’ i.e. the view “that there are many incomplete bodies of truths that cannot be embraced
together” (Rovane 2008, 45; see also Rovane 2013).
39
For references see section 3 of this chapter. In all fairness, I should stress that Street’s discussion is mainly
focused on what she calls “ideally coherent eccentrics”, such as Hume’s example of a person who would prefer the
destruction of the whole world to his finger being scratched; Rawls’ example of a grass counter, a person whose
sole pleasure it is to count blades of grass; Gibbard’s example of an ideally coherent Caligula whose only aim it is
to maximize the suffering of others; and finally, Parfit’s example of a person with Future Tuesday Indifference,
that is, a person who always prefers agony on a future Tuesday to trivial pain on any other day of the week (Street
2009, 273). Street’s basic claim is that these kinds of persons are philosophers’ fiction and that real people are not
like this. This is supposed to show that the inability of her Humean constructivism to claim that these characters
are irrational or immoral is not a real problem. This, however, neglects that the really troubling cases for Humean
constructivism are not philosophers’ fictions but real people who, for instance, deny human rights to certain
minorities.
REFLECTIVE EQUILIBRIUM
61
justice; rather, they have different conceptions of that very same concept. It might well be
the case that when X and Y are talking about lying and untruthfulness, they are talking at
cross purposes, because one has the concept of vranyo whereas the other does not. But
Velleman gives no proper defence of the much stronger claim that examples of the sort he
discusses would carry over to the more fundamental moral concepts like justice or harm.
However, even if these positions are valid on their own terms, there is still an important
reason to give priority to the idea of moral objectivity. The reason for this is that moral
relativism is always only a second-best option. By saying that moral relativism is always a
second-best option, I mean that any argument in favour of moral relativism necessarily
includes a negative argument, i.e. an argument against the possibility of providing a
plausible account of moral objectivity
Moral relativists argue for moral relativism by offering an argument by elimination,
that is, moral relativism is vindicated because by elimination no other position can be
shown to be plausible. For instance, Street claims that her position is
defined negatively in contrast to Kantian constructivism: it is characterized by a
general skepticism about arguments which seek to establish moral values as
following from the standpoint of “agency” as such. Therefore, the obvious way,
and perhaps the only way, to argue for Humean as opposed to Kantian
constructivism is negatively—by examining the best attempts to establish such
substantive values and showing why they fail (2012, 45).
And although Velleman partly rests his case for relativism on an argument about the
existence of a plurality of conceptual schemes, his defence of relativism is crucially based on
his rejection of moral objectivist views. He claims that “no one has ever succeeded in
showing any one set of norms to be universally valid” (2013, 45). He states: “I am a moral
relativist ... [because] I deny that there are universal norms of any kind, and that there are
necessarily ubiquitous norms of morality” (2013, 94; see also Harman and Thomson 1996,
8–11). Moreover, consider Harman’s description of relativism: 40 “moral relativism argues
that there is no objective way to establish that a particular morality is the correct morality,
40
He makes similar statements in his classic ‘Moral Relativism Defended’, where he claims, “I simply assume that
the Kantian approach is wrong ... In other words, I assume that the possession of rationality is not sufficient to
provide a source for relevant reasons, that certain desires, goals or intentions are also necessary” (Harman 1975,
9).
CHAPTER 2
62
and concludes that there is no reason to believe in a single true morality” (2015, 860). The
important point here is that even moral relativists acknowledge that moral objectivity is the
default conception of morality, in the sense that moral objectivity has to be disproven in
order to argue for moral relativism.41 This means that the question of whether or not moral
objectivity can be vindicated is of philosophical interest, whether or not one actually thinks
that moral objectivity is something worth wanting in the first place.
Of course, I do not want to deny that moral relativism is also a positive position.42
Moral relativism is a positive position insofar as it does not deny (pace global moral
scepticism) that there are any justified moral claims. A moral relativist claims that moral
claims can be justified, but stresses that the validity of moral claims is only valid relative to a
contingent standard. The dialectical position of moral relativism is thus roughly as follows:
a moral relativist has to show, negatively, that there are no objective moral claims (pace
moral objectivism). A moral relativist has to show, positively, that there are nevertheless
justified moral claims (pace moral scepticism).
What I have tried to show is that even if one does not agree that moral objectivity is
something worth wanting, it is important to analyse the possibility of vindicating moral
objectivity, because alternative, relativist conceptions of morality can only be justified by an
argument from elimination. The question is thus if there are any objective moral claims and
how these could possibly be justified. This is the question I will turn to in the following
chapters.
6. Conclusion
In this chapter, I have critically evaluated the most recent attempts to defend the method of
reflective equilibrium against common objections to the methodology. I have argued that
proponents of reflective equilibrium still fail to show that their methodology can safeguard
moral objectivity. In addition, I have argued that reflective equilibrium is not the only game
in town once intuitionism has been discarded. Moreover, I have suggested that we should
not be too easily satisfied with settling for moral relativism, because even if one is willing to
41
A comparison with error theory might be helpful. Error theorists deny that there are any justified moral claims.
Error theorists argue for this sceptical position by showing that existing justifications of morality fail (for an
overview see Joyce 2015). Error theory is thus first and foremost a negative position. I think that moral relativism
is a negative position in a similar way (but see the next paragraph in the text).
42
This is therefore where the analogy with error theory collapses.
REFLECTIVE EQUILIBRIUM
63
face the potential normative costs of moral relativism, the only argument for relativism is
an argument against the possibility of vindicating moral objectivity.
For all these reasons, proponents of reflective equilibrium should be interested in the
question of whether an alternative, transcendental methodology can be made to work,
before they accept moral relativism. This is the question I will turn to next.
Chapter 3
Transcendental Arguments
1. Introduction
The question we now face is this: is there a moral methodology that can succeed where
reflective equilibrium fails, i.e. is there a method which can vindicate moral objectivity
(pace reflective equilibrium) without relying on self-evident moral truths (pace
intuitionism)?
This chapter introduces transcendental arguments as such an alternative moral
methodology. Transcendental arguments can be contrasted with both intuitionism and
reflective equilibrium. Intuitionism holds that there are self-evident moral truths, so a
moral claim is either self-evidently true or can be derived from such a truth. The method of
reflective equilibrium claims that moral justification is constituted by the coherence
between our contingently held moral and non-moral beliefs. Transcendental arguments, on
the other hand, try to show that anyone ought to accept certain moral principles given
certain inescapable aspects of a person’s self-understanding, where ‘self-understanding’
refers to one’s first person or second- person perspective as, for instance, an agent or as
someone who is engaged in certain types of interaction.
The goal of this chapter is twofold. The first goal is simply to introduce this
transcendental approach to ethics. Although different transcendental arguments have been
put forward in ethics, there is comparatively little reflection on what makes an argument in
ethics a transcendental argument.
1
Before assessing the plausibility of specific
1
One important exception is Illies (2003). In addition, there is a growing literature on so-called constitutive
arguments in meta-ethics (Tubert 2010; Katsafanas 2013), which, as I will argue below, could be understood as a
type of transcendental argument.
CHAPTER 3
66
transcendental arguments, it is therefore helpful to get an idea of the general structure of
the argument. An account of the general structure of a transcendental argument in ethics,
independent of more specific arguments, allows one, for instance, to distinguish between
challenges that apply to all transcendental arguments in ethics and challenges that only
apply to specific transcendental arguments.
The second goal of the chapter is to show that some of the objections to transcendental
arguments, which go back to Stroud’s (1968) paper on the topic, are merely reasons against
transcendental arguments that are concerned with the external world, not with morality.
The reason for discussing Stroud’s objection is that his objection is often taken to
undermine transcendental arguments in general, even though Stroud puts forward his
objection in the specific context of transcendental arguments in theoretical philosophy
which aim to justify claims about the external world. The reason why Stroud’s objection is
taken to undermine transcendental argument in general has to do with the fact that in the
Stroudian-inspired literature on transcendental arguments, typically no distinction is made
between transcendental arguments in theoretical philosophy and those in ethics.2
The underlying goal of this chapter is thus to show that, on the one hand,
transcendental arguments in ethics should be understood against the background of a more
general transcendental approach to philosophy and, on the other hand, that the differences
between transcendental arguments in ethics and theoretical philosophy are significant and
need to be kept in mind. Very roughly, transcendental arguments in theoretical philosophy
try to justify (true) beliefs about the external world. Transcendental arguments in ethics,
however, do not justify beliefs about the external world but rather are meant to justify
moral commitments. Because I am ultimately interested in the plausibility of
transcendental arguments in ethics, and more specifically the question of whether
transcendental arguments can justify moral claims, I will not take a stance on the
plausibility of transcendental arguments in theoretical philosophy. Instead, my aim is to
show that common objections to transcendental arguments in theoretical philosophy do
not apply to transcendental arguments in ethics, at least insofar as they are used to justify
Kantian constructivist conclusions, and that transcendental arguments in ethics should
therefore be assessed on their own terms. This is the goal for the remainder of the thesis.
The structure of the chapter is as follows: first, I briefly describe the main features of
transcendental argumentation. In section 3, I discuss Stroud’s objection to transcendental
2
For exceptions see section 3.
TRANSCENDENTAL ARGUMENTS
67
arguments and argue that the objection only applies to a subclass of transcendental
arguments, so-called strong transcendental arguments for ontological conclusions. In
section 4, I argue that transcendental arguments used to justify moral realist accounts of
objectivity are susceptible to Stroud’s objection and, consequently, in section 5, I argue that
transcendental arguments for Kantian constructivist objectivity are not susceptible to this
objection.
2. What is a Transcendental Argument?
As I have already mentioned in the introduction to this thesis, it is difficult to define
transcendental arguments in such a way that the definition applies to all transcendental
arguments that have been put forward, especially if this definition is meant to apply to both
transcendental arguments in theoretical philosophy and transcendental arguments in
ethics. Until now, I have defined a transcendental argument as a kind of argument which
starts from an inescapable aspect of someone’s self-understanding and which subsequently
explores the necessary conditions of the possibility of this self-understanding, where selfunderstanding refers either to the first person or the second-person perspective, such as
one’s self-understanding as an agent (first person) or the self-understanding of someone
who engages in certain types of interaction (second person). To some extent this
description might be illuminating because it makes clear what a transcendental argument is
not: a transcendental argument is not a third-person description of some fact or the logical
analysis of a concept. At the same time, this definition raises more questions than it
answers, because it is not obvious what it means to argue from the first person or second
person.
In this section, I try to start clarifying what a transcendental argument is. I do so by
using an analogy to illustrate how a transcendental argument functions, the famous
Epimenides paradox of a Cretan who claims that all Cretans are liars.3 Subsequently, I try to
generalize from this example to the structure of transcendental arguments in general. In
this discussion, my focus is mainly on transcendental arguments from the first person,
which means that not everything I say about transcendental arguments in general
3
For an overview of different discussions of this liar paradox, see Beall and Glanzberg (2014).
CHAPTER 3
68
necessarily applies to transcendental arguments from the second person.4 I will briefly come
back to transcendental arguments from the second person towards the end of this section,
but a comprehensive discussion of transcendental arguments from the second person will
have to wait until chapter 5. With these qualifications in mind, let me now turn to the
Epimenides paradox.
Imagine a Cretan who claims that ‘all Cretans are liars’. The statement ‘all Cretans are
liars’ is, taken by itself, not necessarily unintelligible (we can understand the sentence) or
necessarily (empirically) false. There is, however, something odd happening if the
statement ‘all Cretans are liars’ is made by a Cretan. The problem is that the Cretan, by
saying that all Cretans are liars, is stuck in a self-contradiction. After all, an implication of
the statement that all Cretans are liars is that the Cretan who makes the statement is also a
liar. However, if the Cretan is a liar, the statement that all Cretans are liars cannot be true.
The statement that all Cretans are liars can only be true if the Cretan is not a liar, but in that
case the statement would be false. This means that the Cretan cannot coherently claim nor
believe that all Cretans are liars, because this statement cannot be true if he is himself a
Cretan.5
A Cretan who states that all Cretans are liars thus contradicts himself insofar as the
making of the statement contradicts what is implied by the statement itself. The
Epimenides paradox is one example of how self-contradictions can arise from the firstperson perspective. This self-contradiction can be further explained in different ways, and
in what follows I focus on an explanation which stresses the importance of the first-person
perspective.
It is important to stress that the statement ‘all Cretans are liars’ only leads to a selfcontradiction if this claim is understood as being made by a Cretan. As a Cretan, the Cretan
cannot coherently claim that all Cretans are liars. This means that the self-contradiction
can only arise if we understand the statement ‘all Cretans are liars’ from the first-person
4
This depends to what extent arguments from the second person could be reformulated in terms of the firstperson perspective of someone who engages in certain forms of interaction, or whether an irreducible secondperson standpoint exists.
5
This does not mean that the Cretan cannot actually say that all Cretans are liars (because obviously he can say it)
but that he ought not to say it. The reason why he ought not to say it is that he becomes unintelligible both to
himself and to others (I will come back to the nature of this ‘ought’ below and in the next chapter when I discuss
Sharon Street’s objection to Kantian constructivism). In addition, the Cretan might simply be making a joke, in
which case he would also not be contradicting himself. The self-contradiction arises only if the Cretan aims to say
something true.
TRANSCENDENTAL ARGUMENTS
69
perspective of the Cretan who makes the statement.6 Independently of the first-person
perspective of the Cretan, there is nothing contradictory in claiming that all Cretans are
liars. Of course, this does not mean that the self-contradiction can only be described from
the first-person perspective – in fact, I have just given a third-person description of the
Cretan. The point is rather that the reason why there is a contradiction in the case of the
Cretan can only be understood by appreciating that a Cretan, from his first-person
perspective as a Cretan, cannot coherently claim that all Cretans are liars.
Transcendental arguments work in a similar way: they try to show that certain claims
or judgements cannot be accepted or rejected without one contradicting oneself. In this
sense, a transcendental argument is thus a self-reflexive argument: it is an argument which
tries to show what a person necessarily has to accept or reject insofar as he or she
understands herself in a certain way (Düring and Düwell 2015). The defining feature of a
transcendental argument, whether it is employed in theoretical philosophy or in practical
philosophy, is thus that it proceeds from the internal, first-person perspective (or from the
internal, second-person perspective). There are, however, two important differences
between the example of the Cretan and transcendental arguments. First, instead of focusing
on particular statements (such as the statement that all Cretans are liars), transcendental
arguments typically focus on experience, speech, thought or action in general. 7 The
difference between the example of the Cretan and transcendental arguments is that,
whereas only a Cretan cannot without self-contradiction claim that all Cretans are liars, the
conclusions of a transcendental argument are supposed to apply to all agents, thinkers,
speakers or beings that have experiences.
The second difference is that whereas the self-contradiction involved in the case of the
Cretan can be easily recognized because his making the statement that all Cretans are liars
directly contradicts the content of the statement, transcendental arguments typically try to
show that there are self-contradictions that are less visible. This is done by showing,
6
The same point could be made in more second-personal terms by stressing the (performative) contradiction
between the semantic content of the statement of the Cretan and the ‘discursive commitments’ of the Cretans as a
language user (Heath 2003, 295). In other words, from a second-person perspective one might point out that there
is a contradiction between the content of his statement, i.e. that he is a liar, and his making this statement, because
making a statement commits one to telling the truth.
7
Sometimes a distinction is made between so-called global transcendental arguments and local transcendental
arguments (cf. Stern 2000, 82). A local transcendental argument concerns the conditions of the possibility of a
particular practice. A global transcendental argument, on the other hand, explores the conditions of the possibility
of experience, thought or action as such. Following this distinction, showing the self-contradiction involved in the
case of the Cretan could be labelled as a local transcendental argument. In this thesis, I use transcendental
arguments to refer to global transcendental arguments.
CHAPTER 3
70
typically in a sequence of argumentative steps, that an experience, thought or action has
certain, possibly non-obvious and non-trivial, necessary conditions of possibility, which,
insofar as one is engaged in action, or insofar as one is thinking or having experiences,
should necessarily be accepted on pain of self-contradiction (i.e. on contradicting that one
is an agent, thinker or experiencer).
Consider, for instance, Descartes’s famous cogito argument. Descartes claims that
existence is a necessary condition of the possibility of thinking and that because he thinks
he must exist:
1) I think.
2) Existence is a necessary condition of the possibility of thinking.
3) Therefore, I must (believe that I) exist.
If it is true that existence is a necessary condition of the possibility of thinking, Descartes,
and any other thinker, cannot without self-contradiction deny that he exists, at least insofar
as he is thinking.
A prominent example of a transcendental argument in ethics is Christine Korsgaard’s
(1996) transcendental argument for the value of humanity. Korsgaard summarizes her
transcendental argument as follows:
Since you are human you must take something to be normative, that is, some
conception of practical identity must be normative for you. If you had no
normative conception of your identity, you could have no reasons for action, and
because your consciousness is reflective, you could then not act at all. Since you
cannot act without reasons and your humanity is the source of your reasons, you
must value your humanity if you are to act at all (1996, 123).
This argument raises many questions: what does Korsgaard mean by the value of
humanity? Why can only your humanity function as the source of reasons? Why must you
take something to be normative for you? I will come back to these questions in chapter 6
when I discuss the details of Korsgaard’s argument on the value of humanity, but here I just
want to focus on the transcendental structure of her argument. Korsgaard’s transcendental
argument can be roughly summarized as follows:
TRANSCENDENTAL ARGUMENTS
71
1) Agency is inescapable for A (i.e. you, me or any other human being).
2) Valuing your (own) 8 humanity is a necessary condition of the possibility of
agency.
3) Therefore, A must value his or her (own) humanity.
Korsgaard’s claim is thus that although it is not immediately obvious, valuing your (own)
humanity is a necessary condition of the possibility of agency. And given that agency is
somehow inescapable, you must value your (own) humanity.
In general, a transcendental argument is thus an argument which aims to show that a
possibly non-obvious and non-trivial Y is a necessary condition of the possibility of X –
where, given that X is somehow inescapable for A, it logically follows that A must accept Y.
A transcendental argument thus has the following structure:
1) X is inescapable for A.
2) Y is a necessary condition of the possibility of X.
3) Therefore, A must accept Y (e.g. ‘(believe that) one exists’ or ‘value one’s
humanity’).
In what follows, I will call the X that is inescapable for A the ‘starting point’ of a
transcendental argument; Y is the ‘transcendental commitment’, which follows from the
‘transcendental conditional’ (Y is a necessary condition of the possibility for X).9
Note that this general definition of transcendental arguments allows for many different
applications (as the examples of Descartes and Korsgaard already illustrate). First, the
8
I add ‘own’ in brackets in order to emphasize that this first step of Korsgaard’s argument is only meant to show
that one has to value one’s own humanity, and not yet that one also has to value the humanity of other agents (this
is shown in a second step of the argument). I briefly distinguish the different steps in Korsgaard’s argument in
chapter 4. In chapters 5 and 6 I will critically discuss different attempts by Korsgaard to justify the claim that one
also has to value the humanity of other agents.
9
Christian Illies (2003) makes a distinction between an explorational and a retorsive transcendental argument. An
explorational argument starts from the existence of some fact and argues to the necessary conditions of the
possibility of this fact. A retorsive transcendental argument, on the other hand, tries to show that denying the
transcendental commitment constitutes a performative self-contradiction because its denial rests on certain
necessary presuppositions that contradict the asserted proposition. The general characterization of transcendental
arguments put forward in this section aims to illustrate that explorational and retorsive transcendental argument
are not two alternative types of transcendental arguments but are two sides of the same coin: a transcendental
argument explores the necessary conditions of the possibility of an inescapable feature of our practical selfunderstanding, which implies that a denial of the transcendental commitments constitutes a self-contradiction
(insofar as it contradicts an essential feature of one’s self-understanding).
CHAPTER 3
72
argument could be applied to different starting points, such as thinking, experiencing,
acting or speaking, as long as it can convincingly be shown that it forms an inescapable
aspect of our self-understanding.10 Second, transcendental arguments can be used to argue
for a plurality of claims: claims about the external world and the existence of other minds,
but also claims about rational principles such as the principle of non-contradiction, or
moral principles such as the categorical imperative. Third, the definition allows for
different interpretations of the status of the transcendental commitment: is the
transcendental commitment supposed to say something about the world or only something
about our conceptual scheme, i.e. can or should transcendental arguments justify
ontological claims (such as Descartes’s cogito argument) or ‘merely’ make claims about
what we should believe or what we should do (e.g. valuing your humanity)?11 These details
will be filled in differently by specific transcendental arguments and, as I argue below, how
one fills in the details will impact the plausibility of a transcendental argument.
Because of these different possible applications, it should not be surprising that
transcendental arguments have been put forward both in theoretical philosophy and in
practical philosophy. Typical examples of transcendental arguments in the history of
theoretical philosophy include Aristotle’s argument for the law of non-contradiction,
Descartes’s cogito argument and Kant’s transcendental deduction of the categories.
Examples of transcendental arguments in more recent theoretical philosophy can be found
in the work of Ludwig Wittgenstein, Peter Strawson, Hillary Putnam and Donald
Davidson.12
Transcendental arguments in ethics can arguably be traced back to the third section of
Immanuel Kant’s (1998) The Groundwork of the Metaphysics of Morals (from now on: The
Groundwork). There have been several attempts in the second half of the twentieth century
at formulating transcendental arguments in ethics (for references see below). In addition, in
10
This raises the question of what it means for a starting point to be inescapable. I will come back to this question
in the next chapter.
11
Sometimes transcendental arguments are more specifically defined in terms of the ambition that they will lead to
true beliefs. See e.g. Quassim Cassam’s definition of the form of a transcendental argument: “there is experience; a
necessary condition of the possibility of experience is the truth of P; therefore P” (Cassam 1987, 355). I think the
problem with this definition is that it excludes the possibility of what I’ll refer to as ‘modest’ transcendental
arguments (see below) and, more importantly, it excludes most, if not all, transcendental arguments in ethics as
transcendental arguments (see section 4.2).
12
Reading Wittgenstein’s private language argument, for instance, as a transcendental argument is not
uncontroversial. The same applies to most other transcendental arguments mentioned in the text. The point of
mentioning these arguments is only to illustrate the wide range and diversity of arguments that are typically
considered to be transcendental arguments.
TRANSCENDENTAL ARGUMENTS
73
meta-ethics there is growing interest in so-called constitutive arguments (for an overview
see Tubert 2010; Katsafanas 2013). In a recent review article, Ariela Tubert defines
constitutivism as follows: “the basic idea behind constitutive arguments is that there are
certain features constitutive of belief or action, which guarantee that insofar as we act or
believe at all, we are committed either to following principles like the law of noncontradiction or the instrumental principle, or to aiming at something like truth or the
good” (Tubert 2010, 656). On Tubert’s definition, constitutivism is just another name for a
transcendental argument. Throughout the thesis, I therefore assume that constitutive
arguments are transcendental arguments.
There are important differences between different attempts at putting forward a
transcendental argument in ethics. The first concerns moral metaphysics. Although most
transcendental arguments in ethics have been put forward by Kantian constructivists, there
are also several attempts to justify moral realism on the basis of a transcendental arguments
(Illies 2003; Enoch 2011b). Later in this chapter, I will argue that transcendental moral
realism should be rejected because transcendental arguments, given their first-personal
nature, cannot lead to the justification of mind-independent moral facts.
There is also an important distinction to be made between different transcendental
arguments that are put forward by Kantian constructivists. Authors such as Karl-Otto Apel
(1975; 1980), Jürgen Habermas (Habermas 1990a; 1994), Onora O’Neill (1986), Stephen
Darwall (2006) and, on one possible interpretation of her work, Christine Korsgaard (1996;
2009) have attempted to justify objective moral principles on the basis of the necessary
preconditions of certain types of interaction such as communication, argumentation or
reasoning or from an irreducible second- person perspective. Their claim is thus that
engaging in a certain type of interaction is inescapable for A and that therefore A is
committed to whatever are the necessary preconditions of this type of interaction.
Others, such as Alan Gewirth (1978) and, on another possible interpretation of her
work, Korsgaard (1986; forthcoming) have attempted to justify objective moral principles
in terms of the necessary preconditions of agency. They claim that it is inescapable for A to
understand herself as an agent, that A is therefore committed to whatever are the necessary
preconditions of her own agency and that these preconditions include a categorical and
universal principle of interpersonal morality. In this thesis, I put forward a qualified
defence of the view that moral claims can be defended as the necessary preconditions of
agency (and not interaction). In the remainder of this chapter and in the next chapter, I will
CHAPTER 3
74
therefore focus exclusively on transcendental arguments from the first person. I will come
back to transcendental arguments from the second person in chapter 5.
3. Transcendental Arguments in Theoretical Philosophy
As I have already mentioned in the introduction to the thesis, transcendental arguments are
a fringe phenomenon in ethics. A possible reason for this could be that recent discussions
about transcendental arguments in theoretical philosophy have been largely negative13 and
that in these discussions typically no distinction is made between the plausibility of
transcendental arguments in theoretical philosophy and in practical philosophy.14 In this
section, I briefly discuss the reasons for scepticism about certain transcendental arguments
in theoretical philosophy. The aim of this discussion is not to take a stance in these debates,
but merely to show, in the remainder of the chapter, that whatever are the objections to
(some) transcendental arguments in theoretical philosophy, they do not apply to
transcendental arguments in ethics, at least insofar as transcendental arguments in ethics
are used to justify Kantian constructivist conclusions. This, in turn, should take the sting
out of the idea that transcendental arguments in ethics are prima facie unpromising.
Recent discussions about transcendental arguments in theoretical philosophy have
focused mainly on the status of the transcendental commitment, in light of the objections
raised by Barry Stroud (1968). In his criticism of transcendental arguments, Stroud focuses
specifically on transcendental arguments that try to provide an answer to (ontological)
scepticism about the existence of the external world and the possibility of secure knowledge
of it or the existence of other minds. There is, according to Stroud, a distinction to be made
between that which one must accept and that which exists, or between justified belief and
true belief. Consequently, his objection to these so-called strong transcendental arguments
13
The pioneering article on this topic is in Stroud (1968). Most of the recent work on transcendental arguments in
ethics could be understood as an attempt to analyse the potential of transcendental arguments in light of the
objections raised by Stroud. For recent discussions of transcendental arguments in theoretical philosophy see
Stern (1999; 2000; 2013a).
14
This applies to most discussions about transcendental arguments in theoretical philosophy. See e.g. Stroud
(1968; 1994), Cassam (1987; 1996) Stern (2000) and all papers on transcendental arguments in the edited volume
by Stern (1999). There are exceptions. Stern (2013a) explicitly notes the difference between transcendental
arguments in theoretical philosophy and in practical philosophy, and he acknowledges that transcendental
arguments in practical philosophy might not be susceptible to the Stroudian worries about transcendental
arguments in theoretical philosophy (for a similar claim see Skidmore 2002). See also Stern (2011) for a
reconstruction of Korsgaard’s (1996) transcendental argument for the value of humanity.
TRANSCENDENTAL ARGUMENTS
75
– those concerned with justified true belief – is that a transcendental argument can only
justify the conclusion that we are necessarily committed to the belief that the external world
exists, but it cannot lead to the conclusion that the external world actually exists.15 The
reason for this is that “the sceptic can always very plausibly insist that it is enough to make
language possible if we believe that S is true, or if it looks for all the world as if it is, but that
S need not actually be true” (Stroud 1968, 225).16 Stroud’s point is thus that, taken by itself,
a transcendental argument does not justify any ontological conclusions.
Stroud suggests that there might be ways in which a transcendental argument could
cross this so-called bridge of necessity which exists between belief and true belief. He
suggests, for instance, that a transcendental argument could rely on a form of
verificationism, according to which a proposition can only be intelligible if it is truth apt
(Stroud 1968, 247). If verificationism is true, there is, at least assuming that some of our
world-directed beliefs are meaningful, a direct relation between beliefs and truth, because
beliefs can only be meaningful if they can also be true. This crosses the bridge of necessity
because the gap between justified beliefs and justified true beliefs is closed. Stroud, however,
argues that although this principle succeeds in crossing the bridge of necessity, it would
make the transcendental argument itself superfluous because the verificationist principle by
itself constitutes a rejection of global scepticism about the external world.17 In other words,
if verificationism is true, there is no need to rely on a transcendental argument in order to
refute the ontological sceptic (for a discussion see Stern 2000, 44–49). Moreover,
verificationism is itself a controversial position.
Another way in which the bridge of necessity could be crossed is by relying on an
idealist metaphysics. If we accept idealism, the gap between our beliefs and the world would
also dissolve because truth would be constituted by what we believe (Stern 2000, 49–58).
Against the background of idealism, Stroud’s objection becomes powerless, because it
presupposes a distinction between our beliefs and the world. Robert Stern, however, argues
15
The reason why Stroud focuses on these so-called strong transcendental arguments has to do with Peter
Strawson’s (1975) attempt to put forward a strong transcendental argument in The Bounds of Sense. Strawson’s use
of transcendental arguments has been hugely influential in recent discussions on transcendental arguments.
16
Of course, someone using a transcendental argument might resist this conclusion and claim that it is really the
existence of Y which is a necessary condition of the possibility of X, but this claim seems to be hard if not
impossible to substantiate (cf. Cassam 1987). I will come back to this possible reply in the next section when I
discuss David Enoch’s transcendental argument for moral realism.
17
Stroud offers the following reason: “If the skeptic’s claim makes sense it must be false, since if that proposition
could not be known to be true or known to be false it would make no sense” (Stroud 1968, 247).
CHAPTER 3
76
that the problem with this response is similar to relying on a verification principle (2000,
49–58). That is, by relying on an idealist metaphysics, the transcendental argument itself
becomes redundant, because if one is an idealist, ontological scepticism about the external
world does not make sense in the first place. In addition, Stern stresses that, just like
verificationism, idealism is itself a controversial metaphysical position.
So although there might be ways to cross the bridge of necessity, this bridge cannot be
crossed without making the transcendental argument itself superfluous by relying on
controversial doctrines such as verificationism or idealism. Ultimately, Stroud’s objection
to transcendental arguments is thus that they either fail to answer ontological scepticism or
become redundant because they have to rely on additional premises such as the verification
principle or idealism, which by themselves constitute answers to ontological scepticism and
which are themselves controversial.
Stroud’s objection is widely accepted in theoretical philosophy, and for the sake of
argument I will assume that these kinds of Stroudian worries do indeed pose a problem for
transcendental arguments in theoretical philosophy. Note, however, that Stroud’s objection
applies exclusively to transcendental arguments that aim to justify ontological claims. If a
transcendental argument does not have an ontological claim as its conclusion, it is not
susceptible to Stroud’s objection. Nevertheless, Stroud’s objection is often interpreted as
undermining transcendental argumentation in general, and not just strong transcendental
arguments. The reason for this is the widespread assumption that more modest
transcendental arguments, i.e. arguments which do not aim to justify true beliefs but which
‘merely’ aim to show what we should believe or experience, or which capacities we should
have (Stern 2000, 10–11), are somehow less ‘interesting’ than strong transcendental
arguments (Stern 2000, 65).18
The assumption that modest transcendental arguments are less interesting can be
traced back to a fundamental assumption about the function of a transcendental argument.
The assumption is that transcendental arguments should be understood first and foremost
as anti-sceptical arguments (see e.g. Stroud 1968, 243; Cassam 1987, 355; Stern 2000, 3–6).
From this perspective, any transcendental argument that does not try to defeat the
ontological sceptic seems to be less interesting by definition, because ontological scepticism
is, at least in these debates in which realism is taken to be the default metaphysics, typically
18
Stern does believe that there are ‘interesting’ modest transcendental arguments (Stern 2000, 125). He does,
however, seem to share the assumption that the most interesting transcendental argument would be a strong
transcendental argument.
TRANSCENDENTAL ARGUMENTS
77
assumed to be the most radical form of scepticism and, therefore, the only genuinely
interesting (philosophical) opponent.
The assumption that the supposedly most interesting transcendental argument in
theoretical philosophy is a strong transcendental argument has, however, been challenged,
and I think for good reasons (see e.g. Ameriks 1978; Cassam 2003). It should not come as a
big surprise that an argument which explores the necessary conditions of the possibility of
our self-understanding, i.e. our first-person perspective as agents, thinkers or experiencers,
cannot lead to any strong ontological claims about the world. For how could an argument
from the first person lead to impersonal or third-personal conclusions about the way the
world is? But this does not necessarily mean that the transcendental argument is somehow
less interesting or superfluous.
Karl Ameriks, for instance, responds to Stroud by claiming that “not every interesting
argument has to be a refutation of extreme scepticism” (1978, 282).19 Positively, one might
say that transcendental arguments do not just have an anti-sceptical function, but also have
what one might call a ‘constructive’ function in contributing to our practical selfunderstanding, for instance by reflecting on the necessary preconditions of our conceptual
scheme, the structure of language or the possibilities of experience. Even if this all proves to
be an illusion induced by an evil demon, learning something about the illusion in which we
inescapably live might be very interesting. By not exclusively focusing on its anti-sceptical
potential but on its constructive potential, it might be easier to see the potentially
interesting conclusions that could be generated by modest forms of transcendental
argument.
Of course, much more can be said about Stroud’s objection to transcendental
arguments in the context of theoretical philosophy. What I have tried to show, however, is
that Stroud’s objection applies exclusively to strong transcendental arguments and that the
question of whether or not this undermines transcendental arguments in general hinges on
19
Ameriks distinguishes between what he calls ‘regressive’ and ‘progressive’ transcendental arguments. Ameriks
introduces this distinction in the context of the interpretation of Kant’s transcendental deduction in The Critique
of Pure Reason. The distinction between regressive and progressive transcendental arguments refers to different
readings of the appropriate starting point of the transcendental deduction. On the progressive reading, the
argument starts from the idea that we are self-conscious and consequently argues that “one can be self-conscious
only if there is an objective world of which one is aware” (Ameriks 1978, 227). On a regressive reading, on the
other hand, the starting point is that we have objective, empirical knowledge and consequently that “for us there is
objectivity, and hence empirical knowledge, only if the categories are universally valid” (Ameriks 1978, 277). In
other words, a regressive transcendental argument accepts “empirical knowledge as a premise to be regressively
explained rather than as a conclusion to be established” (Ameriks 2003, 54).
CHAPTER 3
78
an understanding of the ambition and function of specific transcendental arguments (and,
perhaps, the meta-philosophical question of what arguments are and what would make
them interesting, more generally). In what follows, I will turn to transcendental arguments
in ethics. First, I will show that Stroud’s objection does apply to attempts in ethics to justify
realist conclusions on the basis of a transcendental argument. In the remainder of this
chapter, I will argue that Stroud’s objections do not apply to Kantian constructivism
because Kantian constructivists do not have to make a claim about the existence of moral
facts.
4. Transcendental Arguments in Ethics
4.1 Transcendental Arguments and Moral Realism
In the introduction to this thesis, I have stressed that I will not directly engage with moral
realism. Instead, the aim of the thesis is to show that it is possible to justify objective moral
judgements without relying on realist assumptions. Although this does not by itself refute
moral realism, showing that moral objectivity can be vindicated without relying on moral
realist commitments might be attractive even to those theorists who are initially attracted to
moral realism.
Before turning to Kantian constructivism, I discuss, in this section, attempts to justify
moral realist conclusions on the basis of a transcendental argument (for two prominent
example see Illies 2003; Enoch 2011b; see also Regan 2002, 273–74 who hints at a
transcendental argument for Mooran realism which closely resembles Enoch’s argument).20
The goal of this discussion is to show that transcendental arguments for moral realist
conclusions are susceptible to Stroudian worries about transcendental arguments and,
subsequently, that transcendental arguments for Kantian constructivist conclusions are not
20
One could wonder, however, whether Illies is really a transcendental moral realist, depending on how exactly
one defines moral realism. On the one hand, Illies defines moral realism as the ontological view that “there are
moral facts, which exist independently of our evidence for them” (2003, 4). On the other hand, he also claims that
one can be a moral realist without accepting a robust realist metaphysics. For instance, Illies argues that Habermas
should be labelled as a moral realist because, even though Habermas denies the existence of mind-independent
moral facts, he holds that there exist rationally justified norms and he claims that these norms are true to the
extent to which they form the necessary preconditions of rational argumentation (Illies 2003, 5f4). I think it is
more helpful to distinguish cognitivism (the epistemological claim that moral judgements can be correct and
incorrect) from moral realism (the ontological claim that the standard of correctness is moral facts). Illies thus
seems to conflate cognitivism and realism and in this way simply blocks the possibility of Kantian constructivism.
TRANSCENDENTAL ARGUMENTS
79
susceptible to Stroud’s objection, precisely because of the differences between moral realism
and Kantian constructivism.
Stroud’s objection to strong transcendental arguments in theoretical philosophy
already suggests that the prospects of transcendental moral realism succeeding are slim.
After all, any transcendental argument for an ontological conclusion, including arguments
for moral realism, seems to be susceptible to Stroud’s objection. Some transcendental moral
realists, however, think that Stroud’s objection can be answered (Illies 2003, 61; Enoch
2011b, 78). In this section, I therefore critically discuss one such attempt at justifying moral
realism through a transcendental argument. I do so by focusing on David Enoch’s
transcendental argument for his robust normative (moral) realism, i.e. the view that there
exist mind-independent normative facts, including moral facts.21 I end the section by briefly
indicating why I think that my objections to Enoch apply to any transcendental argument
for moral realism.
The starting point of Enoch’s transcendental argument, which he labels ‘the argument
from the deliberative indispensability of normative truths’, is the idea that we typically
deliberate about what to do and that this is non-optional from our first-person perspective
(Enoch 2011b, 76).22 Subsequently, Enoch tries to show that in deliberation we necessarily
have to rely on normative truths (and not just on certain beliefs or on certain principles of
action). Insofar as deliberation is truly inescapable, and insofar as the existence of
normative facts is a necessary condition of the possibility of deliberation, robust normative
realism must be true.23 Enoch’s argument can be summarized as follows:
1) Deliberation (X) is inescapable for A.
2) The existence of normative truths (Y) is indispensable for deliberation (X) (i.e. the
21
Enoch calls his argument ‘the argument from deliberative indispensability’, and he considers this argument to be
a kind of transcendental arguments: “my indispensability argument for Robust Realism may be thought of as a
kind of transcendental argument” (Enoch 2011b, 79).
22
Enoch claims that the relevant kind of non-optionality is ‘rational non-optionality’ (2011b, 78). I will come back
to different notions of non-optionality or inescapability in the next chapter when I address Enoch’s objection to
Kantian constructivism (Enoch 2006; Enoch 2011a). For the sake of argument in this section, I assume that
deliberation is non-optional and that we have settled on which type of non-optionality is appropriate in order to
be able to focus on the status of the transcendental commitment.
23
Note that even if this argument succeeds, it is an argument for the existence of robust normative truths, but not
yet for robust moral truths (Enoch 2011b, 50). Enoch thinks that the argument can be supplemented in order to
reach the conclusions that moral truths exist as well. I do not discuss Enoch’s argument for distinctively moral
truths. The reason for this is that I think the argument for robust normative truths already fails and there is
therefore no reason to consider the extension to moral truths.
CHAPTER 3
80
existence of normative truths is a necessary condition of the possibility of
deliberation).
3) Therefore, A must accept that normative truths exist (Y).
In order to support the premise that deliberation presupposes the existence of normative
truths, Enoch puts forward the example of a student who tries to decide whether to join a
law firm, apply to graduate school to study philosophy or to do something else altogether
(2011b, 72–73). Enoch’s argument proceeds as follows: first, he claims that the student, or
anyone who deliberates, has to presuppose that these kinds of deliberative questions have
answers. Second, and more importantly, Enoch claims that you, the person deliberating,
also have to presuppose that the correctness of an answer is not up to you: “this is
something you are trying to discover, not create. Or so, at the very least, it feels like when
deliberating” (Enoch 2011b, 73). In other words, Enoch claims that when deliberating you
have to assume that there exist facts of the matter about the choices you are making – or
maybe more accurately, Enoch claims that it feels as though facts of the matter about the
choices you are deliberating about exist. Finally, Enoch argues, the idea that you are trying
to discover answers to deliberative questions commits you to the existence of normative
truths and hence to his robust moral realism: “deliberation … is an attempt to eliminate
arbitrariness by discovering (normative) reasons, and it is impossible in a believed absence
of such reasons to be discovered” (Enoch 2011b, 74).
Enoch stresses that this argument from deliberative indispensability is analogous to socalled explanatory indispensability arguments in science. The example he uses is the
explanatory indispensability of the existence of electrons: 24
[W]hy should we believe in, say, electrons? One common – and plausible – answer
runs like this: There are many inferences to the best explanation the conclusion of
which entails the existence of electrons: our best scientific theories quantify over
electrons; we ought to believe that these theories are at least approximately true
(they are, after all, our best theories, our best explanations of numerous
phenomena; and they are also – now in non-comparative terms – fairly good), so
we ought to believe that electrons exist (Enoch 2011b, 54).
24
Note that in this quotation Enoch does not claim that electrons exist (strong transcendental argument), but only
that we have to believe that electrons exist (modest transcendental argument). He does, however, claim that belief
implies existence (Enoch 2011b, 78). I come back to this below.
TRANSCENDENTAL ARGUMENTS
81
This argument from explanatory indispensability thus has the following transcendental
structure:
1) Explanation (X) is inescapable for A.
2) The existence of electrons (Y) is indispensable for explanation (X) (the existence of
electrons is a necessary condition of the possibility of explanation).
3) Therefore, A must accept that electrons exist (Y).
Enoch claims that we normally accept arguments from explanatory indispensability, such
as the argument for the indispensability of electrons for scientific explanation, and that in
the absence of a principled distinction between deliberative indispensability and
explanatory indispensability we have to accept that robust normative truths exist. Enoch
formulates this argument in the form of a rhetorical question to his readers: “can you think
of any reason for grounding ontological commitments in explanatory indispensability that
is not really more general, a reason for grounding ontological commitment in
indispensabilities of other kinds as well?” (Enoch 2011b, 55). Enoch clearly thinks that such
a reason cannot be found.25
I think that Enoch’s argument faces many problems. For one thing, I am not convinced
by Enoch’s phenomenology of deliberation. Enoch’s main argument for the indispensability
of normative truths in deliberation is that it ‘feels like’ there are normative truths when I
deliberate. However, although I agree with Enoch that some answers to deliberative
questions might be better than others and that this is a necessary part of deliberation (we
are talking about reasons, after all, and those can be better or worse), I do not see why this
can only be explained in terms of the existence of mind-independent normative truths.26
Maybe it is better for the student to study philosophy than to join a law firm because this
choice fits better with the person she wants to be. Why could this evaluation only be
explained in terms of mind-independent facts?27 There is thus a gap between the idea that
some answers to deliberative questions are better than others and the claim that irreducibly
25
For a criticism of this point see Cline (2016).
What seems to be happening here is analogous to a certain tendency of mainstream meta-ethics to equate
cognitivism, i.e. the view that moral judgements are truth apt, with (robust) realism, i.e. the idea that mindindependent facts are the appropriate truth-makers.
27
An alternative account of what is going on in this case is, for instance, the interpretivist account of practical
reasoning proposed by Charles Taylor (1985). See also Lenman’s (2013) argument against the idea that normative
truths are indispensable for deliberation.
26
CHAPTER 3
82
normative truths are the most appropriate truth-makers.
Enoch does not really provide any argument for his assumption that only mindindependent normative truths can account for the idea that some choices are better than
others. He simply stipulates that this is the case by drawing an analogy between deliberation
and finding answers to factual questions, stating that
it is worth noting how similar the phenomenology of deliberation is to that of
trying to find an answer to a straightforwardly factual question: when trying to
answer a straightforwardly factual question (like what the difference is between the
average income of a lawyer and that of a philosopher) you try to get things right,
to come up with the answer that is – independently of your settling on it – the
right one. When deliberating, you also try to get things right, to decide as –
independently of how you end up deciding – it makes most sense for you to decide
(Enoch 2011b, 73).
Now, maybe for Enoch it feels as if there is a normative fact of the matter when he
deliberates, but he has not provided the reader with any reason why this is an indispensable
element of deliberation in general and not just an element in the deliberation of the robust
normative realist David Enoch. To me, the phenomenology of settling factual questions
seems to be very different from the phenomenology of deliberation about what to do.28
Making up my mind about whether there is a coffee machine on my worktop feels very
different from making up my mind about important career choices. Without a convincing
argument for the analogy between deliberation and answering factual questions, robust
normative realism is thus simply presupposed in the argument rather than argued for.
But here I want to focus on a more fundamental problem with Enoch’s argument
which applies to any transcendental argument for a moral ontological conclusion. For the
sake of argument, let us assume that Enoch is right that it feels as though mindindependent normative truths are indispensable for deliberation from the perspective of
someone who deliberates. Now, recall that Stroud’s objection to strong transcendental
28
My objection here is an instantiation of what Cassam (1996) calls the problem of misconception. The problem of
misconception arises when someone does not actually accept a transcendental commitment but still seems to be
able to engage in the relevant practice (such as deliberation). As I will argue below, and in the next chapter,
Kantian constructivism is not susceptible to the problem of misconception because Kantian constructivism does
not claim that believing Y is a necessary precondition of agency, but that an agent is necessarily committed to
doing Y (which can be valid independent of an agent actually doing Y).
TRANSCENDENTAL ARGUMENTS
83
arguments is that such an argument can at best justify a belief in Y, but can never establish
the existence of Y. That is, one cannot reach conclusions about mind-independent facts
from a radically mind-dependent, first-personal argument. Note that Enoch explicitly
acknowledges the first-personal character of transcendental argument: “the whole point of
the argument of this chapter is the focus on the first person, deliberative perspective”
(Enoch 2011b, 76). The general objection to strong transcendental arguments therefore also
seems to apply to Enoch’s transcendental argument for robust normative realism.
Enoch explicitly addresses Stroud’s objection and puts forward two different lines of
response. In one response, Enoch again stresses the analogy between the usage of inferences
to the best explanation in science and deliberative indispensability in ethics (Enoch 2011b,
78). According to Enoch, electrons, and not just beliefs in electrons, are ‘explanatorily’
indispensable and this, Enoch claims, is commonly taken as an indication that electrons
exist. If we are justified in making a claim about the existence of electrons on the basis of
their explanatory indispensability, Enoch continues, we should also be justified in making
claims about the existence of normative truths on the basis of their deliberative
indispensability.
However, Enoch’s attempt to shift the burden of proof by way of analogy with scientific
objects obviously fails to engage with the core of Stroud’s argument. Stroud’s objection is
first and foremost directed against transcendental arguments that try to make ontological
claims about the external world. The argument from explanatory indispensability for the
existence of electrons is an example of such an argument. Stressing the analogy between the
explanatory indispensability of electrons and the deliberative indispensability of normative
facts therefore does not help to overcome Stroud’s objection.
Enoch’s second response has to do with the relation between belief and existence, i.e.
the relation between the belief in normative truths and the existence of normative truths. At
one point in his book, Enoch seems to grant that what he calls ‘directly indispensable’ in
both the explanation and the deliberation is indeed the belief in electrons or the belief in
robust normative facts and not the electrons or the normative facts themselves: “what is
directly indispensable in the explanatory case too is the belief in electrons” (Enoch 2011b,
78). However, Enoch continues by noting that the content of the indispensable belief is the
existence of electrons or normative truths. Subsequently, his argument, in the case of
normative truths, proceeds as follows: “the content invoked in deliberation is in terms of
the relevant normative truths, not (except in unusual cases) in terms of beliefs about them.
And for this reason what is needed for the deliberation to succeed is for there to be
CHAPTER 3
84
normative truths” (Enoch 2011b, 78). Thus, Enoch concludes that “in both cases, the
indispensability of the relevant objective things (electrons, or normative truths) follows
from the indispensability of the belief in those things for the relevant projects” (Enoch
2011b, 78). Enoch’s argument thus seems to be that because the content of the
indispensable belief is a normative truth, these normative truths must exist.
But this response simply begs the question. For why would it be the case that if you
have to believe in the existence of Y, Y also has to exist? Again, Enoch fails to argue for a
fundamental assumption. He claims: “you cannot believe that there are electrons, or
normative truths, without believing that there are electrons, or normative truths” (Enoch
2011b, 78). But this will not do. Even if one has to believe that there are normative truths, it
does not mean that normative truths actually exist.
Ultimately, Enoch seems to rely on something like a verification principle to justify the
existence of normative facts. That is, Enoch simply denies the possibility that deliberation
might rely on a false belief in normative truths, i.e. that one might believe in normative
truths but that these truths do not actually exist (Enoch 2011b, 79). Enoch writes that “in at
least reasonably fortunate circumstances, epistemic justification and truth are reliably
correlated” (Enoch 2011b, 65).29 Of course, if one accepts this premise, Enoch’s argument
for robust normative realism is successful because justification would be correlated with
truth. However, if one accepts this premise, the indispensability argument itself becomes
superfluous. After all, if one thinks that epistemic justification and truth are reliably
correlated, this of itself already entails normative realism, because any justified normative
belief would by definition correlate to a normative fact (assuming that there are at least
some epistemically justified normative beliefs).
Now, one might wonder whether Enoch could retreat to a modest transcendental
argument and admit that the only thing he has established is that we necessarily have to
believe that robust normative facts exist – a possibility which is not considered by Enoch
himself, and rightly so, as I will argue.30 The problem with this suggestion is that this
argument would simply fail as an argument for moral realism as a metaphysical position.
On the modest transcendental argument, one has a mind-dependent reason to believe in
29
Enoch puts forward the speculative evolutionary reason that “creatures whose basic belief-forming methods
were radically unreliable are probably no longer with us” (Enoch 2011b, 65).
30
Note that retreating to a modest transcendental argument is not the same as providing an argument for Kantian
constructivism. As I will argue below, the distinction between strong and modest transcendental arguments does
not apply to a transcendental argument for Kantian constructivism.
TRANSCENDENTAL ARGUMENTS
85
mind-independent normative facts, but the argument does not prove that these facts
actually exist. A modest transcendental argument, in other words, operates exclusively in
the realm of justification or epistemology and therefore does not and cannot arrive at realist
conclusions. In addition, one might wonder what it would mean to only believe in mindindependent normative truths. Believing in mind-independent normative truths does not
seem to mean anything over and above the conclusion that there are better and worse
answers to deliberative questions. But this, as I already noted above, is still a far cry from
normative realism, which after all is first and foremost an ontological thesis.
A second problem with a modest transcendental argument for mind-independent
moral realism is that it opens up the possibility of what Robert Stern calls a sceptical
nightmare (2015, 168). A sceptical nightmare is the situation in which one necessarily has
to believe in Y, but it might be possible that Y does not really exist, which implies that one
might be forced to believe in an illusion. This sceptical nightmare arises because the modest
transcendental argument acknowledges the gap between justified belief and existence and
therefore allows for the possibility that one is necessarily committed to a belief in
something whose existence cannot be established. After all, the modest transcendental
moral realist still believes in the existence of mind-independent normative truths, but has no
resources to know whether they really exist.
In short, the problem with transcendental arguments for normative (moral) realism is
that on the strong interpretation they fail to close the gap between belief in normative
truths and their existence, and on the modest interpretation the argument no longer seems
to constitute an argument for realism as a metaphysical theory and allows for the possibility
of a sceptical nightmare given the necessity of believing in mind-independent normative
truths.
Enoch’s transcendental argument for the existence of mind-independent normative
truths is only one example of a transcendental argument for normative or moral realism.
There are, however, good reasons to suspect that similar problems apply to any attempt to
argue for realist conclusions on the basis of a transcendental argument. The reason for this
should be clear by now: a transcendental argument is a distinctively first-personal one. It
should not be expected that such an argument could lead to proving the existence of mind-
CHAPTER 3
86
independent facts, whether they are normative facts or other kinds of facts.31 In the next
section, I will therefore turn to transcendental arguments against the background of a
constructivist metaphysics and argue that these kinds of arguments are not susceptible to
Stroud’s objection to transcendental arguments. In addition, I will argue that they do not
allow for the possibility of a sceptical nightmare. And finally, I will claim that these
transcendental arguments are not in any way ‘less interesting’ than (strong) transcendental
arguments for moral realism.
4.2 Transcendental Arguments and Kantian Constructivism
The goal of Kantian constructivism is to justify an objective moral principle that prescribes
how one should act. The goal of justifying a moral principle, as opposed to moral facts, has
important implications for the status of the transcendental commitment. Transcendental
arguments in theoretical philosophy typically aim to justify certain beliefs about the
external world. For instance, transcendental arguments in theoretical philosophy argue that
the external world exists, or at the very least that we are justified in believing that the
external world exists. As I have shown in the previous section, transcendental moral realism
adopts this theoretical model of transcendental argumentation to argue for the existence of
mind-independent moral facts.
The claim of Kantian constructivism, however, is not first and foremost that every
agent must have the belief that Y, but rather that every agent or anyone who engages in
certain types of interaction must accept Y as a principle of action. In Kantian
constructivism, the transcendental commitment is thus not a belief about the external
world, but a practical commitment, i.e. a commitment about how one ought to act.
Consider, for instance, Korsgaard’s transcendental argument for the value of humanity.
Korsgaard does not claim that one must believe in the existence of the value of humanity,
31
Arguably, Illies (2003) faces a similar sort of problem. Illies tries to use a transcendental argument to justify
moral realism, and his response to Stroudian scepticism is that it is unintelligible given the methodological
commitments of transcendental argumentation. Illies’ main point is that the Stroudian sceptic would have to raise
his sceptical doubts from a standpoint from which he could make a distinction between necessary presuppositions
and truth, and he claims that such a standpoint is simply not available to the sceptic. He concludes that “if it is
impossible not to think in a certain way, then we cannot ‘think’ that the world could be different from how we
must think” (Illies 2003, 62). However, it seems to me that Illies’ own argument backfires on his own attempt to
justify moral realism through a transcendental argument. For the absence of this standpoint means that Illies
himself is likewise not warranted to conclude that the necessary preconditions do indeed establish the existence of
moral facts.
TRANSCENDENTAL ARGUMENTS
87
but that one should actually value one’s (own) humanity, i.e. that one should act in such a
way as to respect the value of (your) humanity. In order to see the difference between
beliefs about the external world and practical commitments, consider, for instance, the
claim ‘you should pay your taxes’. This claim does not (just) imply that you should have the
belief that you should pay your taxes, but that you should actually pay your taxes (imagine
someone saying to the tax collector, “It is true that I did not pay my taxes, but I had the
belief that I should pay my taxes!”).32
This suggests that Kantian constructivism is not just a modest version of
transcendental moral realism. The distinction between modest and strong transcendental
arguments presupposes that transcendental arguments justify beliefs about the external
world; a strong transcendental argument produces justified true beliefs and a modest
transcendental argument produces justified beliefs. Both strong and modest transcendental
arguments essentially produce beliefs about the external world. The Kantian constructivist
claim is, however, not that we should believe in mind-independent moral facts – versus
thinking that a transcendental argument can prove the existence of mind-independent
moral facts (transcendental moral realism). The explanandum of Kantian constructivism is
not a question of metaphysics or justified belief, but a practical, normative ethical question
about how we ought to act. Therefore, the distinction between modest and strong
transcendental arguments does not seem to apply to Kantian constructivism.
Transcendental arguments that aim to justify practical commitments are not vulnerable
to the Stroudian objection to transcendental arguments, for the obvious reason that they do
not aim to justify true beliefs or the existence of certain moral facts. In addition, there
seems to be no risk of a sceptical nightmare. The fear of a sceptical nightmare only arises if
one has to believe that there are mind-independent moral facts. A constructivist denies that
we have reason to believe in mind-independent facts in the first place, and therefore there is
no risk that we are necessarily committed to believing in an illusion. In other words, in
Kantian constructivism there is no standpoint from which the sceptical nightmare can
arise.
But could it not be the case that Kantian constructivism is susceptible to a variant of
Stroud’s objection and to a variant of the sceptical nightmare? Could it not be the case that
we, for instance, ought not to value humanity even if we necessarily have to value humanity
32
One way to account for these differences is by saying that a transcendental argument in ethics and
transcendental arguments in theoretical philosophy have a different direction of fit. I will come back to this in the
next chapter.
CHAPTER 3
88
insofar as we understand ourselves as an agent? This objection has recently been put
forward by Matthew Silverstein (2012). Anticipating my Kantian constructivist response to
a Stroudian objection to transcendental arguments, Silverstein writes that33
it is not clear how shifting from an inescapable doxastic attitude to an inescapable
practical one makes matters any better vis-à-vis worries about the validity of our
ethical practices. That we are all necessarily committed to respecting one another’s
humanity seems to be perfectly consistent with its not being the case that we ought
to respect one another’s humanity. We might, in other words, be victims of a
collective “practical” delusion, helplessly devoted to pursuing something which
actually ought not to be pursued (Silverstein 2012, 5).
I am not sure what to make of this kind of scepticism. Silverstein suggests that the following
two claims are “perfectly consistent”:34
1) We are all necessarily committed to respecting one another’s humanity.
2) We ought not to respect one another’s humanity.
I fail to see, however, why these to claims would be perfectly consistent. Of course, they
could be made consistent by claiming, for instance, that one’s necessary practical
33
Although Silverstein does not mention Stroud’s paper on transcendental arguments, he discusses a similar
objection: “consider a version of the Kantian strategy according to which what is inescapable for agents is some
normative belief – say, the belief that we ought to respect one another’s humanity. Would knowing that this
normative belief is inescapable provide us with any reason to conclude that it is correct? Would such knowledge
help to allay concerns about the objective validity of our practice of judging that we ought to respect one another’s
humanity? I cannot see how it would, any more than I can see how learning about the inescapability of some belief
regarding the influence of celestial bodies would assure us of the objective validity of some form of astrology. For
the standard of correctness for belief is truth, and showing that a belief is inescapable does not in any way establish
that it is true. That we all believe that p is perfectly and obviously consistent with p’s being false, for there is always
the possibility that we are victims of some sort of collective delusion. Adding that the belief in question is
inescapable – that we all have this belief merely because we are agents – if anything only strengthens the suspicion
that we have all been duped” (Silverstein 2012, 5).
34
The goal of Silverstein’s paper is not so much to defend this kind of scepticism, but rather to show that Kantians
can only reply to this scepticism by engaging in traditional meta-ethics. His goal is thus to criticize Kantian
constructivists who deny that they are engaged in meta-ethics rather than to criticize Kantian constructivism (see
e.g. Korsgaard 2008b, 325n9). Silverstein therefore does not deny that a Kantian constructivist can give a
convincing reply to this kind of scepticism, but he argues that giving a reply commits a Kantian constructivist to a
specific meta-ethics. I am not sure whether I agree with Silverstein. Much depends on how exactly one defines
meta-ethics and normative ethics, and I will not pursue this question here. Also, I disagree with Silverstein about
the initial plausibility of this scepticism: I do not see why the two claims mentioned in the text would seem
“perfectly consistent”.
TRANSCENDENTAL ARGUMENTS
89
commitments could be in conflict with (moral) oughts that are somehow derived from
mind-independent moral facts. However, the fact that these two claims could be made
consistent in this way does not show that they are perfectly consistent. It only shows that
they are perfectly consistent on one specific view about the source of (moral) oughts. Only
if one thinks that realism about morality is the most desirable view of ethics could this
variant of a sceptical nightmare arise.
I would deny, however, that there is a presumption in favour of realism in the ethical
domain.35 The question of ethics is not, first and foremost, a question about (moral) truth,
but a question about reasons for action. It is therefore simply not obvious that realism
would be the most desirable or most robust moral metaphysics. Even if it can be shown that
all agents are necessarily committed to certain moral claims, it is not clear that the
additional, realist metaphysics would add anything substantive to the objectivity of these
claims.36 The strongest, or most relevant, type of scepticism in the context of ethics, or
practical philosophy in general, is thus not one that questions the existence of certain
metaphysical facts, but one that questions the rational bindingness of moral demands. The
moral sceptic does not ask whether moral principles are literally true, but whether he or she
has reason to adopt those principles as principles of action.
This is all compatible with the fact that a particular reconstruction of our practical
commitments, e.g. the reconstruction of these commitments in terms of valuing one’s
humanity, might turn out to be wrong, and that it therefore fails to state what we ought to
do – if valuing one’s humanity is not a necessary condition of the possibility of agency, one
ought not to value one’s humanity. Although it is true that a valid transcendental argument
would be indubitable and therefore entails claims about what we ought to do, the specific
transcendental arguments that are put forward by different theorists are obviously fallible.37
Transcendental arguments should thus be understood as fallible attempts to describe our
necessary practical commitments. I therefore agree with Charles Taylor, who claims that “a
valid transcendental argument is indubitable; but it is hard to know when you have one, at
35
Note that this is not the same as denying moral realism; it is denying the presumption in favour of realism,
which is quite common in meta-ethics.
36
I think (most) realists would agree with this, but that they would deny that there is a way to justify categorical
and universal moral judgements without reference to moral facts.
37
Here I agree with Habermas (1990a, 95–98), who claims that transcendental arguments are fallible. Habermas,
however, distinguishes fallible, or what he calls weak transcendental, arguments from what he calls strong
transcendental arguments, which, according to Habermas, lack this idea of fallibility. As I explain in the text, I
think that all transcendental arguments are fallible and that Habermas’s distinction between strong and weak
transcendental argument is therefore a false dichotomy.
CHAPTER 3
90
least one with an interesting conclusion. But then that seems true of most arguments in
philosophy” (Taylor 1978, 165).
Let me end this section by stressing that understanding transcendental commitments
in terms of practical commitments has important implications for the status of the moral
principles that are justified through transcendental arguments, more specifically their status
as objective moral principles. Recall that there are different ways of conceptualizing
(maximal) moral objectivity. On an ontological or realist conception of objectivity,
objectivity is constituted by the correspondence between moral principles and moral facts.
Insofar as transcendental arguments in ethics cannot be used to justify moral realist
conclusions, transcendental arguments cannot lead to realist moral objectivity. On a
constructivist conception of moral objectivity, on the other hand, objectivity is constituted
by a practical point of view that each individual should adopt independent of his or her
contingent beliefs or desires. If a transcendental argument can show that every agent is
necessarily committed to certain principles of action, including principles about how to
treat others, the argument does guarantee this type of constructivist objectivity.
Transcendental arguments should thus be understood as an attempt to justify a practical or
constructivist account of objectivity, meaning that the moral principles that are justified
through a transcendental argument do not represent a mind-independent reality, but
express universal and categorical practical commitments, for which the existence of moral
facts is not necessary.
In this section, I have provided a rough description of Kantian constructivism and its
reliance on transcendental arguments. Much of the detail of this view will have to be added
in the chapters to come. What I have tried to show in this section is that there are
important differences between the empirical or theoretical domain and the practical or
ethical domain, and that Kantian constructivism can therefore not be dismissed on the
basis of scepticism about the plausibility and potential of strong and modest transcendental
arguments in general. Instead, Kantian constructivism should be assessed on its own terms.
This will be the topic of the remainder of this thesis.
5. Conclusion
In this chapter, I have introduced transcendental argumentation as a method of moral
justification and I have argued that transcendental Kantian constructivism is not
susceptible to common worries about transcendental arguments in theoretical philosophy.
TRANSCENDENTAL ARGUMENTS
91
In addition, I have argued that transcendental arguments for moral realism are doomed to
fail, in light of Stroud’s objection to strong transcendental arguments. The upshot of this
discussion is that if there is a plausible transcendental argument in ethics, it vindicates a
distinctively Kantian constructivist conception of moral objectivity and this Kantian
constructivism is not susceptible to common objections to transcendental arguments in
theoretical philosophy.
In the following chapters, I will analyse different transcendental arguments in light of
the general dilemma for Kantian constructivism. In the next chapter, I will first turn to the
objection that Kantian constructivism cannot lead to any categorical normative claims.
Part 2: Categoricity
Chapter 4
General Meta-Ethical Objections to
Kantian Constructivism
1. Introduction
The potential general dilemma for Kantian constructivism states that either the starting
point of a transcendental argument is inescapable, but in that case no interesting moral
conclusions can follow from the starting point, or that Kantian constructivism succeeds in
justifying interesting conclusions – but in that case the starting point turns out to be merely
optional. In the remainder of this thesis, I will critically evaluate existing transcendental
arguments against the background of this potential dilemma. As I have already explained in
the introduction to this thesis, I will do so in three steps, by discussing the following:
1) To what extent a transcendental argument succeeds in justifying any categorical
normative claim.
2) To what extent it succeeds in justifying a categorical and universal principle of
interpersonal morality.
3) To what extent it succeeds in justifying substantive norms of action.
In this chapter, I discuss the first step, i.e. the justification of normative claims that any
agent has to accept independent of his or her contingent preferences or desires. I do so by
engaging with recent meta-ethical discussions in which it is argued that transcendental
arguments cannot justify any categorical normative judgements, let alone judgements of
substantive and interpersonal morality, because they are supposed to be susceptible to (a
CHAPTER 4
96
variant of) one or both horns of the potential general dilemma for Kantian constructivism.
I distinguish and respond to four closely related but distinct meta-ethical objections.1
These objections have two things in common. First, the objections are supposed to apply to
transcendental arguments in ethics in general, or at least all transcendental arguments from
the first person, and the idea is thus that the objections undermine transcendental
arguments in ethics independent of the details of a specific transcendental argument (e.g.
independent of the specific starting point of a transcendental argument and the specific
conclusions that are defended through a transcendental argument). Second, the objections
state that transcendental arguments cannot justify any categorical normative claim
independent of the more specific concern of transcendental arguments in ethics to justify
universal principles of interpersonal or substantive morality (i.e. the second and third steps
mentioned above).2
In order to distinguish the different objections, recall the structure of a transcendental
argument in ethics:
1) X is inescapable for A (e.g. agency is inescapable for A).
2) Y is a necessary condition of the possibility of X (e.g. valuing your own humanity
is a necessary condition of the possibility of agency).
3) Therefore, A must Y (therefore, A must value his or her own humanity).
Transcendental arguments in ethics (or at least arguments from the first person) typically
start from a conception of agency (X), and they try to show that the necessary
preconditions of agency include the acceptance of certain categorical normative claims (Y).
The four objections can be distinguished by specifying which part of the argument they
seek to undermine. The first objection, what I will call the ‘escapability objection’, criticizes
the first premise and states that agency, independent of how it is precisely understood, is
escapable and that therefore the conclusions of a transcendental argument are only
contingently valid (see e.g. Silverstein 2015).
The second objection is David Enoch’s (2006; 2011a) ‘shmagency objection’, which
1
Most of the discussions in meta-ethics are framed in terms of normative reasons. I will use ‘normative claims’
and ‘normative reasons’ interchangeably.
2
See e.g. Street (2012, p. 48f15), who stresses that her argument is not concerned with criticizing the attempts of
Kantian constructivists to justify the (universal) principle of interpersonal morality. Instead, she focuses on the
justification of categorical reasons, independent of the question of whether these entail a principle of interpersonal
morality. See also Enoch (2006, 170).
GENERAL META-ETHICAL OBJECTIONS
97
states that it only follows that one must accept whatever are the necessary preconditions of
agency if there is a reason to be an agent in the first place. On my reconstruction of Enoch’s
objection, his claim is that there is a hidden premise in the argument (1*: ‘A has a reason to
X’) and that if this premise is made explicit it becomes clear that a transcendental argument
presupposes another source of normativity, i.e. a realist reason to be an agent, to get off the
ground. According to Enoch, this shows that categorical normativity cannot be grounded
in the self-understanding of an agent. The difference between the first and the second
objection is the following: the shmagency objection grants that agency is inescapable in
some sense, and subsequently claims that the inescapability of agency by itself is not
sufficient to generate categorical reasons. The escapability objection grants that if agency is
inescapable it will lead to categorical reasons, but denies that agency is inescapable.3
The two other objections are directed, respectively, at premise 2 and 3 of a
transcendental argument. What I will call the ‘no-normativity objection’ is directed at
premise 2 of the transcendental argument: ‘Y is a necessary condition of the possibility of
X’. Street (2012) denies that the necessary conditions of the possibility of agency express a
normative relation – not because we need a reason to be an agent in the first place (pace
Enoch), but because she thinks that the necessary conditions of possibility express merely a
conceptual relation (or rely on a realist, normative truth). The final objection, what I will
call ‘the contingency objection’, is directed at premise 3 of the argument and states that not
every agent actually accepts Y, e.g. not every agent actually values her humanity, and that
this fact undermines the validity of the argument.
The main goal of the chapter is to provide a qualified response to these objections, and
in this way I aim to clarify the Kantian constructivist project in ethics. This response is
qualified because although I aim to show that the objections fail as general objections to
transcendental Kantian constructivism, providing a comprehensive response to these
objections requires an assessment of the details of specific transcendental arguments in
ethics. This will be done in the chapters that follow. This chapter thus functions as a
transition chapter between the discussion of transcendental arguments in general (in
chapter 3) and the evaluation of specific transcendental arguments (in the following
chapters).
Before proceeding, one point of clarification is in order. Most of the objections to
3
As I will argue below, the shmagency objection and the escapability objection are often conflated but should be
distinguished from one another.
CHAPTER 4
98
transcendental arguments in meta-ethics are raised in the context of Christine Korsgaard’s
(1996) argument for the value of humanity in The Sources of Normativity (referred to from
now on as The Sources).4 Both Enoch and Street, arguably the two most influential critics of
Kantian constructivism, however, explicitly stress that their objections to Korsgaard are
meant to apply to Kantian constructivism in general (Enoch 2006, 171; Street 2012, 45). I
therefore discuss their objections as objections against Kantian constructivism in general,
and not as arguments about the details of Korsgaard’s argument. The goal of this chapter is
thus not to defend Korsgaard’s specific variant of Kantian constructivism, but to explore
whether there is a variant of Kantian constructivism that is not susceptible to these
objections.
In addition, as I have already noted in the previous chapter, I focus exclusively on
transcendental arguments from the first person and do not consider whether and how
transcendental arguments from the second person could respond to these objections. My
aim is not to provide a comprehensive overview of the different ways in which Kantian
constructivist could respond to these objections. Instead, I formulate what I take to be the
most promising Kantian constructivist response to these objections. This should be
sufficient because the objections are framed as objections to Kantian constructivism in
general and not just as objections to specific variants of Kantian constructivism.
The structure of the chapter is as follows. In section 2, I briefly introduce the structure
of Korsgaard’s argument for the value of humanity as an example of the kind of argument
that is the object of the objections discussed in this chapter. In section 3, I respond to the
escapability objection by discussing in which sense agency is inescapable. In section 4, I
respond to the shmagency objection. In section 5, I respond to the no-normativity
objection. And finally, in section 6, I respond to the contingency objection.
2. Korsgaard’s Argument for the Value of Humanity
Although the aim of this chapter is not so much to defend one specific transcendental
argument but to show that general meta-ethical objections to Kantian constructivism fail to
convince, it might be helpful to briefly outline the main structure of Korsgaard’s argument
for the value of humanity as an objective moral value. The reason for focusing on
Korsgaard’s transcendental argument is that, as I have already mentioned in the
4
Other explicit targets include, most notably, David Velleman and Connie Rosati (see e.g. Enoch 2006).
GENERAL META-ETHICAL OBJECTIONS
99
introduction to this chapter, even though the objections that I discuss in this chapter are
put forward as general objections to Kantian constructivism or transcendental arguments,
most of these objections are formulated in the context of Korsgaard’s Kantian
constructivism and her transcendental argument for the value of humanity. It might
therefore be helpful to briefly discuss the main structure of Korsgaard’s argument (although
a detailed discussion of Korsgaard’s argument will have to wait until chapters 5 and 6). By
distinguishing between different steps of Korsgaard’s argument, I will also further clarify
how the discussion in this chapter relates to the remainder of the thesis (where I will discuss
questions of interpersonal morality and substantive morality).
In the previous chapter, I briefly introduced Korsgaard’s transcendental argument.
Korsgaard tries to show that any agent, from her first-person perspective, necessarily has to
value her own humanity. In chapter 6, I will discuss the question of what exactly ‘valuing
humanity’ means or entails. What is important for now is that Korsgaard thinks that there
is a categorical reason to value one’s own humanity, insofar as one understands oneself as
an agent, whatever ‘valuing’ one’s ‘humanity’ exactly means. This argument has the
following structure:
1) Agency (X) is inescapable for A.
2) Valuing your (own) humanity (Y) is a necessary condition of the possibility of
agency (X).
3) Therefore, A must value his or her (own) humanity (Y).
Korsgaard, however, stresses that if this argument is successful it “shows only (or at most)
that you must place a value on your own humanity, but not yet that you therefore have
obligations to other human beings” (Korsgaard 1996, p. 130). In other words, if you must
accept whatever are the necessary preconditions of your agency, this does not imply
anything about interpersonal morality, i.e. it does imply that you must also value the
humanity of other agents. An additional argument is needed to arrive at claims of
interpersonal morality.5
Moreover, even if one could provide an argument for interpersonal morality, this
would not yet be sufficient to vindicate a non-trivial account of moral objectivity. After all,
5
Korsgaard herself has provided different arguments for this second step. In chapter 5, I will discuss her argument
from the second person. In chapter 6, I will discuss her argument from the first person.
CHAPTER 4
100
the value of humanity might be nothing more than an empty formalism which does not
generate any substantive constraints on action. Korsgaard and other Kantian
constructivists, however, claim that the necessary preconditions of agency entail certain
substantive constraints on action. Korsgaard, for instance, claims that “it follows from this
argument [the argument for the value of humanity] that human beings are valuable.
Enlightenment morality is true” (Korsgaard 1996, 123). The final step of the argument
would thus be to show that whatever the necessary preconditions of agency are (e.g. valuing
humanity), they lead to concrete and substantive moral permissions, prohibitions,
requirements or constraints.
Based on this brief reconstruction, we can distinguish between the three steps of
Korsgaard’s argument, or any other transcendental argument from the first person:
1) The first step tries to establish that A must value his or her own humanity, or
whatever are the necessary preconditions of agency.
2) The second step tries to establish that A (or any other agent) must also value the
humanity of other agents, or whatever are the necessary preconditions of agency.
3) The third step tries to show that the value of humanity (or whatever the necessary
preconditions of agency are) entails substantive norms of action.
The potential general dilemma for Kantian constructivism is applicable to all three steps; in
all three steps, it has to be shown that there is an inescapable conception of agency which
can lead to one of these conclusions. For instance, even if one thinks that it is possible to
show that one must value one’s own humanity insofar as one is an agent, one might deny
that the necessary preconditions of agency entail a principle of interpersonal morality
and/or that this principle entails any substantive norms of action.
One might wonder which part of this three-step argument is actually a transcendental
argument. I think it is uncontroversial that the argument for the first step is a
transcendental argument. Whether or not the argument for the other steps qualify as
transcendental arguments depends on the details of different arguments that are put
forward, and I will come back to this question in the chapters that follow. Very roughly, I
think that most of the arguments given for the second step are transcendental arguments
themselves or could be conceived as part of the overall transcendental argument for a
principle of interpersonal morality. The third step, on the other hand, is not, as I have
already briefly discussed in the introduction to the thesis, a transcendental argument, but it
GENERAL META-ETHICAL OBJECTIONS
101
tries to show that it is possible to somehow derive substantive norms of action from a
supreme principle of morality – although, as I will argue in chapters 8 and 9, the
justification of a principle of interpersonal morality through a transcendental argument and
the application of this principle should not be understood as completely independent steps.
The focus of the remainder of this chapter is on the first step: the argument which tries
to establish that any agent has to value her own humanity, or whatever are the necessary
preconditions of agency. I will now turn to several meta-ethical objections which all aim to
show that transcendental arguments for categorical claims or reasons are susceptible to
variants of one (or both) of the horns of the potential general dilemma for Kantian
constructivism.
3. The Escapability Objection
The first premise of a transcendental argument states that there is some inescapable aspect
of our self-understanding, such as agency. The first objection that I will discuss denies that
agency is inescapable. It is possible to distinguish between a global and a local variant of
this escapability objection. A global escapability objection claims that there is not a single
conception of agency, however minimal, that is inescapable, where agency is understood,
very roughly, as acting for a reason (Tiffany 2012, 227; Silverstein 2015, 1138). Agency,
according to the global escapability objection, is just a contingent form of selfunderstanding. A local escapability objection, on the other hand, claims that a specific
conception of agency, for instance a conception which stresses the importance of a
diachronic self-conception, is escapable, leaving open the possibility that there might be
other, more minimal conceptions of agency that are inescapable.
In this section, I am concerned with the global escapability objection as it has been put
forward in recent debates in meta-ethics (cf. Silverstein 2015, 1129). The goal of the section
is to respond to the objection that agency is globally escapable (referred to from now on as
escapable) and to use this objection to clarify in which sense agency might be said to be
inescapable. I will come back to local escapability objections in the remainder of the thesis.
In a certain sense, it is obviously false that agency is inescapable. In a recent paper,
Matthew Silverstein, for instance, correctly argues that “agency is all too easy to escape: I
can escape it temporarily by falling asleep or taking a pill, or permanently by getting hit by a
bus or committing suicide” (2015, 1132). Silverstein admits that some ‘escapes’ from
agency, such as committing suicide, are themselves acts, but he stresses that others, such as
CHAPTER 4
102
falling asleep or being hit by a bus, are not. In addition, even though he grants that
committing suicide is itself an act, this, according to Silverstein, does not show that agency
is in fact inescapable. In order to argue for this conclusion, Silverstein puts forward the
following analogy:
Being in my office is my plight so long as I am in my office. I can escape my office
only by performing some action in my office, such as standing up or scooting my
chair toward the door. Any action I might take in order to escape my office will be
an action performed in my office. Does this show that my office is inescapable in
any interesting way? Hardly. My office is not Alcatraz (2015, 1132).
Thus, he concludes that “we can stop acting whenever we want, even though we may have
to act (one more time) in order to do so” (2015, 1132).6
I think Silverstein is right to suggest that there are different ways in which we can stop
being agents. If one is dead, one is obviously not an agent. But one could also imagine
various science fiction scenarios in which our agency is stretched out over so many bodies
that we no longer have to make choices about what to do. In these cases, our capacities for
acting would be infinite, and the limitedness that is a defining feature of our agency would
be overcome (Korsgaard 1989, 115).
However, pace Silverstein, I do not think that these kinds of examples show that agency
is escapable in the sense that is relevant in the context of justifying normative
commitments. That is, I take it that Silverstein is right in suggesting that agency is not
Alcatraz, but agency does not need to be Alcatraz in order to justify normative
commitments. The point of the transcendental argument is not that there is no possible
way in which we could stop being agents. Instead, agency is inescapable in a more
‘practical’ or ‘phenomenological’ sense
There are two closely related reasons to think that agency is inescapable in this sense
6
One of the main goals of Silverstein’s paper is to argue that agency is escapable. Another goal of his paper,
however, is to argue that despite the fact that agency is escapable, constitutivism can still provide an account of the
source of normativity because there is a “deep metaphysical connection between agency and normativity”
(Silverstein 2015, 1141). In this section, I only focus on Silverstein’s objection against the inescapabilty of agency.
The reason for this is that I have to admit that I do not fully understand his theory about the metaphysical
connection between agency and normativity and that insofar as I do understand his metaphysical account it seems
to be incompatible with a transcendental, first-personal justification of morality and therefore lies outside the
scope of this thesis.
GENERAL META-ETHICAL OBJECTIONS
103
(here I follow Gewirth 1978, 25). First, agency is the inescapable subject matter of morality
and practical reasoning, insofar as all practical precepts, including moral precepts, have
action as their object, independent of the content of these precepts (see e.g. Gewirth 1978,
25). Deryck Beyleveld, for instance, writes that “the properties attributed to an agent are
those that a being must have in order to be an intelligible subject of practical precepting (a
being that can ask what it has reason to do)” (Beyleveld 2013, 212f26). Practical precepts,
including moral principles, tell us how to act, and in this sense all practical precepts have
action as their subject matter.7 Second, agency is inescapable for us insofar as we act or
insofar as we engage in practices. Thus, insofar as one engages in any human practice, one
has to accept that one is an agent. This is also where the analogy between stopping being an
agent and leaving one’s office fails. If one leaves one’s office one is still an agent, whereas if
one is no longer an agent one cannot be engaged in any human activity.
Action is thus inescapable from the perspective of practical reasoning, including moral
reasoning, and from the first-person perspective of someone who engages in human
activities. Showing that one might somehow stop being an agent therefore does not
undermine the transcendental argument for practical commitments, but only shows that if
one is not engaged in any human practice one is not committed to any normative
commitment. This conclusion should be no more troubling for an account of normativity
than the idea that normativity does not apply to billiard balls.
Consider, for instances, Silverstein’s example of suicide. Silverstein suggests that agency
is not inescapable, because obviously we can in this way stop being agents, i.e. by
committing suicide. However, pace Silverstein, examples such as suicide are the perfect
illustration of why agency is inescapable in the sense that is most relevant in the context of
morality. Even someone who ends her life is committed to whatever are the conditions of
the possibility of agency. Of course, an additional, separate question is whether suicide
would constitute a violation of these conditions (I do not think it does). In other words, any
normative question, even the question of whether one should end one’s agential life by
committing suicide, presupposes that one understands oneself as an agent. It is true that
one is no longer an agent after one has committed suicide. This implies that one is no
7
Korsgaard makes a similar point: “[T]he normative question is a first person question that arises for the moral
agent who must actually do what morality says. When you want to know what a philosopher's theory of
normativity is, you must place yourself in the position of an agent on whom morality is making a difficult claim.
You then ask the philosopher: must I really do this? Why must I do it? And his answer is his answer to the
normative question” (Korsgaard 1996, 16).
CHAPTER 4
104
longer the object of normative concerns, but it does not show that a transcendental
argument on the basis of agency fails to justify categorical normative commitments.
This discussion raises the question of how exactly to understand the inescapability of
agency. I think this is a question that has received surprisingly little systematic attention
from proponents of transcendental arguments in ethics, even though this question seems to
be central to understanding the Kantian constructivist project. Proponents of
transcendental arguments have tried to describe the relevant inescapability of agency in
different ways. In recent work, Korsgaard describes the inescapabilty, or what she calls the
‘necessity’, of agency as follows:
So action is necessary. What kind of necessity is this? Philosophers like to
distinguish between logical and causal necessity. But the necessity of action isn’t
either of those. There’s no logical contradiction in the idea of a person not acting,
at least on any particular occasion. You could not fail to act, in all the ways I’ve
just described, if there were. And although particular actions, or anyway particular
movements, may have causes, the general necessity of action is not an event that is
caused. I’m not talking about something that works on you, whether you know it
or not, like a cause: I am talking about a necessity you are faced with. Now
sometimes we also talk about rational necessity, the necessity of following the
principles of reason. If you believe the premises, then you must draw the
conclusion. If you will the end, then you must will the means. That’s rational
necessity, and it’s a necessity you are faced with, so that comes closer. But the
necessity of action isn’t quite like that either, for in those cases we have an ifclause, and the necessity of action is, by contrast, as Kant would say,
unconditional. The necessity of choosing and acting is not causal, logical, or
rational necessity. It is our plight: the simple inexorable fact of the human
condition (Korsgaard 2009, 1–2).
Thomas Nagel claims that “I have no confidence that it is a necessary truth that we are
constituted as we are, in the fundamental respects which give rise to our susceptibility to
moral considerations. But if we were not so constituted, we should be unrecognizably
different, and that may be enough for the purpose of the argument” (Nagel 1979, 19).
Habermas claims that (communicative) agency “is so intimately interwoven with the
intersubjective form of life to which subjects competent in speech and action belong that
GENERAL META-ETHICAL OBJECTIONS
105
there is no way we can either posit or bypass it” (Habermas 1990b, 95). That is, Habermas
claims that dropping out of the ‘game’ of agency is, given the kind of beings that we are,
simply “inconceivable, even as a thought experiment” (Habermas 1990b, 100). Finally, Alan
Gewirth claims that escaping from agency means “disavowing and refraining from all
intentional action and from the whole sphere of practice” (Gewirth 1978, 29). Beyleveld
takes this to imply that “the phenomenology of our experience and agency makes it, in
practice, impossible to assume that we are not experiencing a ‘real’ world, or that we are not
[agents]. This fact gives our starting point phenomenological necessity” (Beyleveld 1991,
118).
Although there are important differences between these different authors, they also
have important things in common. First, they all highlight the fact that the starting point of
a transcendental argument should be understood as an aspect of our first-person
perspective. The claim is not that there is no way in which we could stop being agents, but
that we cannot help but understand ourselves as agents. In other words, the claim is that
insofar as we engage in practices and insofar as we have a first-person perspective on the
world, we have to understand ourselves as agents. The importance of the first person is, I
think, best illustrated by Korsgaard, who writes:
The reflective structure of the mind is a source of “self-consciousness” because it
forces us to have a conception of ourselves ... From a third-person point of view,
outside of the deliberative standpoint, it may look as if what happens when
someone makes a choice is that the strongest of his conflicting desires wins. But
that isn’t the way it is for you when you deliberate. When you deliberate, it is as if
there were something over and above all of your desires, something which is you,
and which chooses which desire to act on (Korsgaard 1996, 100).
Korsgaard’s point is that only from the first person, deliberative perspective can it be
understood that we have to understand ourselves as agents. This claim is compatible with
saying that from the third-person perspective it is not at all clear that we are agents.
Second, the authors claim that agency is not just an arbitrary or contingent feature of
our self-understanding, but that it is in an important sense unconditional (it is our ‘plight’).
So what exactly is this important sense in which agency is inescapable? Surely, as should
have become clear by now, agency is not ‘empirically’ inescapable, because one can stop
being an agent. Neither is agency ‘metaphysically’ inescapable, because one might well
CHAPTER 4
106
imagine a possible world in which we do not have to make choices about how to act, for
instance because we might have a plurality of bodies with which to act (Korsgaard 1989).
Agency is only inescapable from our first-person perspective against the background of
certain conditions, e.g. in the case of having a plurality of desires and of having only one
body with which to act. I think it is most accurate to describe this kind of inescapability as
phenomenological inescapability.8
Talking about phenomenological inescapability raises the further question of how to
determine whether something is phenomenologically inescapable. Is phenomenological
inescapability determined by the self-reports of individuals – what they say about how they
experience or understand themselves – or are there other truth conditions of the correct
description of our phenomenology? What can be said, for instance, to a person who denies
that he or she understands herself as an agent – or, to use an analogy from transcendental
arguments in theoretical philosophy, someone who denies that he or she has any
experiences? I think that the kind of phenomenology that is relevant for a transcendental
argument is not so much determined by what people say but is shown in how they act.
Someone who explicitly denies that he or she is an agent has an incoherent selfunderstanding insofar as in denying that he or she is an agent she has to presuppose that he
8
Jonathan Lear’s (1998) work on transcendental arguments explores this tension between inescapability and
contingency in transcendental arguments in the context of theoretical philosophy (Lear 1998). What I take to be
Lear’s main insight into the nature of transcendental arguments is that transcendental arguments always start from
what he calls, following Wittgenstein, “our form of life”, (Lear 1998, 249) and that although our form of life is in
some sense contingent, it is at the same time necessary for us insofar as we cannot step outside our own form of
life. In order to illustrate this point, Lear uses the following example:
“consider, for example, the alternative answers to the following questions:
‘What does 7 + 5 equal?
(a) 12.
(b) Anything at all, just as long as everyone is so minded.
What follows from P and If P, then Q?
(a) Q.
(b) Anything at all, just as long as everyone is so minded.’” (Lear 1998, 249)
Subsequently, Lear goes on to argue that to both questions (a) gives the correct answer. His reason for this
conclusion is that even though the correctness of (a) depends on our specific form of life (which might seem to
indicate that (b) is the correct answer to the questions), there is in fact no position from which to coherently
conceive other possible forms of life in which (a) would not be true. Lear writes: “[T]hese counterfactuals cannot
for us express real possibilities; for, in this broad context, there is no coherent notion of people being ‘otherminded’. The possibility of there being persons who are minded in any way at all is the possibility of their being
minded as we are. Our problem is that being minded as we are is not one possibility we can explore among others.
We explore what it is to be minded as we are by moving around self-consciously and determining what makes
more and less sense. There is no getting a glimpse of what it might be like to be ‘other-minded’, for as we move
toward the outer bounds of our mindedness we verge on incoherence and nonsense” (Lear 1998, 250). Lear’s
seemingly paradoxical conclusion is thus that transcendental conditionals are both contingent and necessary at the
same time.
GENERAL META-ETHICAL OBJECTIONS
107
or she is an agent, because denying one’s agency is itself an act (compare the paradox of the
Cretan). The same can be said of someone who denies that he or she understands herself an
agent while ordering a cappuccino or travelling to work.9
Much more needs to be said about the phenomenology of agency, and I will briefly
come back to this question in chapter 9. For now, I think it is sufficient to stress that the
starting point of a transcendental argument should be understood from the first-person
perspective. Once this is acknowledged it is far less plausible that some sort of experience of
agency is escapable. We cannot but understand ourselves as agents, and the person who
does not or who no longer understands herself as an agent would not be the object of any
practical precepts in the first place.
Of course, this leaves open the further questions of which specific conception of agency
is inescapable and whether this inescapable conception of agency can lead to any interesting
normative conclusions. In this section, I have merely tried to show that global escapability
objections to Kantian constructivism fail. I will discuss more local escapability objections in
chapter 9. The reason for not discussing more local escapability objections at this point is
that for the justification of abstract reasons for action, including a principle of interpersonal
morality, such as the value of humanity, it is sufficient to work with a relatively
indeterminate conception of agency. The question of how exactly this conception of agency
should be specified becomes particularly relevant in the context of justifying substantive
norms of action.
4. The Shmagency Objection
4.1 Enoch’s Objection to Kantian Constructivism
One of the most influential recent critics of Kantian constructivism is David Enoch (2006;
2011a). Although, as I have discussed in the previous chapter, Enoch himself uses a
9
There is a further question of how someone else could determine whether another person has an incoherent selfunderstanding. If the phenomenological inescapability of agency is determined from the first-person perspective,
this seems to be in principle inaccessible to others. Despite everything that I have said, the person who denies that
he or she is an agent while ordering a coffee might be a philosophical zombie from someone else’s point of view. I
think this raises important questions of moral status which cannot be answered within the scope of this thesis. Let
me stress, however, that in the absence of philosophical zombie scenarios, we are normally perfectly capable of
determining whether or not another person is an agent. In addition, there might be precautionary reasons to treat
someone as an agent even if we can never be absolutely sure whether someone is an agent.
CHAPTER 4
108
transcendental argument to argue for moral realism, he criticizes attempts to use a
transcendental argument to justify categorical reasons. The objects of his criticism are those
theories that aim “to derive a full first-order normative (or perhaps just moral) theory from
what is constitutive of action” (Enoch 2006, 170).10 Enoch tries to show that constitutive
features can only generate categorical normative reasons if there is a reason to be an agent
in the first place. In the absence of this reason, the conclusions of a transcendental
argument would be conditional on the contingent fact of agents caring about their agency,
and in that case the conclusions of the argument would only be hypothetically and not
categorically binding. In other words, Enoch believes that a transcendental argument can
only work if we introduce an additional premise (1* below):
1) X is inescapable for A.
1*) X is also normatively inescapable for A, i.e. A has reason to X.
2) Y is a necessary condition of the possibility of X.
3) Therefore, A must Y.
Enoch claims that making explicit premise 1* shows that Kantian constructivism cannot be
a fundamental theory of normativity, because it has to rely on a normative reason to be an
agent to get off the ground.11 If Enoch is right, a transcendental argument, or what he calls a
10
Note that strictly speaking his objection is not directed at transcendental arguments, or at least in the way in
which I defined transcendental arguments in the previous chapter. Enoch defines constitutive arguments as those
arguments that try to derive normativity “from what is constitutive of action” (2006, 170). This might be taken to
suggest that constitutive arguments try to derive normativity from the concept of action. As I have argued in the
previous chapter, however, transcendental arguments do not aim to derive normativity from the concept of action,
but from the self-understanding of agents, i.e. from what any agent is necessarily committed to from his or her
first- person perspective as an agent. It would be premature, however, to conclude that Enoch’s objection does not
apply to transcendental arguments. First, Enoch considers Korsgaard’s argument for the value of humanity as one
of the main examples of the kind of argument he has in mind, and Korsgaard explicitly stresses the importance of
the first person nature of her transcendental argument. Second, in his own work on transcendental arguments in
the context of moral realism, Enoch acknowledges the first person nature of transcendental arguments in general
and claims that his own transcendental argument might also be susceptible to the shmagency objection, or at least
if it did not include a normative reason to engage in the relevant project, i.e. agency or deliberation (2011b, 62,
62f33). Finally, Enoch explicitly notes that he understands constitutive arguments as “variants of transcendental
arguments” (2006, 192f51). In what follows, I will therefore assume that Enoch’s objections to constitutive
arguments also apply to Kantian constructivism.
11
It is not always clear whether Enoch holds that premise 1* should complement premise 1 or whether it should
replace premise 1. Most of his writing suggests the former, because, as I will argue below, Enoch grants the
inescapability of agency, but he also hints at the latter interpretations. Consider, for instance, the following remark
in the context of his own transcendental argument for moral realism: “[I]t’s opting out of the deliberative project
as a whole that may not be an option for us. But even if it is an option for us, it does not seem to be a rational
option for us” (Enoch 2011b, 70–71).
GENERAL META-ETHICAL OBJECTIONS
109
constitutive argument, therefore “cannot do the fundamental [normative] work it was
supposed to do” (2006, 192).12
Enoch introduces his objection to transcendental arguments by imagining a sceptic,
what he calls ‘a shmagent’, as saying, “Classify my bodily movements and indeed me as you
like. Perhaps I cannot be classified as an agent without aiming to constitute myself [or
without doing whatever is a necessary condition of the possibility of agency]. But why
should I be an agent?” (Enoch 2006, 179).13 A shmagent is thus someone who accepts that
agency has certain necessary conditions of possibility, but who denies that these generate
any categorical normative reasons, because he or she simply does not have a reason to be an
agent.
In order to argue for this conclusion, Enoch draws an analogy with chess. Enoch claims
that if one is playing chess one is committed to the constitutive rules of chess, such as the
goal of trying to checkmate one’s opponent. If not, one would not be playing chess:
checkmating one’s opponent is one of the features that makes playing chess a game of chess
instead of just a random movement of pieces on a board. However, Enoch suggests that
even though checkmating your opponent might be one of the constitutive aims of chess,
this does not mean that you necessarily have a reason to checkmate your opponent. This, so
Enoch argues, follows only if one has a reason to be playing chess in the first place. That is, a
violation of the constitutive principles of chess is only normatively problematic if one has a
reason to play chess. If not, one is not rationally committed to its constitutive aims and
principles. According to Enoch, essentially the same idea applies to the constitutive features
of agency. One might have no reason for being an agent (one might be a happy shmagent),
and therefore one might not necessarily be committed to its constitutive features.
Enoch’s objection to Kantian constructivism (or constitutivism, as he calls it) is thus
that whatever are the necessary preconditions of agency, these can only generate reasons for
action insofar as one has a reason to be an agent in the first place. It is important to stress
12
I think this conclusion needs to be qualified in an important way because it seems to me that Humean
constructivism or constitutivism (e.g. Street 2012) is not susceptible to Enoch’s general objections to
constitutivism, because a Humean believes that all normative reasons are a function of contingent evaluations,
such as caring about one’s agency. In what follows, I will limit the discussion to Kantian versions of constitutivism
which aim to justify categorical normativity.
13
Enoch also defines a shmagent as a “nonagent who is very similar to agents but who lacks the aim (constitutive
of agency but not of shmagency) of self-constitution.” (Enoch 2006, 179). I think this is misleading because, as I
argue below, Enoch grants both that a shmagent is also an agent and that agency has certain constitutive aims.
CHAPTER 4
110
that Enoch not only accepts that agency has certain necessary preconditions (2006, 177–78)
but also that agency itself is phenomenologically inescapable, which implies that the
shmagent is also an agent:
“Perhaps,” Korsgaard’s sceptic may say, “I cannot opt out of the game of agency,
but I can certainly play it half-heartedly, indeed under protest, without accepting
the aims purportedly constitutive of it as mine.” The kind of necessity the game of
agency has to enjoy in order to solve the problem we are now in is normative
necessity. Invoking other necessities here will just not do (Enoch 2006, 188; see
also Enoch 2011a, 216).
The shmagent, Enoch suggests, is an agent and he or she does not question the
(phenomenological) inescapability of agency, i.e. the shmagent acknowledges that he or she
cannot opt out of the game of agency, or at least not in a relevant sense (pace Silverstein).
The shmagent, however, questions the normative necessity or the normative inescapability
of playing the game of agency.14 This is also where the analogy between chess and agency
breaks down.15 Chess is escapable and, presumably, there is also no necessary reason to play
chess. Agency, on the other hand, is, according to Enoch, inescapable, but there might be
no reason to be an agent despite its inescapability. Summarizing his shmagency objection in
other writings, Enoch stresses that “here as everywhere else, normative input is needed if
normative output is to be secured” (Enoch 2011a, 62).
The message should be clear by now: according to Enoch, the phenomenological
inescapabilty of X is not sufficient to generate normativity, or at least not the categorical
normativity both the moral realist and the Kantian are after. In the absence of a normative
reason to X, i.e. a reason be an agent, the necessary conditions of agency should only be
accepted by those who contingently happen to care about their agency in the first place
(just like the constitutive rules of chess are only normative for people who have a reason to
play chess).
Does Enoch’s shmagent undermine Kantian constructivism? That is, assuming for the
14
Cf. “the kind of non-optionality [i.e. inescapability] directly relevant here is normative. The point is that the
project [e.g. agency or deliberation] is one we have strong reason not to disengage from, perhaps one that we
rationally ought not to disengage from” (Enoch 2011a, 62).
15
See also Ferrero (2009, 308), who argues that the disanalogy between chess and agency takes away much of the
initial appeal of the shmagency objection.
GENERAL META-ETHICAL OBJECTIONS
111
moment that Enoch is right that a transcendental argument needs to rely on a reason to be
an agent, does this show that Kantian constructivism fails? Although this might seem rather
obvious (which might also explain why this question is rarely if ever discussed in the
literature on the shmagency objection), it might nevertheless be helpful to briefly discuss
the two closely related arguments that Enoch provides for this conclusion. First, and
negatively, Enoch argues that if Kantian constructivism aims to ground normativity in what
is constitutive of agency, and if we need a reason to be an agent in order to have a reason to
accept the constitutive features of agency, Kantian constructivism is incomplete as a
method used for the justification of (categorical) normative reasons. Enoch writes that “the
constitutivist strategy cannot give us the whole story of normativity. It is no longer possible
that all practical reasons are grounded solely in what is constitutive of agency. Normativity,
in other words, cannot be grounded in what is constitutive of agency” (2006, 186–87).
Enoch’s point is thus that Kantian constructivism cannot on its own terms justify a reason
to be an agent. Instead, Kantian constructivism has to presuppose this reason to be an agent
in order to get off the ground. This shows that Kantian constructivism cannot give us the
whole story of normativity.
Second, and closely related to the first argument, Enoch argues that a robust normative
realist truth could ground a reason to be an agent, but only by undermining Kantian
constructivism as a fundamental theory of normativity:
[T]hings would have been different, of course, had it been possible to invoke here
a normative truth, robustly realistically understood, to the effect that we all have a
reason to be agents. With this normative claim in hand, the rest of the
constitutive-of-agency line can be pursued rather safely. And nothing I say here
goes to deny that even with such a normative claim in hand questions about the
constitutive aspects of action may be interesting and indeed normatively relevant.
But introducing such a normative truth – itself not accounted for by the
constitutivist strategy – defeats the motivations underlying this strategy: it renders
Korsgaard unable to cope with the skeptic who is left indifferent to the purported
reason she has to be an agent (2006, 187).16
16
See also “the most natural way of defending the normative non-arbitrariness of such things [e.g. agency] is by
invoking a general, constitutivism-independent reason to be an agent” (Enoch 2011a, 229).
CHAPTER 4
112
Enoch thus strongly suggests that only robust normative realism could ground a normative
reason to be an agent and that this shows that Kantian constructivism is parasitic on robust
realism.
In other words, the shmagency objection undermines Kantian constructivism because
Kantian constructivism cannot justify a reason to be an agent on its own terms. Instead,
Kantian constructivism (implicitly) has to rely on a robust moral realist reason to be an
agent and thereby undermines its own ambition, i.e. to provide a justification of objective
moral judgements without relying on moral realism. How convincing is Enoch’s criticism
of Kantian constructivism?
4.2. Why Only Moral Realists Believe in Shmagents
In this section, I put forward two responses to Enoch’s objection. The first response is that
even if we assume, for the sake of argument, that Enoch is right in claiming that a
transcendental argument has to rely on a reason to be an agent, it is not obvious that this
would undermine the Kantian constructivist project. The reason for this is that it is unclear
how exactly Enoch understands this ‘reason to be an agent’ and why it would be impossible
for a Kantian constructivist to justify such a reason on its own terms. Let me stress that the
point of this first response is not so much to defend such a Kantian constructivist reason to
be an agent (because ultimately I think this is neither possible nor necessary), but to point
out that, even on its own terms, Enoch’s criticism of Kantian constructivism is incomplete
as it stands. The second, and more fundamental, response is that transcendental arguments
do not have to rely on a reason to be an agent in order to generate categorical reasons and
that Enoch fails to give a non-question begging argument for the contrary claim.
The first response to Enoch’s objection is that it is not obvious that even if we assume
that a transcendental argument has to rely on a reason to be an agent in order to generate
categorical reasons, this would undermine Kantian constructivism.17 Enoch simply assumes
that Kantian constructivists cannot, on their own terms, provide a reason to be an agent. I
17
In Taking Morality Seriously, his book published after the publication of Agency, Shmagency, Enoch stresses that
the relevant kind of inescapability is “normative” (2011b, 62) or “rational” (2011b, 71) inescapability, but he does
not elaborate on what exactly he means by this. In fact, in a footnote to his claim that a transcendental argument
should start from rational inescapability, Enoch admits that he is himself not exactly sure what he means by this:
“[L]et’s be honest: it’s not that I’ve said so little because ‘saying more would take me too far astray’, or any such
thing. I just don’t know what more to say” (Enoch 2011a, 71f51). If Enoch himself cannot explain what exactly he
means by saying that a starting point of a transcendental argument should be normatively or rationally
inescapable, then why should we think that it undermines Kantian constructivism?
GENERAL META-ETHICAL OBJECTIONS
113
agree with Enoch that it is hard to see how a Kantian constructivist could possibly justify a
reason to be an agent on Kantian constructivist terms, given that the starting point of the
transcendental argument is not a normative premise but the phenomenological
inescapability of agency. It therefore seems rather mysterious how a Kantian constructivist
could justify a reason to be an agent, because transcendental arguments always start from a
phenomenologically inescapable starting point and not from a normatively inescapable
starting point.18 In addition, I also agree with Enoch that if a Kantian constructivist were to
implicitly or explicitly rely on a moral realist reason to be an agent, this would undermine
the Kantian constructivist project.
But things are not as obvious as they might look in the first instance. In order to
illustrate this point, let me briefly come back to Enoch’s own transcendental argument for
normative realism. The reason for going back to this argument is to show that there is
tension between Enoch’s criticism of Kantian constructivism, in which he at least strongly
suggests that the reason to X (e.g. deliberate or be an agent) should be understood in moral
realist terms, and his own transcendental argument for normative realism, in which he
explicitly denies that the reason to X should be understood in realist terms. Much of the
force of Enoch’s criticism of Kantian constructivism is derived from his suggestion that
Kantian constructivism has to rely on a moral realist reason to get off the ground. In the
absence of the need to rely on a moral realist reason to be an agent, Kantian constructivism
does not seem to be susceptible to the shmagency objection.
Recall that in Taking Morality Seriously Enoch tries to give a transcendental argument,
the argument from deliberative indispensability, for the existence of mind-independent
normative truths. His transcendental argument on the basis of deliberative indispensability
starts from the normative or rational inescapability of deliberation (see e.g. Enoch 2011b,
62 & 71) – if it did not start from normative inescapability it would be susceptible to his
very own shmagency objection. However, if this normative reason were a robust realist
reason (as he suggests in his (2006) paper on the shmagency objection), his argument from
deliberative indispensability would be obviously question-begging, because in that case his
18
One could interpret the conclusion of Korsgaard’s argument, i.e. that you must value one’s own humanity, as
showing that one has a reason to be an agent, depending, of course, on how exactly one understands the concepts
of ‘value’ and ‘humanity’ (see chapter 6). However, even if this is the best interpretation of the conclusion of
Korsgaard’s argument, it is not sufficient to respond to Enoch’s objection. The problem is that the argument for
the value of your own humanity/the reason to be an agent would still start from the phenomenological
inescapability of agency. Rather paradoxically, Enoch would, in that case, stress that one needs a reason to be an
agent in order to argue transcendentally for a reason to be an agent.
CHAPTER 4
114
argument for robust realism would simply presuppose robust realism. Enoch, however,
anticipates this objection in the following way:
[M]y indispensability argument for Robust Realism will rely on a normative
premise (that the deliberative project is rationally non-optional). But isn’t it
problematic – question-begging, perhaps – to use a normative premise in an
argument for Robust Metanormative Realism? Against most alternative views, the
answer is “no”, for the premise is the normative proposition itself, not a robust
realist understanding of it (Enoch 2011b, 62).19
If I understand him correctly, Enoch’s point is that although any transcendental argument
has to rely on a normative premise as a starting point, this normative premise is neutral
among (most) meta-ethical positions. If this is true, this means that his transcendental
argument for robust normative realism does not beg the question, because it would not
presuppose realism, but the normative proposition that deliberation is rationally
inescapable.
It should be clear by now that there is a tension between Enoch’s suggestion that
Kantian constructivism needs to rely on a moral realist reason to be an agent and his claim
that the reason to deliberate is meta-ethically neutral. I think this tension presents Enoch
with a dilemma. He could allow for the possibility that the reason to be an agent is metaethically neutral, but in that case the shmagency objection does not seem to undermine
Kantian constructivism. After all, just like Enoch’s claim that the reason to deliberate is
neutral among meta-ethical positions, the Kantian constructivist can likewise claim that the
reason to be an agent is meta-ethically neutral. In that case, Kantian constructivism could
still be the fundamental (meta-ethical) theory of normativity, even if the Kantian
constructivist would admit that a transcendental argument has to rely on a reason to be an
agent. Alternatively, Enoch could return to his suggestion that the reason to be an agent
should be understood as a robust realist reason after all, but in that case his own
transcendental argument for robust realism would be question-begging in the way
described above (as Enoch himself admits would be the case). In other words, either
Kantian constructivism is off the hook (if the reason to be an agent is understood as a
19
The exception is error theory: “against error theorists there may be some question-begging involved, but not of
an objectionable kind” (Enoch 2011b, 62). For the sake of argument, I assume that Enoch is right that even in
relation to error theory his argument is not objectionably question begging.
GENERAL META-ETHICAL OBJECTIONS
115
normative proposition) or the shmagency objection also undermines Enoch’s own
argument for robust moral realism (if the reason to be an agent is understood in terms of
robust realism).
Of course, from this it does not follow that the shmagency objection fails, because it
might well be that it ultimately undermines both Kantian constructivism and Enoch’s
transcendental argument for robust realism. It does, however, make clear that Enoch needs
to clarify what exactly he means by ‘a reason to be an agent’ before his objection can be
successful. I will not, however, pursue this question further here because I think that there
is a second and more fundamental problem with Enoch’s objection.
So let me now turn to my second response to Enoch’s objection. Until this point, I have
assumed, for the sake of argument, that Enoch is right that a transcendental argument has
to rely on a reason to be an agent. I think, however, that there are good reasons to question
this conclusion. The question is this: why should we think that a transcendental argument
needs to rely on a reason to be an agent for it to justify categorical normative reasons?
In ‘Agency, Shmagency,’ Enoch assumes that if we can somehow ‘imagine’ a sceptic
who asks ‘Why should I be an agent?’, this would be sufficient to prove that a
transcendental argument needs to rely on a reason to be an agent to get off the ground.
Enoch argues that since the question ‘Why should I be an agent?’ “seems to make sense”,
we are “justified in taking it at face value” (2011a, 228). Although I agree with Enoch that
the shmagency question ‘makes sense’, I think that the fact that a question ‘seems to make
sense’ is not sufficient for it to pose a problem for Kantian constructivism.
My reason for believing this is that although the shmagency question makes sense, it
only poses a problem for Kantian constructivism if the question is one that can and ought to
be answered by any account of normativity. However, one will only be inclined to think that
this question can be and needs to be answered if one thinks that there are normative
reasons that exist independent of the perspective of agents, i.e. if one thinks that moral
realism is true.20 In the absence of a commitment to moral realism, it seems to be sufficient
to show that, from her first-person perspective, any agent necessarily has to accept the
necessary preconditions of agency, such as valuing her humanity. My point is thus that the
20
This should be qualified because, as I have argued above, it is not evident that a reason to be an agent implies
that there are robust normative reasons. However, as I have also argued above, if a reason to be an agent is metaethically neutral it does not pose a problem for Kantian constructivism in the first place. I therefore assume that
Enoch wants to say that a reason to be an agent commits Kantian constructivism to robust realism because this
would be the only way in which the shmagency objection constitutes an objection to Kantian constructivism.
CHAPTER 4
116
only person who will think that this is insufficient is the moral realist, who believes that
normativity has to come from mind-independent facts. 21
Kantian constructivists should only be worried by the shmagency objection if they
would think that there is such a thing as a mind-independent reason to be an agent. The
problem with Enoch’s objection is thus that thinking that we need to give a reason to be an
agent presupposes a meta-ethical standpoint that Kantian constructivists deny is available,
or whose availability is at least bracketed by Kantian constructivism.
22
Kantian
constructivists do not believe that there are normative reasons that are independent of the
first-person perspective of agents. On the Kantian constructivist conception of normativity,
the question ‘Why should I be an agent?’ cannot have an answer, but this is no
embarrassment, because the Kantian constructivist does not believe that there are agencyindependent reasons in the first place (and if an answer can be given on Kantian
constructivist grounds, then raising the question obviously poses no problem for Kantian
constructivism). Enoch, however, is a robust realist, i.e. he believes that there are mind- or
agency-independent reasons for action. For Enoch, the question ‘Why should I be an
agent?’ ought to have an answer. However, because the shmagency objection crucially
hinges on the existence of agency-independent reasons, it is not an argument against
Kantian constructivism, but simply assumes that Kantian constructivism is wrong from the
very start. Of course, this does not prove that Kantian constructivism is the correct theory
of normativity, but proves only that the plausibility of the shmagency objection itself
depends on the plausibility of moral realism and cannot therefore be used as an
independent argument against Kantian constructivism. As a consequence of this, the
Kantian constructivist can justifiably shrug and ignore the moral realist who asks ‘Do I have
a reason to be an agent?’.
Nothing I have said so far shows that an agent, simply insofar as he or she understands
him- or herself as an agent, necessarily has to accept certain normative judgements. Nor
will my reply convince the moral realist according to whom the shmagency question does
and should have an answer. The only point I want to make in this section is that the fact
that one can intelligibly raise the question of whether one should be an agent does not show
that a Kantian constructivist ought to take this question seriously. Instead, whether or not
one finds the shmagency objection convincing will ultimately depend on whether one is a
21
Ferrero (2009, 311) hints at a similar response by saying that “there is no standpoint external to agency that the
shmagent could occupy and from which he could launch his challenge.”
22
Tubert (2010, 663–64) hints at a similar response to Enoch.
GENERAL META-ETHICAL OBJECTIONS
117
moral realist or a Kantian constructivist in the first place.
As I have explained in the introduction to this thesis, the goal of this thesis is not so
much to argue against moral realism but to constructively investigate the possibility of
justifying moral objectivity without relying on (robust) moral facts. There might, therefore,
even be a reason for moral realists to suspend judgement about the shmagency objection.
After all, if it is possible to justify moral objectivity without relying on moral facts, this
might even convince some moral realists to become constructivists. If not, then at least the
discussion of the shmagency objection shows that we have reached rock bottom: the realist
and the Kantian constructivist differ fundamentally on their overall moral–metaphysical
outlook. The consequence of this, though, is that the shmagency objection ends up only
convincing those with a moral realist outlook.
5. The No-Normativity Objection
The escapability objection and the shmagency objection both criticize, although in different
ways, the starting point of a transcendental argument. The escapability objection claims
that agency is escapable. The shmagency objection claims that even if agency is
(phenomenologically) inescapable, this is not sufficient to show that one has a reason to be
an agent, i.e. the shmagency objection claims that agency is potentially normatively
escapable and that the normative inescapability of agency should be the starting point of a
transcendental argument.
In the next two sections, I will turn to two objections which grant the starting point of a
transcendental argument, i.e. which grant that agency is inescapable, but which
subsequently deny that any categorical reasons can follow from this starting point through
a transcendental argument. The first objection, which is put forward by Sharon Street, is
that a transcendental argument cannot lead to any normative conclusions (the nonormativity objection). The second objection, put forward by Street and others, states that
not every agent actually accepts Y, e.g. not every agent values her humanity, and that this
fact undermines the validity of the argument (the contingency objection).
5.1 Street’s No-Normativity Objection to Kantian Constructivism
A transcendental argument aims to show that any agent necessarily has to accept whatever
are the necessary conditions of agency and that these necessary conditions of agency
CHAPTER 4
118
include categorical claims or reasons. Korsgaard, for instance, claims that any agent must
value her own humanity. What I will call ‘the no-normativity’ objection states that the
claim ‘if you are an agent, you must Y’ – e.g. Korsgaard’s claim ‘if you are an agent, you
must value your humanity’ – is ambiguous and ultimately incoherent (Street 2012, 51).23
This objection has recently been put forward by Street. In this section, I respond to this
objection and try to clarify the way in which a transcendental argument can lead to
categorical normative claims.
In order to argue for the conclusion that a transcendental argument cannot lead to any
categorical normative claim, Street distinguishes two possible interpretations of the
statement ‘if you are an agent, you must Y’, and she argues that both fail to justify a
categorical normative claim.
On the first, normative, interpretation, such claims “state substantive normative truths”
(Street 2012, 51), meaning that there is a mind-independent reason to value your humanity.
Thus, according to the normative interpretation, the claim ‘if you are an agent, you must Y’
should be read as ‘if you are an agent, you have a mind-independent, substantive reason to
Y’. The problem with this interpretation should be obvious. Insofar as constructivism aims
to provide an alternative for moral realism, it should not presuppose any substantive
normative truths in its argument for objective moral claims. Street therefore concludes that
the normative interpretation is unavailable for constructivists.
On a second, conceptual, interpretation, on the other hand, “these claims state
conceptual, ‘constitutive’ truths, such as ‘To be a parent, one needs to have children’”
(Street 2012, 51). If the claim ‘if you are an agent, you must Y’ is read as a conceptual claim,
the argument clearly does not presuppose any substantive, mind-independent reason. In
this sense, the conceptual claim is available to Kantian constructivists (whereas the
23
Michael Nance (2015) puts forward an almost identical objection to Fichte’s transcendental argument for the
necessity of mutual recognition. He presents the objection in terms of what he calls a ‘modal dilemma’: “The
modal dilemma goes as follows. Fichte understands it to follow from his argument that ‘I must in all cases
recognize the other as a free being, i.e., I must limit my freedom through the concept of the possibility of his freedom’
(ibid.: 39). There is an ambiguity concerning the nature of the ‘must’ in this conclusion. The necessity expressed by
the ‘must’ is either metaphysical or normative. If the necessity is metaphysical, then Fichte’s conclusion is falsified
by empirical reality, since one can evidently be a finite self-positing agent without recognizing the rights of others.
... But if the necessity expressed by the ‘must’ is normative—if his conclusion means ‘I ought in all cases to
recognize the other’—then it does not follow from the premises about self-positing that Fichte establishes in the
preceding theorems of his deduction” (Nance 2015, 615–16). This seems to suggest that Street’s objection to
Korsgaard’s argument is an instantiation of a more general objection to transcendental arguments in ethics. I will
not discuss the applicability of this no-normativity or modal objection to Fichte. Instead I will try to show that
there is an interpretation of transcendental arguments in ethics which is not susceptible to this kind of objection.
GENERAL META-ETHICAL OBJECTIONS
119
normative interpretation is not). The problem with the conceptual interpretation, Street
argues, is, however, that it is merely conceptual, i.e. it does not generate any reasons for
action. Street writes that “an ‘agent’ who doesn’t take anything at all to be a reason does not
have a reason to take something or other to be a reason; rather, ‘he’ or ‘she’ is not an agent
at all” (Street 2012, 51). Street’s objection is thus that the transcendental claim either
presupposes moral realism because it relies on a ‘substantive normative truth’ or does not
generate any reason for action because it is merely a conceptual claim.
Street’s incoherency objection to Kantian constructivism should be distinguished from
the shmagency objection. Although Street also questions the normativity of Kantian
constructivism, her objection is directed at a different part of the Kantian constructivist
argument. Street does not think that one needs a reason to be an agent in order to generate
normativity from the standpoint of agency – in fact, her own Humean constructivism also
presupposes that normative reasons can be generated from the perspective of agents (I
come back to this below). Street’s objection is directed at premise 2 of the transcendental
argument, which is that ‘Y is a necessary condition of the possibility of X’. More
specifically, she denies that the necessary conditions of the possibility of agency express a
normative relation – not because we need a reason to be an agent in the first place, but
because she thinks that the necessary conditions of possibility express merely a conceptual
relation (or rely on a realist, normative truth).
5.2 Categorical Instrumental Reasons
Street’s dilemma regarding Kantian constructivism is that either it cannot justify normative
reasons or it relies on a mind-independent normative reason. However, this supposed
dilemma overlooks the possibility that there might a third way to interpret the claim that ‘if
you are an agent, you must Y’. There might be normative reasons that are not derived from
mind-independent reasons but which are generated from the first-person perspective of the
agent.
In this subsection, I elaborate one possible way in which Kantian constructivism could
argue for this conclusion. The proposal is to understand the necessary conditions of the
possibility of agency in terms of categorical instrumental reasons. I argue that although
most instrumental reasons are hypothetical, insofar as they depend on pursuing a
contingent end, there are also instrumental reasons involved in pursuing any end
whatsoever. These are what I call categorical instrumental reasons. Note that this reply
CHAPTER 4
120
crucially hinges on the acceptance of the principle of instrumental rationality. I briefly
come back to the justification of the instrumental principle towards the end of the section.
The goal of this subsection is, however, limited. I aim to show that Street, or any other
philosopher with Humean (constructivist) sympathies, has to accept the coherency of the
Kantian constructivist project, given their own philosophical commitments, which include,
as I argue below, the acceptance of the principle of instrumental rationality. A more
comprehensive defence of Kantian constructivism will follow in the remaining chapters.
The account of Kantian constructivism presented in this subsection is very much in
line with Alan Gewirth’s (1978) interpretation of transcendental arguments (I will discuss
Gewirth’s transcendental argument in detail in chapter 7). It is questionable whether
Korsgaard would or could agree with this account of Kantian constructivism.24 In this
subsection I am not, however, so much concerned with this exegetical question, but with
trying to provide one possible Kantian constructivist response to Street’s objection.
Consider, first, contingent instrumental reasons. The principle of instrumental
reasoning states that if I want to achieve an end, I have a reason to take the necessary means
to that end or give up the end. For instance, if I want a drink, and there is only water
available, I have a reason to drink water. Of course, once I discover that there is only water
available, I might revise my end and decide that I would rather be thirsty than drink water,
but insofar as I want a drink, I ought to drink the water.
Contingent instrumental reasons are obviously contingent on pursuing a specific end.
Only someone who wants a specific end has a reason to take the necessary means to that
end, and once someone discovers what the necessary means to that end are, he or she might
decide to revise the end instead of taking the necessary means to that end. So although
contingent instrumental reasons are compatible with constructivism, insofar as
instrumental reasons do not presuppose substantive normative truths but only agents who
pursue ends, it is not sufficient to argue for constructivist, objective reasons.
In order to argue for Kantian constructivism, an additional step is required. The
24
Korsgaard argues that “the instrumental principle cannot stand alone” (1997, 251). Instead, Korsgaard claims
that the instrumental principle can only generate normativity if there are already ends that are good. This
conception of the role of the instrumental principle is incompatible with my response in this section, because my
response claims that (categorical) instrumentality can generate normativity without presupposing anything about
the goodness of ends. At the same time, however, Korsgaard also believes that the distinction between the
instrumental principle and the categorical imperative might “break down” because “moral or unconditional
principles and the instrumental principles are both expressions of the basic requirement of giving oneself a law,
and bring out different implications of that requirement” (1997, 250f73). This seems to bring her view closer to my
view.
GENERAL META-ETHICAL OBJECTIONS
121
additional step is the recognition that there might not only be instrumental reasons that are
conditional on the acceptance of a contingent end but that there might also be instrumental
reasons involved in adopting any end whatsoever. For instance, being alive is a necessary
means for engaging in any action whatsoever. Independent of whether one wants a drink,
wants to play chess, study philosophy or start a family, etc., a necessary means for acting,
whatever one’s contingent ends, is that one is alive. Even someone who commits suicide has
to be alive in order to achieve his or her end (cf. Gewirth 1978, 136–37). In this sense, one
has a reason to want to be alive, independent of one’s contingent ends.
Insofar as one has a reason to take the necessary means to achieve an end, and insofar
as there are necessary means for engaging in any action whatsoever, there are categorical
reasons to take the necessary means for engaging in action as such. The only difference
between contingent and categorical instrumental reasoning is that whereas contingent
instrumental reasons only exist insofar as someone wants a specific end, categorical
instrumental reasons exist insofar as there are beings who want to achieve any end
whatsoever. The categorical instrumental means are categorical insofar as they do not
depend on one’s contingent desires or preferences, but only on one’s agency as such. Pace
Street (2012, 53) and many others, saying that a reason is categorical therefore does not
have to mean that a reason is purely non-instrumental. It could be, but it need not be.
Instead, saying that a reason is categorical means that it should be accepted independent of
one’s contingent preferences or desires, not that it should be accepted independent of being
an agent.
By forcing the Kantian constructivist in the dilemma to choose between relying on
substantive normative truths or expressing merely conceptual truths, Street neglects the
crucial fact that Kantian constructivists try to justify objective reasons from the perspective
of the self-understanding of agents. Korsgaard, for instance, describes the importance of the
first-person perspective as follows:
Value, like freedom, is only directly accessible from within the standpoint of
reflective consciousness. And I am now talking about it externally, for I am
describing the nature of the consciousness that gives rise to the perception of
value. From this external, third-person perspective, all we can say is that when we
are in the first-person perspective we find ourselves to be valuable, rather than
simply that we are valuable. There is nothing surprising in this. Trying to actually
see the value of humanity from the third-person perspective is like trying to see the
CHAPTER 4
122
colours someone sees by cracking open his skull. From outside, all we can say is
why he sees them (Korsgaard 1996, 124).
In other words, it would be a mistake to think that the claim ‘if you are an agent, you must
Y’ expresses merely a conceptual truth. This would be to take an external, third-person
perspective on instrumental reasoning. The claim ‘if you are an agent, you must Y’ is a
claim about what any agent is necessarily committed to from his or her first-person
perspective as an agent, i.e. as someone who wants to achieve ends (whatever the
contingent content of those ends).
Note that when I say that the transcendental claim (‘you must Y’) is not merely a
conceptual truth, I do not want to deny that the instrumental principle might also be a
conceptual truth. What is important, however, is to notice that even if the instrumental
principle is a conceptual truth, it does generate normativity if it is applied to the firstperson perspective of an agent. From the first-person perspective of someone who tries to
achieve certain purposes, the instrumental principle generates normativity because it claims
that one ought to take the necessary means to one’s goals. This is what I mean when I say
that a transcendental must is not merely a conceptual truth – this contrasts with the claim
that ‘to be a parent one needs children’, which is merely a conceptual truth, because it does
not generate any reasons for action even when understood from the first-person
perspective of an agent.
In my response to Street, I have relied on the normativity of the instrumental principle.
One might legitimately raise the question of the source of the normativity of the
instrumental principle. I will briefly come back to this question at the end of this section,
but let me highlight that, as I have already stressed at the beginning of this section,
responding to Street’s specific objection does not require arguing for the instrumental
principle. The reason for this is that Street herself accepts the normativity of the principle of
instrumental reasoning, and she believes that it can be given a constructivist foundation
(2008, 228). Consider, for instance, her remark that according to her Humean
constructivism, “the agent might of course still be making an instrumental mistake as
dictated by the terms of his other values” (Street 2012, 55f23).25 In other words, Street’s
objection does not question the normativity of the instrumental principle, and because she
25
In fact, in other writings, Street acknowledges that the instrumental principle can generate reasons for action,
but at the same time she claims that the instrumental principle is a conceptual truth (see Street 2008, 228f37).
GENERAL META-ETHICAL OBJECTIONS
123
herself relies on the normativity of the instrumental principle it seems legitimate to accept
its normativity in the context of discussing her objection to Kantian constructivism.
This has a surprising implication. For if I am correct in believing that a transcendental
argument should be understood in terms of categorical instrumental reasoning, this
actually presents Street, and not the Kantian constructivist, with a dilemma. One option is
for Street to reject the normativity of the instrumental principle, but this rejection backfires
insofar as Street’s Humean constructivism also presupposes the normativity of the
instrumental principle. Rejecting the instrumental principle therefore not only undermines
Kantian constructivism but also Humean constructivism, which seems to imply that
denying the instrumental principle might lead to scepticism about normativity (at least
assuming, as Street also does, that moral realism is off the table as well).26
Alternatively, Street could accept the instrumental principle, but then she would have
to accept that Kantian constructivism is a coherent project, because the normativity of
categorical instrumental reasoning is not more mysterious than the normativity of
contingent instrumental reasoning. Opting for the second horn of the dilemma, however,
similarly undermines Street’s Humean constructivism. After all, Street’s argument for
Humean constructivism is essentially premised on the rejection of Kantian constructivism.
Street writes:
[Humean constructivism] is defined negatively in contrast to Kantian
constructivism: it is characterized by a general skepticism about arguments which
seek to establish moral values as following from the standpoint of “agency” as
such. Therefore, the obvious way, and perhaps the only way, to argue for Humean
as opposed to Kantian constructivism is negatively—by examining the best
attempts to establish such substantive values and showing why they fail (Street
2012, 45).
Accepting that Kantian constructivism is a coherent position therefore not only implies
that Street’s objections against Kantian constructivism fail but that her main argument in
favour of Humean constructivism fails as well. In short, Street has to choose between
26
Another option for Street would be to retreat to what Bernard Williams (1979) calls the “sub-Humean model”
(1979, 102), which states that any desire also constitutes a reason. This, however, also seems to amount to
scepticism about normativity insofar as on the sub-Humean model one cannot make a mistake about one’s
reasons.
CHAPTER 4
124
scepticism about normativity or Kantian constructivism. It should be clear which horn of
the dilemma is to be preferred.
But what does a Kantian constructivist have to say to a sceptic about instrumental
normativity? The dominant Kantian constructivist justification of the instrumental
principle is to argue that accepting the instrumental principle is itself constitutive of
pursuing a purpose or having an end, i.e. that it is constitutive of agency.27 The argument
would go roughly as follows:
1) Agency is inescapable for A, i.e. I inescapably try to achieve purposes/will an end
from my first-person perspective (X).
2) Taking the necessary means to an end (Y) is a necessary condition of trying to
achieve a purpose/willing an end (X).
3) Therefore, A must take the necessary means to A’s ends (Y).
I will come back to a transcendental justification of the instrumental principle in chapter 7,
although I already ought to warn the reader that I do not try to provide a full defence of the
instrumental principle in this thesis. Instead, in most of this thesis, I assume that such a
justification can be given and subsequently explore the possibility of justifying claims about
interpersonal and substantive morality. The reason for this is that although a justification of
the instrumental principle is not uncontroversial (for a recent overview see Kolodny and
Brunero 2016), worries about Kantian constructivism typically focus on its ambition to
justify categorical and universal claims about interpersonal morality and not so much on its
ability to justify instrumental normativity. In other words, I do not try to fully convince
those who are sceptical about instrumental normativity in this thesis, although I will give
some reasons for accepting the instrumental principle in chapter 7. But given that most of
us are not sceptical about instrumental normativity in the first place, this is not a major
concession.
The point of this section is to show that Kantian constructivist normativity is not more
27
See e.g. “willing an end just is committing yourself to realizing the end. Willing an end, in other words, is an
essentially first-personal and normative act. To will an end is to give oneself a law, hence, to govern oneself ...
Willing an end is equivalent to committing yourself, first-personally, to taking the means to that end ... What is
constitutive of willing the end is not the outward act of actually taking the means but rather the inward, volitional
act of prescribing the end along with the means it requires to yourself” (Korsgaard 1997, 245). See also (Skidmore
2002; Beyleveld and Bos 2009).
GENERAL META-ETHICAL OBJECTIONS
125
mysterious than, for instance, Humean normativity.28 The transcendental conditions that
can lead to normative reasons are those claims that are instrumentally necessary to be an
agent. However, even if my argument succeeds, I will merely have shown that Kantian
constructivism is a coherent project. Whether or not Kantian constructivism ultimately
provides a plausible account of moral objectivity depends on its success in justifying
specific, substantive reasons for actions. I have already mentioned one example: being alive
seems to be a necessary means for engaging in any action whatsoever. That is, it is
impossible to pursue any end whatsoever if one is not alive. One therefore has a substantive
reason to want to be alive insofar as one engages in any action whatsoever. This is, however,
but one example of a necessary means for agency. In chapters 8 and 9, I will discuss the
question of how to identify the necessary means of agency.
6. The Contingency Objection
The fourth and final general objection to Kantian constructivism that I discuss in this
chapter is that not every agent actually accepts Y, e.g. not every agent actually values his or
her humanity. This objection has been put forward by Street as a second objection,
independent of her no-normativity objection. A similar objection has been formulated by
Quassim Cassam (1996) in the context of transcendental arguments in general, or at least in
the context of transcendental arguments in theoretical philosophy, and his formulation of
the contingency objection, what he calls ‘the problem of misconception’, is the main focus
of this section. I argue that, whatever the plausibility of the contingency objection in the
context of theoretical philosophy, this objection misfires as an objection to Kantian
constructivism because Kantian constructivists do not, and do not have to, claim that every
agent actually accepts Y.
Street’s variant of the contingency argument proceeds as follows. First, she remarks
that “one thing that makes the Kantian suggestion seem plausible is that as a contingent
28
Another way to make the same point is to understand transcendental commitments as limiting cases of internal
reasons (cf. Williams 1995, 220f3). An internal reason is generally understood as a reason which can be derived,
via a valid deliberative route, from the existing motivational set of a person, i.e. the existing beliefs, desires and
preferences of a person (Williams 1979). A transcendental argument is a limiting case of such an internal reason,
because although it does not presuppose the existence of any mind-independent reason (pace the external reason
theories), reasons are not derived from the substantive content of the person’s motivational set (as Humeans,
including Street, typically assume), but from the conditions of the possibility of having a motivational set at all,
independent of its contingent content.
CHAPTER 4
126
matter, most actual valuing creatures do as a matter of fact take themselves to be valuable”
(Street 2012, 53). Subsequently, Street claims that even though many agents value their own
humanity, not all agents value their humanity. Street’s point is thus that there might be an
agent that “does not take itself to be valuable” (Street 2012, 53). Therefore, Street argues, the
conclusions of Kantian constructivist arguments are not universally and categorically valid,
but merely express the contingent moral commitments of those that already have Kantian
moral sympathies in the first place.
In order to argue for this conclusion, Street comes up with a thought experiment which
involves a non-human agent, “a reflective social insect”, which lives “on another planet in
another galaxy” (Street 2012, 53). Street imagines this social insect telling the Kantian: “You
may have strong unreflective tendencies to value yourself non-instrumentally (most
animals do)—but I don’t: I honestly don’t care a hoot for my own survival” (Street 2012,
54). Street’s general objection to Kantian constructivism seems to be that for any Y, there
will always be agents who do not accept Y.
Street’s objection is very similar to Cassam’s (1996) objection to transcendental
arguments in theoretical philosophy. Recall that transcendental arguments in theoretical
philosophy aim to show that a justified belief (modest transcendental argument) or a
justified true belief (strong transcendental argument) in Y is a necessary condition of the
possibility of X, for instance that believing in the Kantian categories (Y) is a necessary
condition of the possibility of experience (X). Cassam’s objection – which he labels ‘the
problem of misconception’ – is that there will always be persons that seem to qualify as
having X but who do not believe in Y. For instance, there will always be persons who seem
to be perfectly capable of having experiences (X) but who do not actually believe in the
Kantian categories (Y). According to Cassam, this puts pressure on the idea that a belief in
Y is a necessary condition of the possibility of X.
Cassam discusses this general objection to transcendental arguments in theoretical
philosophy in relation to what he calls the Neo-Kantian ‘embodiment requirement on
personal self-consciousness’, which entails that “a necessary condition of personal selfconsciousness … is that one conceives of oneself qua subject as a ‘corporeal object among
corporeal objects’” (Cassam 1996, 276). The Neo-Kantian embodiment requirement thus
claims that embodiment (Y) is a necessary condition of the possibility of self-consciousness
(X). If this is true, this would show that, for instance, Cartesian conceptions of the self,
which understand the self in immaterial, incorporeal terms, are essentially flawed.
According to Cassam, the problem of misconception arises when a Cartesian comes
GENERAL META-ETHICAL OBJECTIONS
127
along who claims that she “genuinely believe[s] that her thinking self is incorporeal” (1996,
278). Cassam argues that this Cartesian poses a fundamental problem for the
transcendental argument because “If S is the proposition that the subject of thought and
experience is a corporeal object among corporeal objects, then someone who claims that
her thinking self is incorporeal is presumably going to want to deny that belief in the truth
of S is a necessary condition of the self-ascription of thoughts and experiences” (Cassam
1996, 279). In other words, if the transcendental argument claims that a belief in Y (e.g.
embodiment) is a necessary condition of the possibility of X (e.g. self-consciousness), and if
there is a person who seems to qualify as understanding him- or herself in terms of X (e.g.
being self-conscious) but who genuinely denies Y (e.g. embodiment), it is hard to maintain
that Y is truly a necessary condition of X.29
In the previous chapter, I have put forward a similar objection against David Enoch’s
transcendental argument for robust normative realism. Recall that Enoch claims that a
(justified true) belief in robust normative facts is indispensable for deliberation. He argues
that deliberation is (normatively) inescapable and that a belief in normative facts is a
necessary condition of the possibility of deliberation. Against Enoch, I argued that I seem to
be perfectly capable of engaging in deliberation, but I genuinely do not believe in the
existence of normative facts. And so Enoch must either accept that normative facts do not
seem to be a necessary condition of deliberation or somehow make clear that I do believe in
the existence of normative facts, contrary to appearances.
Street’s objection to Korsgaard’s transcendental argument for the value of humanity,
and Kantian constructivism in general, can be read as another instance of this general
objection to transcendental arguments in theoretical philosophy. Just like the Cartesian,
Street’s social insect denies the transcendental claim Y (i.e. that one must value one’s
humanity), while presumably counting as an X (i.e. an agent).
Although I agree with Cassam that the problem of misconception might pose a
challenge for transcendental arguments in theoretical philosophy and for arguments in
practical philosophy that aim to justify (true) beliefs, I do not think that it has the same
29
Of course, a proponent of a transcendental argument in theoretical philosophy might try to argue either that the
sceptic, e.g. the Cartesian, does not really qualify as an X (i.e. that the Cartesian is not really self-conscious) or that
the sceptic does not genuinely or really deny the belief in Y. Cassam (1996) argues that these kinds of moves fail. I
do not want to take a stance in this debate on the problem of misconception in relation to transcendental
arguments in theoretical philosophy. My main point is rather that the problem of misconception does not even
arise for transcendental Kantian constructivism because Kantian constructivism does not, or at least not first and
foremost, aim to justify doxastic commitments.
CHAPTER 4
128
impact on Kantian constructivism. Cassam himself mentions the possibility that some
transcendental arguments are not susceptible to his objection “as long as the conclusion of
[a transcendental argument] ... is understood as being essentially normative, it is not
undermined by the fact that some subjects do not believe what they ought to believe about
themselves” (Cassam 1996, 284). Cassam rejects this normative reading of the
transcendental commitment as a possible reconstruction of transcendental arguments in
theoretical philosophy, but he does not consider the possibility that transcendental
arguments in ethics could be reconstructed in this way. In what follows, I argue that the
most plausible reading of Kantian constructivism is along these normative lines and that
Kantian constructivism is therefore not susceptible to this objection.
In order to see why, it is important to look at the reason why Cassam thinks the
problem of misconception is so important in the context of theoretical philosophy in the
first place. According to Cassam, the problem of misconception applies to any ‘beliefdirected’ transcendental argument, i.e. any argument which aims to show that a belief in Y
is a necessary condition of the possibility of X.30 Subsequently, Cassam claims that beliefdirected transcendental arguments are susceptible to the following challenge:
The challenge which belief-directed transcendental arguments face, however, is
that if S is a proposition which the sceptic herself claims not to believe, then she is
hardly likely to concede that believing that S is true is a necessary condition of the
possibility of, say, experience or language (Cassam 1996, 279).
Cassam’s point is thus that if one tries to argue that a belief that Y is a necessary
precondition of X, an individual who does not believe in Y but who cannot be easily
dismissed as lacking in X poses a real problem.
Note, however, that this problem arises specifically in the context of transcendental
arguments which try to justify beliefs about the external world. This is the reason why the
objection applies both to arguments in theoretical philosophy, such as the argument for the
30
Cassam contrasts belief-directed arguments with ‘truth-directed’ arguments, i.e. arguments that aim to justify
true claims about the world. I do not find the distinction between truth directed and belief directed very helpful,
because truth-directed arguments are also belief directed. The difference between strong and modest
transcendental arguments is that strong transcendental arguments hope to justify true beliefs, whereas modest
transcendental arguments are ‘only’ interested in justified beliefs. Nothing important hinges on rejecting Cassam’s
distinction. Understanding both strong and weak transcendental arguments as belief-directed (or doxastic)
transcendental arguments only enlarges the scope of the problem of misconception.
GENERAL META-ETHICAL OBJECTIONS
129
embodiment requirement, and to arguments in ethics which try to justify a belief in the
existence of normative facts, such as Enoch’s indispensability argument for robust
normative realism.
As I have already argued in the previous chapter, however, Kantian constructivism
does not, first and foremost, aim to justify beliefs about the external world. Instead, Kantian
constructivism aims to justify claims about how one ought to act. I think that for this
reason the contingency objection, or the more general problem of misconception, simply
does not apply to Kantian constructivism. Street assumes that Kantian constructivism
would be undermined if not every agent accepts the conclusion of a transcendental
argument for certain normative conclusions (e.g. valuing one’s own agency). However, the
conclusion of transcendental Kantian constructivism is not that every agent actually values
Y or has a belief in the value of Y, but that every agent should value Y.
The problem is thus that Street misunderstands the direction of fit of Kantian
constructivism. Kantian constructivism has a world-to-mind direction of fit, where the
world should conform to the mind, and not a mind-to-world direction of fit, where the
mind should conform to the world. Consider Elizabeth Anscombe’s illustration of the
distinction between directions of fit by way of an example of a man who is walking around
a town with a shopping list while a detective is watching him. Anscombe writes that
if the list and the things that the man actually buys do not agree, and if this and
this alone constitutes a mistake, then the mistake is not in the list but in the man’s
performance (if his wife were to say: “Look, it says butter and you have bought
margarine”, he would hardly reply: “What a mistake! We must put that right” and
alter the word on the list to “margarine”); whereas if the detective’s record and
what the man actually buys do not agree, then the mistake is in the record
(Anscombe 2000, 56).
Anscombe’s point is that whereas the man’s shopping list has world-to-mind direction of
fit, the detective’s record of the actions of the man has a mind-to-world direction of fit.
Importantly, the example makes clear that different directions of fit have different ‘truthmakers’. If the man does not follow the shopping list, the man is at fault and not the
shopping list. The same could be said about Street’s thought experiment regarding the
reflective social insect who does not value his or her humanity (or his or her ‘insectness’).
Street assumes that Kantian constructivism is at fault if there are agents who do not value
CHAPTER 4
130
their humanity. But this would be similar to saying that the shopping list is at fault when
the man in Anscombe’s example fails to buy margarine.
In fact, the normativity of the transcendental claim presupposes at least the possibility
that not everyone accepts Y. The conclusions of Kantian constructivism state what
someone should do. This kind of normativity presupposes the possibility of someone failing
to act according to this principle: it makes no sense to tell a perfectly rational being that
they should do something, because there is no possibility that they would act otherwise.
Only finite rational beings can be wrong. Or as Korsgaard puts it, “there is no normativity if
you cannot be wrong” (Korsgaard 1996, 161). It thus only makes sense to say that every
agent should Y, if there might be agents who do not actually do Y at this point, but who
should.
Of course, this discussion leaves open the possibility that one might legitimately raise
the question of whether one should accept a specific Y insofar as one understands oneself as
an agent, e.g. whether one should indeed value one’s humanity insofar as one is an agent.31
The only point I want to make in this chapter, however, is that the fact that someone denies
Y (e.g. does not value his or her humanity) does not, taken by itself, constitute an objection
to Kantian constructivism. The objection that accepting a specific Y is not a necessary
condition of the possibility of X is a legitimate objection, and this objection can only be
discussed in the context of specific transcendental arguments. However, insofar as the
contingency objection is supposed to be a general objection to Kantian constructivism it
fails.
Understanding the transcendental claim in terms of a practical or normative
commitment also raises problems of its own, though. For instance, although the claim that
one ought, for instance, to value one’s humanity cannot be simply rejected by claiming that
one does not actually value one’s humanity, the ‘transcendental ought’ cannot be
understood completely independent of the psychology of individuals. Understanding how
exactly transcendental commitments relate to (empirical) psychological reality is crucial to
understanding transcendental arguments in ethics (or moral psychology in general), but
discussing this question unfortunately lies outside the scope of this thesis. What seems
uncontroversial, however, is that if it would in general be psychologically impossible for
human agents to Y, e.g. value their humanity, this would most likely constitute a
31
Regan (1999), for instance, raises such an objection in the context of Gewirth’s transcendental argument. Regan
argues that the substance of Gewirth’s Y is more minimal than Gewirth himself suggests. I will return to this in
chapter 9.
GENERAL META-ETHICAL OBJECTIONS
131
convincing objection to the argument because ‘ought implies can’.32 It seems unlikely,
however, that it is psychologically impossible for agents to accept just any Y. Instead, this
objection only potentially applies to specific, overambitious transcendental arguments.
Finally, ‘ought implies can’ is not a problem specific to Kantian constructivism but a
problem that applies to any account of normativity. ‘Ought implies can’ is therefore not a
reason to reject Kantian constructivism, but a reason to analyse whether or not a specific
transcendental argument asks something from an agent that is psychologically impossible.
7. Conclusion
In this chapter, I have responded to four objections to Kantian constructivism which all try
to show that Kantian constructivism fails to justify any categorical reason, independent of
its additional ambition to justify claims of substantive and interpersonal morality.
Importantly, these objections were directed to Kantian constructivism in general,
independent of the details of specific transcendental arguments, and they tried to show that
there is no single (relevantly) inescapable starting point for a transcendental argument in
ethics and that, even if there is an inescapable starting point, no categorical normative
claims can be justified on the basis of this starting point. These objections could thus be
understood as (variants of) one of the horns of the general dilemma for Kantian
constructivism, as applied to the ambition to justify categorical normative claims.
I have tried to show in this chapter that these objections fail as general objections to
Kantian constructivism, leaving open the possibility that there might be specific starting
points, e.g. specific conceptions of agency, that are escapable, or that there are specific
transcendental commitments which can be rejected without contradicting one’s own selfunderstanding. Although I have thus not tried to provide a comprehensive defence of
Kantian constructivism against these common meta-ethical objections, I have tried to show
that their plausibility should be assessed in the context of more specific transcendental
arguments.
The exception, however, is Enoch’s shmagency objection, i.e. the objection that a
transcendental argument has to presuppose a reason to be an agent in order to justify
32
Note, however, that this objection differs from Street’s objection, because Street does not claim it is
psychologically impossible to value one’s humanity. Instead, she claims that accepting the transcendental claim is
contingent.
CHAPTER 4
132
categorical normative claims. I have tried to argue that even if we assume, for the sake of
argument, that Enoch is right in suggesting that one needs a reason to be an agent, it is not
obvious that this would undermine Kantian constructivism. More fundamentally, however,
I think that Enoch has not provided us with a non-question begging reason to think that
Kantian constructivism needs to rely on such a reason to be an agent in the first place.
In the remaining chapters, I will discuss specific transcendental arguments that have
been put forward in ethics. I will do so by focusing first on the justification of categorical
and universal principle of interpersonal morality (chapters 5, 6 and 7) and subsequently on
the justification of substantive morality (chapters 8 and 9).
Part 3: Universality
Chapter 5
Transcendental Arguments from the
Second Person
1. Introduction
In the previous chapter, I have discussed general meta-ethical objections to Kantian
constructivism, which claim that Kantian constructivism cannot justify any categorical
normative claims or reasons. I have argued that these objections fail as general objections to
Kantian constructivism, but that the success of Kantian constructivism in answering these
objections ultimately hinges on the details of specific transcendental arguments.
In the remainder of this thesis, I will therefore turn to specific transcendental
arguments that have been put forward in ethics. In the next three chapters, I will discuss the
success of specific transcendental arguments in ethics in the context of justifying not just
categorical claims but also universal claims, that is, moral claims that govern interpersonal
conduct and which entail moral respect for all agents and not just for a subset of agents.
The reason for this is that Kantian constructivism typically aims to show that (a subset of)
the necessary preconditions of an inescapable aspect of our practical self-understanding
entail claims about how to treat others.
In this context, the distinction between what I have called transcendental arguments
from the first person and transcendental arguments from the second person is of particular
relevance. Transcendental arguments from the first person hold that it is possible to argue
from the claim that A must accept whatever are the necessary preconditions of A’s agency
(e.g. value A’s humanity) to the claim that A is required to assign the same status to the
necessary preconditions of the agency of other individuals (value the humanity of other
CHAPTER 5
136
agents). Transcendental arguments from the second person, on the other hand, claim that it
is impossible to argue from the necessary preconditions of individual agency to
interpersonal morality. Instead, transcendental arguments from the second person propose
to take as their starting point A’s involvement in some form of interaction, such as
communication, argumentation or shared action, and subsequently explore the necessary
conditions of the possibility of interaction. By focusing on the necessary conditions of the
possibility of interaction versus the necessary conditions of the possibility of action, the
idea is that there no need to cross the bridge between the personal and the interpersonal in
the first place, because the necessary conditions of the possibility of interactions are
interpersonal from the very start. Or at least this is what transcendental arguments from the
second person aim to show.
The most recent attempts at putting forward a transcendental argument for
interpersonal morality are arguments from the second person (Apel 1980; Habermas 1990a;
O’Neill 1986; Darwall 2006; and to some extent Korsgaard 1996). The main goal of this
chapter is to critically assess several influential transcendental arguments from the second
person for categorical and universal claims of interpersonal morality. I argue that despite
the existence of important differences between these arguments, they are all susceptible to
(a variant of) one or both horns of the general dilemma for Kantian constructivism: either
the specific kind of interaction that is taken as a starting point of these arguments is
inescapable, but in that case it is unclear why this kind of interaction has any universal
principle of interpersonal morality among its necessary conditions of possibility, or
interaction does have a principle of interpersonal morality among its necessary conditions
of possibility – but in that case the practice seems to be optional and the moral principles
are therefore not categorically valid.
Although I cannot, within the scope of this chapter, provide a conclusive argument
against transcendental arguments from the second person, I try to show that several
influential authors have failed to provide a convincing response to this dilemma and that
the failure is systematic. In the chapters that follow, I will discuss whether and how
transcendental arguments from the first person could lead to the justification of
interpersonal morality where second-personal accounts fail.
The structure of this chapter is as follows. In section 2, I briefly discuss the difference
between transcendental arguments from the first person and transcendental arguments
from the second person. In section 3, I discuss common objections to first person singular
arguments in general in order to indicate the reasons why most theorists have recently
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
137
turned to transcendental arguments from the second person. Subsequently, I critically
discuss different arguments from the second person and suggest that they are susceptible to
the general dilemma for Kantian constructivism: Karl-Otto Apel’s and Jürgen Habermas’s
argument
from
argumentation
(section
4.1);
Onora
O’Neil’s
argument
from
communication (section 4.2); Christine Korsgaard’s argument from the publicity of reasons
(section 4.3); and Stephen Darwall’s argument from the second-person standpoint (section
4.4). Finally, in section 5, I abstract away from these specific arguments and briefly discuss
general reasons why I think that the prospects of a transcendental argument from the
second person for universal and categorical morality look grim.
2. Transcendental Arguments from the Second Person
In the previous chapters, I have relied on an interpretation of Kantian constructivism which
starts from the self-understanding of the individual agent, i.e. the self-understanding of a
being who acts for reasons and who aims to achieve certain purposes. Immanuel Kant’s
justification of the categorical imperative in the third section of The Groundwork is often
considered to provide the prototypical example of such a transcendental argument from the
first person.1 A more recent example is Alan Gewirth’s argument for the ‘principle of
generic consistency’ in Reason and Morality. In addition, the first phase of Christine
Korsgaard’s argument for the categorical imperative (the argument for the claim that you
must value your own humanity) could be understood as starting from the first-person
perspective of the agent (but, as I will show below, she thinks, at least on one possible
reading of her work, that we need an additional (transcendental) argument in order to
arrive at intersubjective morality).2
However, most of the recent transcendental arguments that have been proposed in
ethics start not from A’s self-understanding as an agent, but from A’s involvement in some
form of interaction, such as communication, argumentation or shared action. Very
roughly, transcendental arguments from the second person claim that A is inescapably
involved in some form of interaction and that, insofar as interaction has certain necessary
conditions of possibility, A has to accept whatever are the necessary conditions of the
1
But see e.g. O’Neill (1989b), who puts forward a reading of Kant which stresses the intersubjective nature of this
ethics, in terms of the public use of reason.
2
I will discuss a fully first person interpretation of Korsgaard’s argument in chapter 6.
CHAPTER 5
138
possibility of this kind of interaction. Crucially, the claim is that the necessary conditions of
the possibility of interaction entail the acceptance of a categorical and universal principle of
interpersonal morality.
I have referred to these arguments as transcendental arguments from the second
person to highlight, negatively, that they do not start from the self-understanding of the
individual agent and, positively, that these arguments somehow try to justify morality in
terms of the necessary preconditions of the relation between individuals. The relevance of
the distinction between arguments from the first person and arguments from the second
person in the context of justifying interpersonal morality is as follow: a transcendental
argument from the first person typically tries to justify a categorical reason as a necessary
condition of the possibility of (individual) agency first, and then introduces an additional
argument that is supposed to lead to conclusions about interpersonal morality, including
claims about moral rights and moral duties to others. A transcendental argument from the
second person, on the other hand, begins from A’s involvement in some form of interaction
and, if successful, its conclusions are therefore interpersonal from the very start.
As I will argue below, however, there are important differences between different
transcendental arguments from the second person which should be kept in mind.
Nevertheless, I think that, despite these differences, it might still be useful to group these
different arguments under the label second person arguments, because they at least share
the idea that principles of interpersonal morality cannot be justified in terms of the
necessary conditions of (individual) agency.3
Let me briefly mention some of the arguments that I have in mind. Karl-Otto Apel and
Jürgen Habermas defend what they call a ‘discourse ethical’ theory of morality, according to
which a categorical and universal principle of interpersonal morality can be justified in
terms of the necessary preconditions of argumentation or ‘discourse’. Discourse ethics
suggest that we should move away from what they call the “philosophy of consciousness”,
with its emphasis on the solitary subject, to the “philosophy of language”, which emphasizes
the social embeddedness of individuals (see e.g. Habermas 1990b, 9, 96). For discourse
ethics, the central concept, which forms the starting point of its transcendental argument, is
communication or argumentation and not agency.
A very similar argument is put forward by Onora O’Neill. O’Neill focuses on the
3
I put ‘individual’ in brackets because some of the transcendental arguments might be (re)formulated in terms of
the necessary conditions of, for instance, communicative agency.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
139
“necessary conditions for possible communication or interaction with others, without
restriction” (O’Neill 2013, 223). Although O’Neill, at least insofar as I am aware, never
explicitly states that she uses a transcendental argument, her attempt to show that
reasoning and communication are only possible insofar as one accepts the categorical
imperative could, I think, be understood as a transcendental argument from the second
person (I will come back to this in section 4.2). For O’Neill, the central concept is thus
communication.
Stephen Darwall claims that moral concepts “are all irreducibly second-personal” and
that interpersonal morality can only be vindicated “from the second- person perspective”
(Darwall 2006, x). For Darwall, the central concept is ‘the second-person standpoint’, the
perspective that persons take up when they make and acknowledge claims on one another’s
conduct and will (Darwall 2006, 3). According to Darwall, respecting the equal dignity of all
persons is among the necessary conditions of the possibility of the second-person
perspective.
Finally, as I have already briefly illustrated in the previous chapter, although Korsgaard
uses a transcendental argument from the first person to argue for the claim that one must
value one’s humanity insofar as this argument starts from the individual agent, she, at least
on one possible reading of her work, sides with proponents of transcendental arguments
from the second person when it comes to the justification of interpersonal morality, i.e. the
claim that one must value the humanity of others. Korsgaard criticizes the idea that
interpersonal morality can be justified by reference to the “individual agent” with his or her
“private reasons” (Korsgaard 1996, 133). Instead, Korsgaard claims that morality can only
be justified by reference to “our social nature” and essentially “public reasons” (Korsgaard
1996, 135). For Korsgaard, the central concept for a transcendental argument for
interpersonal morality is thus our social nature and the publicity of reasons, i.e. the idea
that reasons always have normative force, not just for yourself but also for others.4
On this reading of her argument, Korsgaard’s argument might be best characterized as
a ‘hybrid’ transcendental argument. Her argument consists of two distinct phases. In the
first phase of her argument, she employs a transcendental argument from the first person in
order to argue for the conclusion that agents necessarily have to value their own humanity.
In the second phase of her argument, she introduces the second-personal idea that all
4
There is disagreement concerning the question of whether this second phase, in which Korsgaard relies on
Wittgenstein’s private language argument, constitutes a transcendental argument (Stern 2011, 77). For a reading of
the second phase as a transcendental argument see Skidmore (2002, 135).
CHAPTER 5
140
reasons are essentially social reasons and argues that only the social nature of reasons can
generate interpersonal morality. In a review paper on Darwall’s second-person standpoint,
Korsgaard emphasizes the hybrid nature of her position when she claims that “Darwall
characterizes me both as someone who thinks all reasons are second-personal and also as
someone who thinks that ‘moral obligations can be grounded in the constraints of firstpersonal deliberation alone’. That may sound paradoxical but it is basically right”
(Korsgaard 2007, 10).
As should have become clear from this brief sketch, there is something artificial about
trying to group all these authors under the same label, given the substantial differences
between these arguments. Saying that argumentation has certain necessary conditions of
possibility is not the same as saying that interpersonal morality can only be vindicated from
an irreducibly second-person standpoint or that reasons are intrinsically social. In addition,
there are, for instance, important differences between what I have called Korsgaard’s hybrid
transcendental argument for interpersonal morality and, for instance, Darwall’s attempt to
justify all normative reasons in terms of the necessary preconditions of the second-person
standpoint.
What these authors share, however, is the idea that interpersonal morality should
somehow be grounded, first and foremost, in relations between individuals, and not, first
and foremost, in the necessary conditions of the possibility of agency.5 In addition, as will
become clear below, another reason to group these authors together under one label is that
nearly all of these authors crucially rely on Wittgensteinian insights concerning the
publicity of language.6 In short, I use the label ‘transcendental arguments from the second
person’ rather loosely to refer to those transcendental arguments in ethics which do not
start from the self-understanding of the individual agent, but from a certain conception of
the relation between persons. In what follows, I will aim to show that despite the differences
between these authors, they all fail to justify a categorical and universal principle of
interpersonal morality, because they are susceptible to one or both horns of the general
dilemma for Kantian constructivism.
5
Even though they disagree about whether or not this only applies to interpersonal morality (e.g. Korsgaard) or to
all normative claims (e.g. Darwall).
6
One might argue that Darwall is an exception to this rule. However, Darwall relies on Strawson’s notion of
reactive attitudes, which could be interpreted as a Wittgensteinian theory. I will not, however, pursue this
suggestion here.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
141
3. Two Objections to Transcendental Arguments from the First Person
Transcendental arguments from the second person are crucially motivated by the belief that
transcendental arguments from the first person cannot, in principle, lead to the justification
of interpersonal morality. Before discussing specific transcendental arguments from the
second person, I will therefore first briefly discuss two prominent reasons for rejecting
transcendental arguments from the first person. The first objection is that these arguments
are logically flawed: interpersonal moral claims (such as claims about rights and duties)
cannot logically follow from the self-understanding of an individual agent. The second
objection is that even if a transcendental argument from the first person is not logically
flawed, it provides the wrong kind of reason for the acceptance of interpersonal morality.
In this section, I do not assess the plausibility of these objections but merely use them
to indicate the motivation behind the recent turn to transcendental arguments from the
second person for the justification of interpersonal morality. The discussion of the
objections to transcendental arguments from the first person is followed by a critical
analysis of different transcendental arguments from the second person. In the chapters that
follow, I will evaluate to what extent these objections against transcendental arguments
from the first person are convincing.
3.1 The Logical Objection
The most widely shared objection to transcendental arguments from the first person for
interpersonal morality is that they are logically flawed, meaning that an argument which
starts from the self-understanding of an individual agent simply cannot lead to the
justification of moral rights and duties.
This objection is, for instance, put forward by Korsgaard in The Sources (Korsgaard
1996, 134). As I have already noted, Korsgaard thinks that a first person singular
transcendental argument can succeed in showing that I necessarily have to value my own
humanity. That is, she grants that an argument from the first person can lead to certain
categorical normative reasons. She denies, however, that this argument by itself imposes
any obligations on others or towards others.
Korsgaard puts forward this objection by making a distinction between what she calls
‘private reasons’ and ‘public reasons’ (Korsgaard 1996, 133). According to Korsgaard, an
agent has a private reason if a reason has normative force for him or her. An agent has a
CHAPTER 5
142
public reason if the reason also has normative force for others, e.g. when your reason to
value your own humanity is also a reason for others to value your humanity. Korsgaard
argues that the problem with arguments from the first person is that they try to argue from
private reasons to public reasons: the argument from the first person tries to show that I
need to take your reasons into account on the basis of the fact that I need to take my own
reasons into account. This cannot work, according to Korsgaard, because “private reasons
will remain private forever” (Korsgaard 1996, 134).7 Subsequently, she claims,
Consistency can force me to grant that your humanity is normative for you just as
mine is normative for me. It can force me to acknowledge that your desires have
the status of reasons for you, in exactly the same way that mine do for me. But it
does not force me to share in your reasons, or make your humanity normative for
me. It could still be true that I have my reasons and you have yours, and indeed
that they leave us eternally at odds (Korsgaard 1996, 134).
Korsgaard thus claims that although the argument could show that every agent necessarily
has to value his or her own humanity, it does not (and cannot) follow that agents have to
value each other’s humanity or whatever are the necessary preconditions of agency. In
short, Korsgaard claims that a first-personal transcendental argument cannot show that
(moral) egoism is inconsistent.
In putting forward this objection, Korsgaard draws on Bernard Williams’ objection to
transcendental arguments from the first person in Ethics and the Limits of Philosophy.8
There, Williams claims,
The reasons that B has for doing something are not in themselves reasons for
another’s doing anything. The should of practical reason has, like any other, a
second and a third person, but these forms merely represent my perspective on
your or his interests and rational calculations, the perspective of “if I were you.”
Considering in those terms what B should do, I may well conclude that he should
interfere with my freedom (Williams 1986, 61).
7
See also Habermas (1994, 60) and Alexy (2006, 177–78). This argument seems to be widely accepted, even by
critics of Korsgaard. See e.g. Gert (2002, 307).
8
For a defence of Gewirth’s first-personal argument against Williams’ criticism, see Beyleveld (2013).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
143
Williams thus likewise denies that it can follow from a transcendental argument from the
first person that one has reason to take into consideration the necessary preconditions of
someone else’s agency.
Similar objections are put forward by others. Habermas, for instance, criticizes
transcendental arguments from the first person because, according to Habermas, they can
never lead to conclusions about rights (and duties). Habermas, for instance, criticizes
Gewirth’s argument from the first person along these lines because, according to
Habermas, “[Gewirth’s] teleological concept of action is inadequate to provide a
transcendental­pragmatic justification of the notion of a right to such ‘necessary goods’, as
opposed to the idea of the goods themselves” (1990b, 101).
In support of this claim, Habermas approvingly quotes Alasdair MacIntyre, who writes:
Gewirth argues that anyone who holds that the prerequisites for his exercise of
rational agency are necessary goods is logically committed to holding also that he
has a right to these goods. But quite clearly the introduction of the concept of a
right needs justification both because it is at this point a concept quite new to
Gewirth’s argument and because of the special character of the concept of a right.
It is first of all clear that the claim that I have a right to do or have something is a
quite different type of claim from the claim that I need or want or will be
benefiting by something. From the first – if it is the only relevant consideration – it
follows that others ought not to interfere with my attempts to do or have whatever
it is, whether it is for my own good or nor. From the second it does not. And it
makes no difference what kind of good or benefit is at issue (MacIntyre 1985,
2nd:80).
Although Habermas and MacIntyre frame the objections in terms of rights and duties and
not in terms of ‘public reason’, their objection is essentially the same as the objection put
forward by Korsgaard and Williams. They all deny that one can justify interpersonal
demands from the first-person perspective. The objection that transcendental arguments
from the first person are logically flawed should be clear by now: the objection is that an
argument that starts from the necessary preconditions of my own agency cannot lead to
justification of any obligations on others or towards others.
Note again that at this point I do not want to assess the plausibility of this objection.
Rather, the point is merely to illustrate the motivation for people to turn to transcendental
CHAPTER 5
144
arguments from the second person. I will critically evaluate this objection in the later
chapters.
3.2 The Wrong Kind of Reason Objection
The second objection to transcendental arguments from the first person is that even if an
argument from the first person could logically lead to conclusions about interpersonal
morality, it would only show that we are committed to moral principles for the wrong kinds
of reasons. Although this objection is not as prominent as the objection that the argument is
logically flawed, it nevertheless seems to be an intuitive objection that is shared by many
people, even though it is not often discussed in writing. Darwall (and to a lesser extent
Korsgaard) is one of the few people who does explicitly articulate this objection, and in this
subsection I briefly describe Darwall’s wrong kind of reason objection.
In general, ‘the wrong kind of reason objection’ refers to the idea that not all reasons
qualify as the right reason for holding a certain belief. Darwall mentions two examples to
illustrate this general problem (Darwall 2006, 16). The first example is that there might be
pragmatic reasons to believe some proposition, but the fact that there are pragmatic reasons
to believe a proposition does not make the proposition credible. Pragmatic reasons are the
wrong kinds of reasons for believing a proposition to be true. Darwall’s second example is
that there might be moral reasons why it is morally objectionable to be amused by a certain
joke, but that does not mean that the joke is not funny. Subsequently, Darwall puts forward
the following general definition of what constitutes a right kind of reason:
To be a reason of the right kind, a consideration must justify the relevant attitude
in its own terms. It must be a fact about or feature of some object, appropriate
consideration of which could provide someone’s reason for a warranted attitude of
that kind toward it (Darwall 2006, 16).
Darwall claims that this problem can also be applied to the justification of morality. For
instance, if one accepts that morality consists of universal and categorical beliefs, a
justification that refers to the desires of individuals would offer the wrong kind of reason
for accepting moral beliefs (even if our desires would converge on a set of moral beliefs)
(Darwall 2006, 17). Similarly, an argument that is based on self-interest cannot provide a
reason for accepting moral beliefs, because, arguably, morality by definition does not refer
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
145
to conduct that is motivated by self-interest (Korsgaard 1996, 134).9 Darwall suggests,
although he does not explicitly argue for this conclusion, that arguments from the first
person similarly offer the wrong kinds of reasons for (interpersonal) morality.
Darwall’s main reason for the claim that arguments from the first person cannot justify
moral judgements seems to be that ethical concepts in general, and moral concepts in
particular, have an irreducible second-personal nature: “ethical notions … —the culpable,
moral responsibility, and, I argue, moral obligation—all have an irreducibly secondpersonal aspect that ties them conceptually to second-personal reasons” (Darwall 2006,
17).10 If Darwall is right that there is a conceptual relation between moral concepts and
second-personal reasons, a transcendental argument from the first person cannot in
principle provide the right kind of reason for accepting moral beliefs because it would not
‘justify the relevant attitude in its own terms’. Again, the point here is not to evaluate this
objection (I will come back to this objection in the chapters that follow), but merely to
illustrate the reasons why many people reject transcendental arguments from the first
person
There is an important difference between the objection that an argument from the first
person is logically flawed and the wrong kind of reason objection. As discussed above, the
objection that the argument is logically flawed grants that a transcendental argument from
the first person succeeds in showing that I necessarily have to value my own humanity and
only fails when it tries to justify obligations between people. Darwall, on the other hand,
thinks that a transcendental argument from the first person cannot even justify obligations
towards oneself: “the problem, according to me, is not how to get from accountability to
oneself to accountability to others. I don’t think it is possible to have either without the
other. As I see it, the two are a package deal” (Darwall 2007, 56; see also Darwall 2006,
32f12). Darwall’s criticism of transcendental arguments from the first person is therefore
more fundamental, because he denies that any normative concept (and not just normative
obligations between people) can be understood and justified independent of the secondperson standpoint. If Darwall is right, only an argument from the second person can
provide the right kind of reason for morality, including interpersonal morality.
9
Korsgaard seems to suggest that this is the problem with neo-Kantian theories like Gewirth’s.
See also his claim that “[h]olding someone responsible always addresses a demand, and the only way to justify
any such demand is with second-personal reasons, that is, from within the standpoint we must occupy to address
it. Otherwise, we will be invoking ‘reasons of the wrong kind’” (Darwall 2006, 246).
10
CHAPTER 5
146
4. Evaluating Transcendental Arguments from the Second Person
Taken together, the logical objection to transcendental arguments from the first person and
the wrong kind of reasons objection have led recent proponents of transcendental
arguments to look for second-personal starting points to justify interpersonal morality. In
this section, I critically evaluate four different transcendental arguments from the second
person. The goal of this section is not to provide a comprehensive analysis of these
arguments. Instead, the goal is to point to a problem which all of these arguments seem to
have in common and critically assess recent attempts to respond to this problem. The
problem is that these arguments fall prey to one or both horns of a variant of the general
dilemma for Kantian constructivism: either participation in interaction is inescapable, but
in that case it is unclear why this kind of interaction has any universal principle of
interpersonal morality among its necessary conditions of possibility, or interaction has a
(universal) principle of interpersonal morality among its necessary conditions of possibility
– but in that case this kind of interaction seems to be optional. In both cases, the ambition
to justify a categorical and universal account of interpersonal morality fails.
In section 4.1, I will first briefly introduce the dilemma in the context of discourse
ethics, a moral theory developed by Karl-Otto Apel and Jürgen Habermas. Subsequently, I
will discuss in more detail recent attempts to respond to this kind of problem. In section
4.2, I will critically discuss O’Neill’s argument from communication. In section 4.3, I will
discuss Korsgaard’s argument from the publicity of reasons. Finally, in section 4.4, I will
discuss Darwall’s attempt to ground interpersonal morality in the second-person
standpoint.
4.1 The Argument from Argumentation
Discourse ethics is one of the most influential attempts at justifying interpersonal morality
on the basis of an argument from the second person. Karl-Otto Apel recently described the
discourse ethical project of both himself and Habermas as follows:
Common to both of us is a certain—positive and critical-transformative—
connection with Kantian philosophy, e.g., a transformation of Kant’s philosophy
of the “transcendental subject” or “consciousness” in terms of a philosophy of
language and intersubjectivity, and in this respect we take up —both of us— the
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
147
“pragmatic turn” of language-analytic philosophy (2005, 13; seen also Apel 1980,
267).
Discourse ethics, in other words, rejects the focus on the individual agent and replaces it
with a focus on language and intersubjectivity. Subsequently, the goal of discourse ethics is
to show that among the necessary conditions of the possibility of language and
intersubjectivity there is a categorical and universal principle of (intersubjective) morality.11
In my discussion of discourse ethics, I focus mainly on Apel’s (1980) defence of
discourse ethics in ‘the a priori of the communication community and the foundations of
ethics’.12 The reasons for this are that this is one of the most influential statements about the
project of discourse ethics and that other discourse ethicists, such as Jürgen Habermas, by
and large follow Apel’s description and defence of discourse ethics.13 I therefore take Apel’s
defence of discourse ethics to be an, if not the, authoritative statement about this project.
11
Both Apel and Habermas position their moral theories in a broader approach to epistemology and ontology that
is inspired by the linguistic turn and by pragmatism. In this section, I focus exclusively on their moral theories.
12
There is a huge amount of literature on discourse ethics, mainly in the German tradition of philosophy. In this
subsection, I cannot possibly do justice to this literature. The aim of this subsection is therefore not to provide a
comprehensive analysis of discourse ethics. Instead, the aim is to describe the main features of discourse ethics and
subsequently formulate what I take to be the fundamental challenge for this kind of theory in the form of the
dilemma for Kantian constructivism. In the subsections that follow, I will discuss in more detail how recent
attempts to formulate similar transcendental arguments from the second person in Anglo-Saxon philosophy face a
similar challenge, and I claim that they have failed to formulate a convincing response to this objection.
13
There is one fundamental difference between Apel’s and Habermas’s version of discourse ethics. This difference
has to do with the status of the transcendental conditions, the necessary presuppositions of argumentation.
According to Habermas, all philosophical propositions, including the transcendental conditions, are fallible and
empirically testable (Habermas 1994, 82–84; see also Habermas 1990b, 95). This claim could be interpreted in two
ways: sometimes it seems as if Habermas claims that the transcendental conditions themselves are fallible and
therefore contingent, but his claim could also be understood as applying not to the transcendental conditions
themselves but to a philosophical reconstruction of the transcendental conditions. I agree with the latter claim
insofar as every transcendental argument that is put forward could in principle be wrong (see also chapter 3), but I
would disagree if the claim about fallibility is meant to apply to the transcendental conditions themselves. The
reason for this is that there is no position from which Habermas could claim that the conditions of possibility of
argumentation are fallible. See also Apel’s response to Habermas’s fallibility claim (interpreted as a claim about the
transcendental conditions): “the unavoidable presuppositions of argumentation (which cannot be denied without
committing a performative self-contradiction) cannot be fallible and subject to empirical tests, because in case of
falsification they would simultaneously be presupposed in their transcendental function. For the same reason it
makes no sense, I suggest, to suppose that the presuppositions of argumentation could change one good day, for
the question would be: from where—i.e., under which presuppositions—could we think of these presuppositions
as being contingent? Habermas may try to think this from ahistorico-sociological perspective, and in doing so he
may understand himself as a modest and self-critical philosopher. But I would assert that he simply forgets to
reflect on his own necessary presuppositions of argumentation and thereby falls back to “transcendent”
metaphysics, for he takes, as it were, a divine point of view outside the world, from which (he tries) to conceive of
everything, including transcendental conditions of thought, as being just contingent, that is, historical facts” (2005,
14).
CHAPTER 5
148
Apel’s goal is to show that “rational argumentation ... in itself presupposes the validity
of universal, ethical norms” (Apel 1980, 257).14 The argument has roughly the following
basic structure:
1) Rational argumentation (X) is inescapable for A.
2) (The acceptance of) a universal, ethical norm (Y) is a necessary condition of the
possibility of rational argumentation (X).
3) Therefore, A must accept this universal ethical norm (Y).
In what follows, I mainly want to focus on the transcendental claim, i.e. the claim that the
acceptance of a universal ethical norm is a necessary condition of the possibility of rational
argumentation. Before discussing this claim in more detail, it is important to first remove a
possible source of misunderstanding. One might be tempted to think that argumentation is
clearly an optional human practice and that therefore the conclusions of the argument
cannot be anything more than hypothetical imperatives. Apel, however, makes it clear that
he thinks that argumentation is not just a contingent human practice, but in fact
presupposed in “reason” (1980, 262), “cognition” (1980, 265), “acting” (1980, 269) and
“self-understanding” (1980, 271). The reason for this has to do with Apel’s Wittgensteinian
background, according to which any meaningful activity or cognition can only be
understood in the context of a language game. According to Apel, “anyone who speaks or
who simply acts meaningfully is already participating in a potential discussion” (Apel 1980,
275). In other writings, Apel therefore concludes that “man can withdraw from this
institution [i.e. argumentation] only at the price of the loss of the possibility of selfidentification as a meaningfully acting being, e.g. in suicide from existential despair or in
the pathological process of paranoid-autistic loss of self” (Apel 1975, 268). I come back to
the Wittgensteinian elements of Apel’s argument below, but for now let us assume that
argumentation is indeed inescapable.
So assuming that argumentation is inescapable, does it presuppose a universal moral
norm? Apel’s argument for this claim (premise 2 in the argument above) proceeds roughly
in three steps. First, Apel claims that the meaning and validity of arguments can only be
settled in what he calls ‘a community of communication’, or a language community:
14
Apel uses the term ‘basic moral norm’ as a synonym for ‘moral principle’ (see e.g. Apel 1980, 278).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
149
[T]he logical validity of arguments cannot be tested without, in principle, positing
a community of scholars who are capable of both intersubjective communication
and reaching a consensus. Even the de facto solitary scholar can only explicate and
test his line of argument in so far as he is able to internalize the dialogue of a
potential community of argumentation in the critical ‘discourse of the soul with
itself’ (Plato). This proves that the validity of solitary thought is basically
dependent upon the justification of verbal arguments in the actual community of
argumentation (Apel 1980, 258).
Apel thus claims that even a solitary scholar has to presuppose a community of
argumentation. Apel takes this conclusion to follow from Wittgenstein’s argument that a
private language is impossible, because it is impossible to determine the correct application
of a rule all by oneself: “‘one person alone’ cannot follow a rule and create validity for his
thought within the framework of a ‘private language’. This is in principle, public” (Apel
1980, 258).
Second, Apel claims that this ‘actual community of communication’ is only possible if
one also accepts a basic moral norm: “in the community of argumentation it is presupposed
that all the members mutually recognize each other as participants with equal rights in the
discussion” (Apel 1980, 259).15 Apel thus claims that a language community is only possible
if one accepts a basic moral norm, which entails that one recognizes the other members of
the language community as equal participants in the discussion. Recognizing the members
of the language community as equal participants in the discussion is, however, in principle
still compatible with a particularist morality in which one only recognizes the equal rights
of persons with whom one is actually communicating: a restricted community of
communication.
Third, Apel therefore claims that the community of communication should include all
persons, i.e. that one has to presuppose a universal language community and that therefore
the basic moral norm is universally valid:
Since all linguistic utterances and, moreover, all meaningful human actions and
physical expressions (in so far as they can be verbalized) involve “claims” (in
15
He adds that all human needs are ethically relevant: “all human needs – as potential claims – i.e. which can be
reconciled with the needs of all the others by argumentation, must be made the concern of the communication
community” (Apel 1980, 277–78).
CHAPTER 5
150
German “Ansprüche”) and hence can be regarded as potential arguments, the basis
norm of mutual recognition by the participants in the discussion potentially
implies that of the “recognition” of all human beings as “persons” in Hegel’s sense.
In other words, all beings who are capable of linguistic communication must be
recognized as persons since in all their actions and utterances they are potential
participants in a discussion, and the unlimited justification of thought cannot
dispense with any participant, nor with any of his potential contributions to a
discussion (Apel 1980, 259).
Apel’s argument thus proceeds as follows:
1) The meaning and validity of arguments can only be settled in a language
community.
2) This language community can only exist if one also accepts a basic moral norm,
which implies equal moral respect for the members of the language community.
3) The relevant language community that engages in argumentation includes all
persons, i.e. argumentation presupposes a universal language community.
4) Therefore, one has to respect all persons (insofar as they are all members of the
universal language community).
If Apel’s argument succeeds, he will have shown that any person, simply by virtue of being
engaged in argumentation, is necessarily committed to a universal moral norm. Given that
argumentation is inescapable, these universal moral norms would be categorically binding.
So how convincing is Apel’s argument for this conclusion? I think that the main
problem with Apel’s argument is that it is unclear why one is necessarily committed to a
universal language community (step 3 above) and therefore to a universal moral norm (step
4 above), insofar as one engages in argumentation (for a similar criticism of Apel see Illies
2003, 85–86). 16 Note that Apel’s argument crucially relies on Wittgenstein’s private
16
Illies claims, “It [Apel’s argument] can show that everyone is committed qua rational being to some normative,
non-empirical judgments about his reasoning. But these judgements are not yet fully universal ... Apel cannot
demonstrate that they include everyone in the required sense” (Illies 2003, 74).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
151
language argument. 17 Apel presents his project as a “transcendental-philosophical
radicalization of the later Wittgenstein’s work” (Apel 1980, 269). However, even if
Wittgenstein’s private language argument is correct,18 it shows only that meaning depends
on a particular language community (step 1 above), and not that meaning depends on a
universal language community. This means that even if Apel is correct that a language
community can only exist if one accepts a basic moral norm (step 2 above), he fails to
justify a universal moral norm. So even if one agrees with Apel that “arguments ... must
always be understood simultaneously as meaning and validity claims that can only be
explicated and decided upon in an interpersonal dialogue” (Apel 1980, 260), one can still
deny that the interpersonal dialogue should include all persons.
Although the claim that the existence of a language community is a necessary condition
of the possibility of language and argumentation seems to be plausible enough, the claim
that argumentation presupposes a universal language community and that one therefore
has to ascribe equal rights to all others, even to those with whom one is not actually in
communication at this moment, as a necessary condition of the existence of such a
language community seems to be less plausible. There is thus a gap between the claim that a
language community is a necessary condition of the possibility of argumentation, and the
claim that only (moral respect for) a universal language community can fulfil this
function.19
Apel thus seems to present the reader with a false dilemma. Apel suggests that either
one has to accept the universal language community or one is committed to
‘methodological solipsism’, which ultimately results in the “loss of the possibility of selfidentification as a meaningfully acting being” and a “paranoid-autistic loss of self” (Apel
17
Although I cannot go into the details of the private language argument here, the private language argument
roughly claims that an individual cannot by themself determine the meaning of a concept, because there has to be
a way to distinguish the correct from the incorrect application of the concept. This conclusion follows from
Wittgenstein’s view that language is a rule-governed activity and that rules cannot determine their own
application.
18
Illies notes that Apel does not argue anywhere for the validity of the private language argument (Illies 2003, 85).
In what follows I will simply assume the validity of the private language argument, but argue that even if the
private language argument is valid, it cannot ground a universal moral norm.
19
Apel also makes a distinction between a real and an ideal communicative community: “anyone who engages in
argument automatically presupposes two things; first, a real communication community whose member he has
himself become through a process of socialization and an ideal communication community that would basically be
capable of adequately understanding the meaning of his arguments and judging their truth in a definitive manner”
(Apel 1980, 280). But again an argument is missing for why one also necessarily has to presuppose an ideal
communication in addition to the real communication community which is necessary to engage in any
meaningful action.
CHAPTER 5
152
1975, 268). But this overlooks the possibility of acknowledging that meaning and
argumentation are intersubjective while denying that the relevant language game is
necessarily a universal language game.
In the passage I quoted above, Apel suggests that “all beings who are capable of
linguistic communication must be recognized as persons” (Apel 1980, 259 my emphasis).
Apel thus suggests that one should not respect all persons because this respect is among the
necessary conditions of the possibility of argumentation, but because they have certain
capacities. This presupposes that all persons with certain capacities should be respected, but
it is unclear why this principle should be accepted or why this principle is a necessary
precondition of argumentation. Apel claims that these persons should be respected because
“they are potential participants in a discussion, and the unlimited justification of thought
cannot dispense with any participant, nor with any of his potential contributions to a
discussion (Apel 1980, 259).” But again, one might wonder why the fact that someone is a
potential participant in a discussion justifies the idea that respecting this person is a
necessary precondition of the possibility of argumentation per se. Why, in other words,
would argumentation become impossible if I do not respect all potential participants in a
discussion?
In other writings, Apel suggests that a universal language game is necessary because
otherwise the kinds of claims that Wittgenstein makes about all languages and all language
games would be impossible (Apel 1975, 259). In other words, although Apel seems to admit
that a particular, limited language community and respect for particular others would be
sufficient for claims made within a particular language community, a universal language
community, including respect for all persons, is a necessary precondition of claims about
language and language communities (e.g. claims about the rule-governed nature of
language and the impossibility of private language), at least insofar as these claims aspire to
have an objective validity.
However, even if one assumes that a universal language game is a necessary condition
of the possibility of making universal claims about language, this only shows that one is
necessarily committed to a universal language community insofar as one makes these kinds
of claims about language (Illies 2003, cf. 86). But surely making these kinds of claims about
language is only an optional practice; writing books like Philosophical Investigations can
hardly be called inescapable. This claim is therefore insufficient to prove that anyone who
engages in argumentation is necessarily committed to a universal language community
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
153
(including the universal moral norm which, according to Apel, is among the necessary
conditions of the possibility of this universal language community).
This shows that discourse ethics is susceptible to an instantiation of the general
dilemma for Kantian constructivism: either argumentation is inescapable, but in that case it
is unclear why this practice of argumentation has a universal moral norm among its
necessary conditions of possibility, or a universal norm is among the necessary conditions
of the possibility of a certain type of (universal) argumentation (i.e. argumentation about
language) – but in that case, argumentation, i.e. universal argumentation, is only an
optional practice.
In what follows, I will analyse to what extent recent attempts at putting forward similar
arguments have succeeded in overcoming this dilemma.
4.2 The Argument from Communication
In this subsection, I discuss Onora O’Neill’s argument from communication. As I have
already mentioned above, O’Neill, at least insofar as I am aware, never explicitly claims that
she puts forward a transcendental argument for a principle of interpersonal morality. In
this section, I try to show that O’Neill’s argument for a supreme principle of morality, the
principle of followability, has a transcendental structure that is very similar to the discourse
ethical project.20 Subsequently, the goal of this subsection is to evaluate to what extent
O’Neill succeeds where discourse ethics fails. I argue that although she provides additional
arguments for accepting a universal and categorical principle of morality, she ultimately
faces the same objection as discourse ethics, i.e. that she fails to show that one is necessarily
committed to a universal principle of interpersonal morality.21
Just like Apel, O’Neill draws on Wittgenstein’s argument against the possibility of a
private language to argue that communication is an essentially public activity which
presupposes the existence of a language community. O’Neill writes that “it is surely
20
The parallel between O’Neill’s work and discourse ethics is also noted by Illies (2003, 68). One interesting
difference is that whereas Apel (and Habermas) criticize Kant for presupposing an idea of solitary agency, O’Neill
defends her theory as an interpretation of Kant. She writes, for instance, “As I see it, Kant’s view of human reason
as the capacity to orient ourselves in thought cannot be properly articulated without taking seriously his claims
that this is a shared rather than a solitary task, so must meet any necessary conditions for coordination among a
plurality of agents. These conditions are also necessary conditions for political coordination” (2011, 542–43).
21
Let me stress that I cannot possibly do justice to O’Neill’s work on the foundations of ethics in this section. In
this section, I focus exclusively on her argument for the principle of followability and her (recent) replies to the
objection that the principle of followability is not a universal principle of morality.
CHAPTER 5
154
controversial to see the speaking, writing, and related activities of human beings as
primarily expressive, something which can in principle be purely private, indeed solitary,
rather than as primarily communicative” (O’Neill 1986, 527). In addition, just like Apel,
O’Neill thinks that communication is an inescapable practice. She claims that even
“[private] expression is parasitic on communication” (O’Neill 1986, 527). The implication
of this Wittgensteinian picture of communication is that, according to O’Neill, the ‘cost’ of
opting out of communication is “to be left in solitary and thoughtless silence” (O’Neill
1986, 539). In other words, communication, according to O’Neill, is inescapable.
O’Neill’s goal is to argue that what she calls the ‘the principle of followability’ is a
necessary condition of the possibility of communication. The principle of followability
entails, according to O’Neill, that “a principle cannot be taken to be universalizable if it
cannot be viewed as a principle for all, because its universal adaptation (per impossibile)
would render some unable to act, a fortiori unable to adopt that principle” (O’Neill 1996,
163). This principle is put forward by O’Neill as an interpretation of the Formula of
Universal Law of the Categorical Imperative, which according to O’Neill essentially boils
down to the question of whether “what one proposes for oneself could be done by others”
(O’Neill 1989a, 94). In other words, according to O’Neill, it is a necessary condition of the
possibility of communication that any principle or maxim of action could in principle be
adopted as a principle of action by all other agents.
In chapter 9, I will come back to this question of what the principle of followability
implies regarding how one has to treat others. For now, I will just assume that the principle
of followability has certain substantive implications regarding how to act, and subsequently
I focus on the question of to what extent O’Neill’s argument for the principle of
followability is convincing.
O’Neill’s argument for the principle of followability clearly has a transcendental
structure. Recently, for instance, she has described Kant’s (and also her own) project as
follows: “Kant’s focus is on the necessary conditions for possible communication or
interaction with others, without restriction” (O’Neill 2013, 223). The structure of O’Neill’s
transcendental argument can be roughly characterized as follows:
1) Communication (X) is inescapable for A.
2) Accepting the principle of followability (Y) is a necessary condition of the
possibility of communication (X).
3) Therefore, A must accept the principle of followability (Y).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
155
In the remainder of this subsection, I will focus on O’Neill’s arguments for premise 2, i.e.
the claim that the principle of followability is a necessary condition of the possibility of
communication.22 O’Neill argues for this claim in the following way:23
What is to be vindicated is not reason, considered in abstraction from any
particular reasoners, but the reasoning of those who like ourselves have no
prescribed modes of coordination, and find that their native endowment provides
neither algorithm nor instinct for acting or for thinking. What can such beings do?
There is no maxim of reasoning whose antecedent authority can compel them; and
yet they cannot share a world if there is no cognitive order. The most then that
they can do is to reject basic principles of thought and action that are barriers to
cognitive order. A minimal, negative step toward any solution must be to refrain
from adopting plans that others cannot adopt. Those who are to be fellow workers
must at least refrain from basing their action on basic principles that others cannot
share. Those who act on such maxims are not guaranteed agreement, at all points;
but if they wholly reject it, communication and interaction (even hostile
interaction, let alone coordination) will be impossible. To act on this maxim is
simply to make what Kant elsewhere calls the Categorical Imperative the
fundamental principle of all reasoning and acting. It is to base action and thought
only on maxims through which one can at the same time will that they be
universal laws” (O’Neill 1989c, 23).
Let me try to unpack this argument. I think it is helpful to stress that O’Neill’s argument is
first and foremost a negative argument: she denies that there are any pre-established
authorities of reason. What she means by this is that principles of (inter)action should not,
for instance, be based on antecedent authorities such as local traditions, the church or
mind-independent moral facts (1986, 538; 1989c, 35; 1996, 54, 77; 2013, 223). The reason
for this, according to O’Neill, is that “thoughts and actions that depend on unvindicated
22
For critical discussions of O’Neill’s argument for the categorical imperative see Besch (2008), Budde (2009) and
Barry (2013). The authors, however, mainly, or even exclusively, focus on O’Neill’s argument for the categorical
imperative in Towards Justice and Virtue (1996).
23
This is just one place in which O’Neill makes this kind of argument. O’Neill makes the same argument, with
minor variations, throughout her work. See, for instance, O’Neill (1986, 545) (1992, 297–98) (2002, 60), (2013,
222). Towards the end of this subsection, I will briefly discuss a recent (2013) defence of this argument in light of
objections put forward by Barry (2013).
CHAPTER 5
156
authorities will hold only where this authority is accepted, so cannot provide general
understanding or agreement or resolve all conflicts of belief and action” (O’Neill 1992,
294). Positively, she claims that the only thing we can do is to not act on the basis of
principles which all others could not adopt as a principle of action. A failure to act on the
basis of the categorical imperative, the principle of followability, would, according to
O’Neill, “defeat the possibility of action, interaction and communication” (O’Neill 1989c,
24). The principle of followability is thus justified (negatively) because there is no maxim of
reasoning with an antecedent authority and (positively) because if one did not act on the
basis of the principle of followability, communication and interaction would break down.
O’Neill’s argument thus roughly goes as follows:
1) Communication and/or the coordination of action is inescapable for A.24
2) There are no pre-established authorities of reason.
3) Therefore, A must refrain from (inter)acting on the basis of principles that others
cannot share.25
The crucial step of the argument is the move from premise 2 to premise 3, i.e. the move
from the claim that there are no pre-established authorities to the claim that one must not
base one’s actions on principles that others cannot share.
My objection to this argument is similar to my objection to Apel’s argument. Recall the
reason why, according to O’Neill, communication is inescapable. Echoing Apel, O’Neill
claims that the alternative is “to be left in solitary and thoughtless silence” (O’Neill 1986,
539). Just like Apel, O’Neill thus wants to force the reader into a dilemma: either one
accepts the principle of followability or one is left in solitary and thoughtless silence. Either
one adopts the principle of followability as the fundamental principle of reason or one
cannot or does not reason at all.26
But it is unclear why we should think that the principle of followability, which is
supposed to be a universal principle that ranges over all other agents, is a necessary
24
I say ‘and/or’ because O’Neill often uses these concepts interchangeably.
A similar argument has recently been put forward by Ronzoni and Valentini (2008). See, for instance, their
claim that “absent a pre-established authority, anyone can, in principle, turn out to be right” (Ronzoni and
Valentini 2008, 412 my emphasis).
26
Cf. “the thought is simply that we do not even offer others reasons for belief unless we aim to be intelligible to
them, and do not even offer others reasons for action unless we make proposals that we think adoptable by them”
(O’Neill 2013, 222). See also: “anything that is to count as reasoning must be followable by all relevant others”
(O’Neill 1996, 3).
25
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
157
condition of the possibility of communication or coordination per se. True, if one relies on
principles which cannot be followed by some others, it might be impossible for these
particular others to communicate with us and for us to communicate with them, but this
does not, it seems, imply that communication as such becomes impossible and it does not
show why, if we do not adopt the principle of followability, “coordination and
communication fail” (O’Neill, 2002, p. 63). At most it shows that insofar as we aim to
communicate with particular others we have to accept the principle of followability. But
surely communicating with particular others is not inescpable.
Moreover, it is not entirely clear why exactly O’Neill rejects pre-established authorities.
To some extent this rejection already seems to presuppose the principle of followability.
Her reason to reject pre-established authorities is that these “will hold only where this
authority is accepted” (O’Neill 1992, 294). Although this might be true, the limited
application of pre-established authorities is only problematic if one presupposes that
principles of reason should be acceptable by all others. At this point of the argument this
cannot simply be presupposed, because this is exactly what is at stake. In other words, it is
not clear why we should reject pre-established authorities as possible sources of reason, and
O’Neill would have to explain why communication breaks down if one relies on such
authorities.
O’Neill’s dilemma thus seems to be a false dilemma. It seems to be perfectly possible to
communicate while restricting the community of communication to some particular
others. In other words, Wittgensteinian publicity, the idea that meaningful thought and
action presupposes a language community, is not sufficient to justify Kantian publicity, the
idea that the relevant language community includes all other agents and that one should
only act on the basis of principles which could be adopted as principles of action by all
others.27 If one were to deny Wittgensteinian publicity, one would indeed be left in solitary
and thoughtless silence, but it is unclear why denying Kantian publicity has the same
consequence. Or at least, this conclusion does not appear to follow from the
Wittgensteinian argument against private language alone, and it also does not seem to
27
The claim that the language community includes all others and the claim that one has to act on the principle of
followability are two distinct steps. Even if the first claim were true, it does not necessarily follow that one has to
grant an equal moral status to all the members of the language community, in the sense that one has to act only on
a principle which could be adopted as a principle of action by all others. For instance, holding others as a slave is
most likely in conflict with the principle of followability, but is not necessarily in conflict with acknowledging that
slaves are part of a universal language community. In contrast to Apel, O’Neill does not make a clear distinction
between these two claims.
CHAPTER 5
158
follow from the rejection of pre-established authorities.
In a recent text, O’Neill anticipates this kind of objection and responds as follows:
Why, one might ask, does ethical reasoning have to offer reasons to all? Many have
thought otherwise, and have claimed that ethical reasoning may, even must, draw
on the established beliefs, conventions or ideologies of particular communities and
traditions, so will be fit to reach some but not others. Kant saw this as a mistake.
Reasoning that is premised on local or contingent beliefs, conventions or
traditions is incomplete reasoning because it offers reasons only to some and not
to all, and assumes various contingent beliefs as premises. Such reasoning, as Kant
sees it, is inevitably less than fully public. It is heteronomous reasoning, in that it
assumes some other source or authority, and while many admirable acts reflect
heteronomous reasoning, the touchstone of reason is not to rely on extraneous or
arbitrary assumptions. Fully reasoned belief and action seek to rely on reasons that
could be offered to all others, so aim to be fully public in the Kantian sense
(O’Neill 2013, 223).
The negative argument which I discussed above claimed that principles of (inter)action
ought not to rely on pre-established authorities of reason. Here, O’Neill makes explicit what
was implicit in the argument described above, namely that any principle which can only be
followed by particular others (instead of by all others) would presuppose the existence of
such a pre-established authority. Given that, according to O’Neill, we should not rely on
these kinds of pre-established authorities, the only non-arbitrary principle of reason would
be to rely on principles which could be followed by all others. 28 This argument thus
provides the link between the rejection of relying on pre-established authorities and the
acceptance of the principle of followability.
However, even if we grant, for the sake of argument, that we have to deny the existence
of pre-established authorities, O’Neill has not given us any reason to believe that relying on
a restricted community of communication is the same as relying on a non-existing pre-
28
A similar argument is made by Ronzoni and Valentini: “the validity of principles of justice depends on their
being agreed upon under conditions in which everyone is free and equal because there is no moral truth one is
rationally justified to appeal to in order to warrant the exclusion of others from the deliberation process” (Ronzoni
and Valentini 2008, 411). Ronzoni and Valentini thus similarly move from the claim that one cannot appeal to
moral truths to the conclusion that no one should be excluded from the deliberation process.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
159
established authority. Say that in deliberation I exclude those persons who only reason on
the basis of selfish premises. Why would I, in that case, invoke a pre-established authority?29
I do not necessarily have to presuppose the truth of my ideology or religion in order to have
good reasons to exclude certain others from deliberation. More importantly, O’Neill has
not given us any reason to believe that communicating with a universal community of
communication is any less arbitrary than communicating with a restricted community of
communication. Why is it more arbitrary to only reason with some, instead of reasoning
with all? This judgement seems to rely on a moral judgement that we should reason with all
others. I do not see how this kind of universalism could follow simply from what it means
to reason.
Let me conclude this section by briefly discussing a final argument that O’Neill puts
forward to argue for the universality of morality. In a chapter called ‘Distant Strangers’,
O’Neill (2000a) argues that the question of moral status, i.e. the question of whom should
be included in the moral community should be understood as a ‘practical question’:
For practical purposes it might be enough to answer the specific question “To
whom are we (or am I) committed to according moral standing in acting or in
living this way?” In posing this question we ask what assumptions we are already
building into our actions, habits, practices and institutions. If in acting we already
assume that others are agents and subjects then we can hardly deny this in the next
breath (2000a, 191–92).
O’Neill’s point is that in acting we typically presuppose that others have certain agential
capacities,30 and she believes that this practical approach to moral status can show that the
moral community should include all human agents (2000a, 192).
O’Neill illustrates her practical approach by arguing that even the Nazis had to
presuppose that their victims had certain agential capacities, even though they explicitly
denied that their victims had these capacities:
29
Cf. “Even if we agree with the Kantian constructivist that there are no good grounds for accepting any
metaphysically robust conception of agents or norms, and so agree that any principles based upon these are
epistemically unjustified, we may still have other reasons to think that the agents who disagree with us do not
deserve to have their views count as possible disqualifiers of principles we find appealing. We may view them as
selfish, silly, or unintelligent, and so deny that their viewpoints deserve equal standing with ours” (Barry 2013, 25).
30
There are certain parallels between O’Neill’s argument and Darwall’s argument from the second-person
standpoint, which I discuss below.
CHAPTER 5
160
It is, of course, quite common for people to deny that certain others are agents or
subjects. Their actual assumptions will, however, be revealed not by avowals or
denials, but in the ways in which they organize and adjust their action to take
account of other’s capacities to act, to suffer and to be influenced. Notoriously
some Nazis claimed that some of their victims lacked moral standing, that they
were Untermenschen. Yet the Nazi’s actions reveal that they in fact assumed that
those whom they persecuted were intelligent, foresighted, literate agents capable of
complex mental and physical suffering. None of the organization of the
deportations or of the camps makes sense except against these background
assumptions. All the subterfuge, the bureaucratic formalities of deportation, the
rhetoric of belittlement, the techniques of control make sense only on the
assumption that the victims were indeed seen as intelligent agents and vulnerable
subjects (O’Neill 2000a, 193–94).
I think O’Neill’s approach is successful insofar as it aims to show that if one stands in a
certain practical relation towards others, one has to presuppose certain agential capacities
on their part, even though one might explicitly deny their having these capacities. I do not,
however, think that this argument is sufficient to lead to universalist moral conclusions.
The reason for this is that these universalist conclusions only follow if one already
assumes that anyone with certain capacities should be included in the moral community.
The problem, however, is that O’Neill does not argue for the claim that anyone with certain
capacities has an equal moral standing, nor does she explain why this principle is among
the necessary conditions of the possibility of communication. In other words, O’Neill here
presupposes that every being with certain capacities has an equal moral standing. Although
it is not impossible to argue for this principle, it is hard to see how this principle could be
justified in terms of the necessary conditions of the possibility of communication
O’Neill has thus not given us a convincing reason to think that a universal principle of
interpersonal morality is among the necessary conditions of the possibility of
communication. Again, there seems to be a gap between showing that certain normative
principles are among the conditions of the possibility of communication and the claim that
communication is only possible if one accepts the categorical imperative. O’Neill thus
seems to be susceptible to the general dilemma for Kantian constructivism: although she
might be right that communication is inescapable, this kind of communication does not
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
161
have a universal principle of interpersonal morality among its necessary conditions of
possibility.
4.3 The Argument from the Publicity of Reasons
In this subsection, I discuss a final attempt to justify universal interpersonal morality on
Wittgensteinian premises: Korsgaard’s attempt to justify interpersonal morality on the basis
of an argument from the publicity of reasons. Having established through an argument
from the first person that individuals should value their own humanity, Korsgaard tries to
justify claims about interpersonal morality (obligations to respect the humanity of others)
by reference to an account of the public nature of reasons. Korsgaard argues that claims of
interpersonal morality can only be justified once it is established that reasons are
intrinsically public, i.e. that reasons always have normative force not just for yourself but
also for others. Instead of trying to argue from private reasons to public reasons (which,
according to Korsgaard, is the problem with an argument from the first person), she thus
tries to argue that reasons are public all the way down. In order to argue for this conclusion,
Korsgaard brings in “some help from Wittgenstein” (Korsgaard 1996, 136).31
Korsgaard takes the view that Wittgenstein has shown that “private language is
inconsistent with the normativity of meaning” (Korsgaard 1996, 137). The reason for this is
that concepts can only be normative if there is a public standard to distinguish between
their correct and incorrect application. Korsgaard’s main claim is that just as it is a mistake
to think that the meaning of concepts could be private, it is also a mistake to think that
reasons could be private (Korsgaard 1996, 137). She writes:
We could make a parallel argument against private reasons: reasons are relational
because reason is a normative notion: to say that R is a reason for A is to say that
31
Coleman (2005) argues that Korsgaard’s Wittgensteinian argument only tries to show that sometimes reasons
can be public, not that they necessarily are public (Coleman 2005, 321). According to Coleman, Korsgaard’s
argument for the necessity of the publicity of reasons is the argument Korsgaard borrows from Thomas Nagel and
which goes as follows: “Suppose that we are strangers and that you are tormenting me, and suppose that I call
upon you to stop. I say: ‘How would you like it if someone did that to you?’ And now you cannot proceed as you
did before. Oh, you can proceed all right, but not just as you did before. For I have obligated you to stop”
(Korsgaard 1996, 142–43). In recent writings, however, Korsgaard explicitly claims that her publicity of reasons
argument is supposed to be analogous to Wittgenstein’s private language argument (Korsgaard 2009, 196f12). I
will therefore focus on her Wittgensteinian argument, and more specially Korsgaard’s (2009, 191–97) most recent
defence of the argument in light of criticism of the publicity of reasons argument in The Sources.
CHAPTER 5
162
one should do A because of R; and this requires two, a legislator to lay it down,
and a citizen to obey. And the relation between them is not just causal because the
citizen can disobey: there must be a possibility of irrationality or wrongdoing.
Since it is a relation, and indeed a relation in which one gives a law to another, it
takes two to make a reason (Korsgaard 1996, 137–38).
This argument leads Korsgaard to conclude that reasons are “public in their very essence”,
(Korsgaard 1996, 135) which means that “to act on a reason is already, essentially, to act on
a consideration whose normative force may be shared with others” (Korsgaard 1996, 136).
Note that in this passage Korsgaard only states that reasons may be shared with others, not
that they must be shared with others. In other places, however, Korsgaard makes it clear
that she thinks the public reason argument leads to necessary commitment, that is, “what
both enables us and forces us to share our reasons is, in a deep sense, our social nature”
(1996, 135). Because only the latter claim could succeed in justifying a universal and
categorical principle in the first place, I will assume that Korsgaard holds the stronger
claim, i.e. the claim that reasons must be shared with others.
Korsgaard’s argument has roughly the following structure:
1) Acting for a reason (X) is inescapable for A.
2) The publicity of reasons (Y) is a necessary condition of the possibility of having a
reason for action (X).
3) Therefore, A must accept that reasons are public (Y).32
Korsgaard combines the conclusion of her argument from the publicity of reasons with the
conclusion of her argument for the value of humanity: ‘you must value your humanity’.
Korsgaard argues that if one necessarily has to value one’s humanity, and if reasons are
32
One might wonder to what extent this argument is really an argument from the second person insofar as its
starts from agency and not from social practice. At the same time, however, the argument aims to show that
agency is intrinsically social, which brings it closer to arguments from the second person. Whether or not this
argument should be classified as an argument from the second person, Korsgaard’s revised argument in SelfConstitution (2009) on the basis of shared action, which I discuss below, clearly is. Whether or not Korsgaard
defends a transcendental argument from the second person thus depends on how exactly one defines these
arguments and which Korsgaard one reads. The appropriate classification of Korsgaard’s argument is not made
easier by her own suggestion that she is “always making the same argument” (Korsgaard 2009, 76). Ultimately, I
am interested in the substance of her argument, and not in what kind of argument she makes. I will therefore
analyse Korsgaard’s argument, whether or not it actually qualifies as an argument from the second person.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
163
necessarily public, so that the normative force of a reason is shared with others, then every
agent necessarily has to value the humanity both in him- or herself and in all other agents.
Before taking a closer look at Korsgaard’s argument, it might be helpful to mention two
important differences between Korsgaard’s argument and the previous two arguments I
have discussed. The first difference has to with the role of the private language argument.
Both Apel and O’Neill try to justify interpersonal morality by building on Wittgenstein’s
private language argument through an investigation of the conditions of the possibility of
having a language community (which according to Wittgenstein is a necessary condition of
the possibility of meaningful language use).33 Korsgaard, on the other hand, merely puts
forward an analogous argument to Wittgenstein’s private language argument. In SelfConstitution, Korsgaard writes about her own publicity of reasons argument:
I did not intend to suggest that the publicity of reason can be inferred from the
publicity of meanings. I meant rather to be making an argument for the publicity
of reason that is analogous to Wittgenstein’s argument for the publicity of
meaning. Wittgenstein’s argument, as I understand it, is intended to show that
meaning can’t be normative at all—you can’t be wrong—unless it is public. My
argument was meant to show that reasons cannot be normative at all unless they
are public (Korsgaard 2009, 196f12).
According to Korsgaard, interpersonal morality cannot be inferred from the private
language argument, but it can be inferred by applying an analogy to the private language
argument to reasons.
The second distinction is that Korsgaard does not reject an argument from the first
person altogether. It is the combination of the argument from the first person for the
conclusion that one necessarily has to value one’s humanity, together with the argument for
the social nature of reasons, which should lead to categorical and universal moral duties.34
According to Korsgaard, an argument from the first person without the idea of the
publicity of reasons cannot lead to conclusions about universal interpersonal morality,
33
This is only partly true. As we have seen, O’Neill also introduces an argument about the rejection of preestablished authority to argue for the categorical imperative.
34
Cf. “to act on a reason is already, essentially, to act on a consideration whose normative force may be shared
with others. Once that is in place, it will be easy to show how we can get someone who acknowledges the value of
his own humanity to see that he has moral obligations” (Korsgaard 1996, 136).
CHAPTER 5
164
because it can only lead to the conclusion that I necessarily have to value my own humanity
(and it can at most show that other agents also have to value their own humanity). But an
account of the social nature of reasons that does not include the idea that one necessarily
has to value one’s humanity cannot lead to conclusions about categorical interpersonal
morality, because in that case there would be no reason to treat some reasons (the value of
humanity) as more important than others. In other words, the argument from the first
person is supposed to justify categorical reasons, whereas the argument from the publicity
of reasons is supposed to justify universal reasons.
In the remainder of this subsection, I will argue that, despite these differences,
Korsgaard is susceptible to the same objection as Apel and O’Neill: she equivocates between
Wittgensteinian publicity, i.e. publicity as communicability, and Kantian or Korsgaardian
publicity, i.e. publicity as the shareability of reasons between all rational agents (respect for
the value of the humanity of all agents). My objection is that although Korsgaard might
succeed in showing that Wittgensteinian publicity is a necessary precondition of having a
reason, this is not sufficient to justify Kantian publicity. Korsgaard’s argument thus falls
prey to the first horn of the dilemma for Kantian constructivism: she fails to show how
universal and categorical reasons (Kantian publicity) can be justified on the basis of an
inescapable starting point (having reasons for action). Or at least her Wittgensteinian
argument from the publicity of reasons does not lead to conclusions about universal and
categorical universal morality. In addition, I argue that her most recent (2009) defence of
(Kantian) publicity in terms of the necessary preconditions of shared action falls prey to the
second horn of the general dilemma for Kantian constructivism: even if Kantian publicity is
a necessary precondition of shared action, shared action (or at least in the way it is
understood by Korsgaard) is merely optional.
Korsgaard’s equivocation between different conceptions of publicity has been noted by
several commentators on Korsgaard’s work (LeBar 2001; Gert 2002; Skidmore 2002;
Wallace 2009). Korsgaard assumes that a Wittgensteinian argument for the publicity of
reasons implies that the reasons of an individual are not just normative for him- or herself
but also for other individuals. But this conclusion about interpersonal morality does not
seem to follow from the Wittgensteinian argument that Korsgaard relies on. A broadly
Wittgensteinian view of reasons would indeed hold that reasons should in some sense be
public in order to be reasons in the first place. The sense in which reasons should be public
is that there should be a public standard to determine the correct application of a reason
(just as there should be a public standard to determine the correct application of any
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
165
concept). Reasons, in other words, should be communicable (cf. Korsgaard 1996, 137). The
argument for this conclusion is that without reasons being public it would not be possible
to distinguish the correct from the incorrect application of a reason, which essentially
undermines the normativity of reasons, even in the individual case.35 This argument for the
publicity of reasons seems to be analogous to Wittgenstein’s private language argument,
according to which one could not distinguish between the correct and incorrect
applications of any concept without them being public. So if there are to be any normative
reasons, these reasons have to be public, in the sense that their correct application is to rely
on some public standards, which implies that reasons should be communicable. So far, so
good.
However, publicity in this minimal Wittgensteinian sense is not the same as treating
your reasons as my own reasons. The fact that the correct application of a (private) reason
requires certain public standards does not commit me to ‘share’ your reasons beyond being
potentially part of a language community which determines the correct application of your
reasons. The Wittgensteinian idea of the publicity of reasons still leaves it open to me to
treat your reasons as mere obstacles in the pursuit of my own ends. As long as some
language community that can determine the correct or incorrect application of reasons
exists, the necessary preconditions of the normativity of reasons are satisfied. The
Wittgensteinian publicity of reasons, in other words, is not yet a Korsgaardian or Kantian
publicity of reasons.
In Self-Constitution, Korsgaard seems to acknowledge that her extension of the private
language argument to reasons is not sufficient to argue for the publicity of Korsgaardian or
Kantian reasons, because she now puts forward a different, or revised, argument from the
publicity of reasons. Korsgaard puts forward the following argument by way of illustration
in order to move beyond the minimal Wittgensteinian interpretation of the publicity of
reasons:36
35
Cf. “That’s why ‘following a rule’ is a practice. And to think one is following a rule is not to follow a rule. And
that’s why it’s not possible to follow a rule ‘privately’; otherwise, thinking one was following a rule would be the
same thing as following it” (Wittgenstein 2009, para. 202).
36
Korsgaard herself seems to suggest that she is in fact making the same argument and that many readers have a
‘misimpression about her argument’ in The Sources (Korsgaard 2009, 196f12).
Although I think the argument in Self-Constitution is different from the private reason argument as described
above, Korsgaard is correct that traces of the argument in Self-Constitution can be found in the earlier The Sources.
See, for instance, her remark that “it is nearly impossible to hear the words of a language you know as mere noise”
(Korsgaard 1996, 139).
CHAPTER 5
166
Suppose you and I are related as student and teacher, and we are trying to schedule
an appointment. “Stop by my office right after class,” I say, thinking that that will
be convenient for me, and hoping that it will also be convenient for you. It isn’t, as
it turns out. “I can’t,” you say, “I have another class right away.” So I have to make
another proposal. It’s important to see why I do have to do this: it’s because having
the meeting is something that we are going to do together (Korsgaard 2009, 192).
Korsgaard’s point is that the student’s reason necessarily also has to be a reason for
Korsgaard at the same time, because they are planning to perform a shared action, i.e. they
are planning to have a meeting together. According to this argument, the publicity of
reasons, in the Korsgaardian or Kantian sense of your reason being normative for me, is
thus a necessary precondition of the possibility of acting together. Thus, Korsgaard
concludes, “to perform a shared action, each of us has to adopt the other’s reasons as her
own, that is, as normative considerations with a bearing on her own case” (Korsgaard 2009,
192).
This revised argument could be roughly summarized as follows:
1) Shared action (X) is inescapable for A.
2) (Kantian) publicity of reasons (Y) is a necessary condition of the possibility of
shared action.
3) Therefore, A must accept the Kantian publicity of reasons (Y).
Does this revised argument provide a convincing basis for a Korsgaardian or Kantian
publicity of reasons?
Let us grant for the sake of argument that Korsgaard is correct in suggesting that your
reason being normative for me is a necessary precondition of shared action. The relevant
question then becomes whether engaging in shared action is inescapable. That is, whether it
is in fact necessary for Korsgaard (or the student) to engage in shared action and, insofar as
Korsgaard aims to justify a universal moral principle, whether it is necessary to engage in
shared action with all persons. After all, if one is committed to sharing reasons only insofar
as one is engaging in a shared action with someone else, the argument can only justify
universal interpersonal morality if it is inescapable to engage in shared action with all other
persons (and not just with your student).
Korsgaard believes that shared action is in fact inescapable. In order to argue for this
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
167
conclusion, Korsgaard invites the reader to consider the alternative, in which she does not
treat the reasons of the student as her own reasons. Korsgaard suggests that, in this case, she
might respond to the student’s remark that she cannot make it because she has another
class by saying, “[W]ell, just skip it” (Korsgaard 2009, 193).
According to Korsgaard, there is something deeply problematic about this alternative,
for “obviously, we can’t relate at all on those terms. So if that’s how it is, no personal
interaction is going to be possible” (Korsgaard 2009, 193). In this case, the only kind of
interaction that is possible is purely strategic interaction with the student in which her
reasons are treated merely as “possible tools and obstacles, things that might help me to
achieve my ends or get in my way” (Korsgaard 2009, 193). But this, Korsgaard claims,
would amount to “a kind of war, or combat” (Korsgaard 2009, 194).37
However, even if we agree that living in a constant state of war is not a real option for
us, this argument is not sufficient to justify the categorical and universal moral obligations
that Korsgaard wants to argue for. The problem with Korsgaard’s argument is that even if
we grant that we cannot live in a constant state of war and that engaging in shared action
commits us to treating the reasons of the other person as our own reasons, this does not
commit us to engaging in shared action with all other human beings and consequently to
valuing the humanity of all other agents. What, on Korsgaard’s account, would be wrong
with the racist who normally treats other people’s reasons as his own reasons, except when
the other person is someone of another race? The racist is perfectly capable of engaging in
personal interaction and in shared action. His life can hardly be described as a constant
war. But the racist refuses to engage in shared action with specific others and is therefore in
these cases not committed to any of the necessary presuppositions of shared action.
Korsgaard thus presents the reader with a false dilemma (similar to Apel’s and O’Neill’s
false dilemmas). Korsgaard suggests that one has to value the humanity of all agents or
one’s life is a constant war, but this overlooks the possibility that one might only engage in
shared action with certain particular others or only value the humanity of some others.
Korsgaard’s revised argument from the publicity of reasons thus justifies neither
universal nor categorical interpersonal morality. It does not justify universal morality,
because Korsgaard fails to show that it is necessary to engage in shared action with all
others. It does not justify categorical interpersonal morality, because although it might not
37
Korsgaard’s reasoning seems to very similar to Darwall’s argument for the inescapability of the second-person
standpoint, which I will discuss in the next section.
CHAPTER 5
168
be optional to engage in shared action altogether, it is optional when and with whom one
engages in shared action. Whether or not someone is committed to valuing the humanity of
others thus depends on whether the person wants to engage in interaction with the other
person in the first place – even if we accept that completely opting out of interaction is not
a real possibility.
A certain pattern begins to emerge. Although Korsgaard’s argument from the publicity
of reasons and from shared action is very different from the arguments put forward by Apel
and O’Neill, it faces a similar objection. Korsgaard fails to show that there is a kind of
interaction which is inescapable and which at the same time has a universal principle of
interpersonal morality among its necessary conditions of possibility. The Wittgensteinian
publicity of reasons is not sufficient to justify Kantian publicity, i.e. the communicability of
reasons is not the same as treating the reasons of all other persons as normative for you.
And even if it is true that one has to value the humanity of the persons with whom one
actually engages in shared action (such as the example of Korsgaard and her student), it
seems to be merely optional to engage in this kind of shared action in the first place.
4.4 The Argument from the Second Person Standpoint
The previous three authors all draw on Wittgenstein’s private language argument to argue
for the (Kantian) publicity of reasons (although they also draw on additional arguments
that are not derived from Wittgenstein). In this subsection, I discuss Stephen Darwall’s
(2006; 2010) argument from what he calls ‘the second-person standpoint’.38 Instead of
drawing on insights about the publicity of language and argumentation, Darwall argues that
anyone who takes up this second-person standpoint is committed to the categorical
imperative.
Darwall defines the second-person standpoint as the perspective persons take up when
they make and acknowledge claims on one another’s conduct and will (Darwall 2006, 3).
According to Darwall, persons take up this standpoint not only when they explicitly make
claims on one another but also by being part of a form of life in which they are the subject
and object of so-called reactive attitudes, i.e. the distinctive attitudes involved in holding
people responsible (Darwall 2006, 17; Darwall 2010, 216). Importantly, the claims and
38
Darwall claims that his argument is compatible with both (Kantian) constructivism and realism (2006, 293).
However, as I have argued in chapter 3, a transcendental argument cannot lead to realist conclusions. I therefore
think that Darwall’s theory should be interpreted as a form of Kantian constructivism.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
169
attitudes need not be moral claims and attitudes; they can be any kind of claim and attitude
that (implicitly) makes a demand on others (Darwall 2006, 248). Persons are thus engaged
in second-personal interaction when they exchange reasons and make explicit demands on
one another, but also when they feel resentment, anger or forgiveness.
According to Darwall, this second-person standpoint has among its necessary
conditions of possibility a universal and categorical moral principle, and more specifically a
commitment to the equal dignity of all persons. Darwall writes that “the address of secondpersonal reasons of any kind carries presuppositions that, when fully worked out, commit
addresser and addressee alike to their second-personal competence and to an equal secondpersonal authority rooted in that, hence to the equal dignity of [all] persons and to morality
as a form of mutual accountability” (Darwall 2006, 81). 39 Darwall thus presents his analysis
of the necessary preconditions of the second-person standpoint as resulting in an argument
for the categorical imperative (Darwall 2006, 115–18). Darwall’s argument has roughly the
following transcendental structure:
1) The second-person standpoint is inescapable (X) for A.
2) A commitment to the equal dignity of all persons (Y) is a necessary condition of
the possibility of the second-person standpoint (X).
3) Therefore, A is necessarily committed to the equal dignity of all persons (Y).
Let me first briefly discuss how exactly Darwall tries to derive normativity from the secondperson standpoint before turning to the more specific question of how Darwall’s analysis is
supposed to justify a categorical and universal principle of interpersonal morality.
According to Darwall, taking the second-person standpoint has certain “normative
felicity conditions”, which are conditions that prescribe “what must be true for secondpersonal reasons actually to exist and be successfully given through second-personal
address” (Darwall 2006, 4).40 Darwall claims that the second-person standpoint has two
normative felicity conditions. The first condition is that in making a second-personal
address, one has to assume that the addressee has ‘second-personal competence’: the
39
Darwall makes it clear that the equal dignity of persons applies to all persons (Darwall 2006, 126; Darwall 2010,
227).
40
Darwall is not clear about the question of whether the felicity conditions are conditions of something to count as
a second person interaction at all or for something to count as a successful second person interaction. This makes
the argument susceptible to the kind of worries voiced by Katsafanas (2013).
CHAPTER 5
170
capacity and competence to respond to a demand as a reason (Darwall 2006, 21). The
second condition is that any second-personal address presupposes the authority and
responsibility of both the addresser and the addressee as equal, free and rational agents to
determine themselves by second-personal reasons (Darwall 2006, 22). According to
Darwall, taking the second-person standpoint thus commits one both to acknowledging
others as having the competence to engage in a second-person relationship and to
acknowledging the authority to contribute to this relationship as a free and equal person.
Taken together, these two normative felicity conditions show, according to Darwall, that
“the dignity of persons, conceived of as a shared basic second-personal authority, is
inescapably presupposed from the second-person standpoint as a normative felicity
condition of addressing second-personal reasons and, therefore, that autonomy of the will
is as well” (Darwall 2006, 245).
Darwall illustrates this general argument by way of the following example: suppose
someone stands on your foot and you either demand or request this person to remove her
foot from your foot. In that case, Darwall states that “you and she commonly presuppose
that she can freely comply if she finds your request or demand one she could not
reasonably reject, regardless of what she desires or how strongly she desires it” (Darwall
2006, 245). In other words, if you address someone, you have to presuppose that the
addressee is capable of responding to reasons. In addition, Darwall claims that this secondpersonal address not only presupposes the capacity (second-personal competence) to
respond to this address as a person but also the normative authority to engage in this
relation as a free and equal person: “[the second-person standpoint] commits both parties
… to the one normative standing that is inescapable for them both, that of a free and
rational person or will as such” (Darwall 2006, 246). Darwall’s point is that if you make a
request of someone, you have to presuppose that you have the authority to make this
request and that the addressee can acknowledge your authority, and consider your request
or demand, as a free and rational person. Darwall concludes that “this means that a
commitment to autonomy and the moral law, conceived, à la Kant, as laws of freedom for a
realm of ends, is inescapable from the second-personal standpoint” (2006, 246).41
Much more needs to be said about Darwall’s analysis of the second-person standpoint,
41
According to Darwall, this analysis also applies to hierarchical relationships (2006, 258), including the relation
between a sergeant and a private (2006, 260), and even to the relation between a slave and a slaveholder (2006,
265). For the sake of argument, I assume that Darwall is right to claim that his analysis also applies to these kinds
of hierarchical relations.
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
171
but this should be sufficient to get a rough idea of the kind of argument that Darwall is
trying to make. In the following, I will assume, for the sake of argument, that Darwall is
right to suggest that in taking the second-person perspective one is necessarily committed
to the autonomy and dignity of both the addressor (yourself) and the addressee (the other
person(s)) of a second-personal address. I argue, however, that even if certain normative
claims follow from the second-person standpoint, Darwall does not show that the secondperson standpoint entails a categorical and universal principle of interpersonal morality.
That is, Darwall fails to argue that among the felicity conditions of the second-person
standpoint is a commitment to the dignity of all persons.
It is not clear how exactly Darwall tries to argue that the second-person standpoint
commits one to a categorical and universal principle of interpersonal morality. The reason
for this is that Darwall is not entirely clear about the sense in which the second-person
perspective is supposed to be inescapable or what it is exactly that is presupposed by the
second-person perspective. In what follows, I therefore discuss two possible reconstructions
of Darwall’s argument for this conclusion and argue that both are susceptible to (a variant
of) one of the horns of the general dilemma of Kantian constructivism: either the secondperson standpoint is escapable or the second-person standpoint is in some sense
inescapable, but in the latter case it is unclear why it would have a universal principle
among its necessary conditions of possibility.
On the first reconstruction of Darwall’s argument, one is committed to respecting the
dignity of persons insofar as one is actually involved in a second-person relationship with
another person.42 On this reconstruction, the argument could only justify a categorical and
universal principle of interpersonal morality if it were somehow inescapable to engage in a
second-person relationship with all others.
Consider, again, the example of someone who stands on my foot. Imagine that instead
of asking the other person to remove her foot from my foot (and thereby constituting a
second-person relationship with the other person), I simply force the other person from my
foot (just like I would remove a piece of wood from my foot). In this case, it seems that I am
not committed to the felicity conditions of the second-person standpoint, because I do not
stand in a second-person relationship in the first place, and I am therefore not committed
42
I do not want to claim that this is the most charitable reconstruction of Darwall’s argument. The point is merely
to show that there are different ways to read Darwall and that these different ways of reading Darwall have their
own problems. There is, however, textual support for this interpretation, as I show in the text.
CHAPTER 5
172
to respect the dignity of the person who stands on my foot. 43 In other words, the secondperson standpoint seems to be escapable, and the normative conclusions that follow from
the second-person standpoint therefore only have to be accepted by those who actually
engage in this standpoint.44 Darwall anticipates this objection when he writes:
A natural response to this argument may be to agree that if the pre- supposition
that addressees can freely acknowledge and make authoritative demands of
themselves from such a common standpoint is part of the very idea of al address,
then perhaps it does commit us to the equal dignity of free and rational persons,
but then to deny that al address, so understood, is anything we need have much of
a stake in. (Darwall 2006, 138)
Darwall’s response to this objection is ambiguous. On the one hand, he reluctantly
acknowledges that the second-person standpoint is only an optional standpoint. He writes,
for instance, that “if there is nothing that purports to be authentic second-personal address
of the kind we have been concerned with, the argument simply does not apply” (Darwall
2006, 266).
In addition, in the context of a discussion about the relation between a slaveholder and
a slave, Darwall rather surprisingly acknowledges explicitly that the slaveholder does not
necessarily have to recognize the slave’s dignity, but that the slaveholder only has to do so
insofar as he takes up the second-person perspective towards the slave:
I am not saying that if a slaveholder claims to the world at large that he has
authority over his slaves, this commits him to recognizing his slaves’ dignity
either. In this case, he addresses his claim not to his slaves but to others, and
whereas my argument would then require us to conclude that he must presuppose
their dignity, it would not entail that he must presuppose that of his slaves. He
might simply be claiming that his slaves are his property, like his cattle, and that he
43
This interpretation is, for instance, supported by the following remark: “Not every form of address, however, not
even every form of ‘demanding’ address, is second-personal or invokes second-personal reasons, in my sense.
Only the address of putatively valid claims and demands, which consequently presuppose an authority to address
them, is. If someone tries to force his way past you, saying ‘Move over buddy!’, although his utterance is
grammatically second-personal (having ‘you’ as its implied subject), it is not the kind of address with which I am
concerned—no putative authority or second- personal reasons need be involved” (Darwall 2010, 217 my emphasis).
44
A similar objection is also raised by Korsgaard (2007, 20–21).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
173
therefore has, within certain limits, rights to do with them as he will, which claims
his addressees must accept. Such a claim would clearly not commit him to his
slaves’ dignity as persons any more than he would be committed to any such
doctrine with respect to his cattle. My claim is rather that if he purports to address
a second-personal reason based on this claim to his slaves, then this secondpersonal address would commit him to the pre-supposition that he and they share
an equal (second-personal) standing just as free and rational persons (Darwall
2006, 267).
In this passage, Darwall thus seems to acknowledge that the recognition of the dignity of
other persons (including the dignity of slaves) is contingent on taking the second-person
perspective. This implies that the moral conclusions that Darwall defends cannot be
categorically valid, but that they are only valid for those who happen to address each other
in a second-personal way. In addition, the conclusions will not be universally valid because
they only entail moral respect for the persons with whom one actually engages in a secondperson relationship.
On the other hand, Darwall goes to great lengths to try to defend the inescapability of
the second-person standpoint. He stresses, for instance, that although it might indeed be
possible to not engage in second person interactions, it is highly uncommon among human
beings to not so engage. He writes that “even tyrants like Stalin rarely reject the secondperson standpoint outright” (Darwall 2006, 266; See also Darwall 2006, 139). Thus, Darwall
concludes that “it is actually quite rare, I think, for human beings to reject the secondperson standpoint outright, since, for one thing, our emotional lives are full of reactive
feelings and attitudes that essentially involve it” (Darwall 2006, 140).
Whether or not this reply is sufficient depends on what exactly is presupposed by the
second-person standpoint. If one is only committed to respecting the dignity of the
particular addressee of a second-person address, the claim that one cannot reject the
second-person standpoint outright is not sufficient. In order to see why, we have to make a
distinction between rejecting the second-person standpoint outright and rejecting engaging
in second-personal practices in specific instances. Although we might agree with Darwall
that it is very uncommon for a human being to never engage in any second-personal
address whatsoever, it seems less uncommon not to engage in second-personal addresses
towards specific others, or in specific instances. In other words, Darwall might be right that
it is an inescapable part of our phenomenology that we engage in second-personal
CHAPTER 5
174
relationships, but it does not seem to be phenomenologically inescapable to engage in a
second-personal relationship with all human beings.
This ambiguity in the notion of the escapability regarding second-personal interactions
is a problem for Darwall’s argument, because the moment we treat other humans as objects,
i.e. the moment we do not exchange reasons with others or exhibit reactive attitudes, there
seems to be no reason, in Darwall’s theory, to respect those other human beings as persons
with dignity (as he himself seems to admit in the example of the slaveholder quoted above).
So although one might not escape the necessary preconditions of the second-person
standpoint altogether (because sooner or later one will take up the second-person
standpoint towards a particular other), one might escape moral demands in specific
instances by being reluctant to engage in particular second-personal interactions. Darwall’s
argument thus seems to be susceptible to a specific variant of the escapability objection: it
seems to be optional when and towards whom one takes up the second-person standpoint.
The acceptance of the preconditions of the second-person standpoint is therefore
contingent on one’s actually taking up the second-person standpoint. If this is correct, the
moral principles that follow from the second-person standpoint are neither categorical
(because they are contingent on one’s willingness to engage in specific second-person
relationships with others) nor universal (because one might refuse to engage in second
person practices with specific others).
Darwall’s argument could, however, be reconstructed in a different way in order to
block this conclusion. This brings me to the second possible reconstruction of Darwall’s
argument for categorical and universal morality. Above, I assumed that one is only
committed to respecting the dignity of a person if one engages in a second-person
relationship with this person. This was the reason why I claimed that it is not enough for
Darwall to claim that one cannot reject the second-person standpoint outright; he has to
show that taking up the second-person standpoint is inescapable in specific instances as
well.
At some point, however, Darwall seems to suggest that one is not just committed to
respecting the dignity of the person(s) with whom one actually engages in a second person
interaction, but that in taking up the second-person standpoint in a specific instance, one is
at the same time committed to the dignity of all other persons. That is, he suggests that
when I ask you to remove your foot from my foot, I am not just committed to respecting
your dignity but also the dignity of all other persons. Darwall writes that “any secondpersonal authority at all, which itself partly is a standing to hold free and rational agents
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
175
responsible, must be complemented by a second-personal authority that anyone can have
just in being second-personally competent” (Darwall 2006, 273–74).
If one is committed to respecting the dignity of all persons by taking up the secondperson standpoint, and if one cannot reject the second-person standpoint outright, Darwall
succeeds in justifying a categorical and universal principle of interpersonal morality. After
all, in that case, it would not be necessary to actually engage in a second-person relationship
with all others in order to justify a categorical and universal principle; it would be sufficient
if one were engaged in at least some second-person relationships.45 In that case, Stalin
would also be committed to a universal and categorical principle of interpersonal morality.
Darwall’s argument for the claim that by taking up the second-person standpoint one is
committed to respecting the dignity of all person is the following:
Any second-personal authority at all can exist only if it can be rationally accepted
by free and rational agents as such. But for that to be true there must be grounds
for such an acceptance, and whatever interests free and rational agents have as
such would have to be among such grounds. It is conceptually necessary,
moreover, that free and rational agents have an interest in not being subject to
others’ arbitrary will since that would, by definition, interfere with the exercise of
their free and rational agency. As this interest must be among the grounds that
free and rational agents have for accepting any authoritative demands at all, it
necessarily supports a demand, as free and rational, against being subject to
demands that cannot be so justified. It follows, again, that second-personally
competent agents have the authority to demand that they not be subject to mere
impositions of will, that is, to demanding (coercive) conduct that cannot be
justified second-personally (Darwall 2006, 274).
I am not sure whether I completely understand this argument, and unfortunately Darwall
says very little in support of it. But insofar as I do understand this argument, it seems to
consist of the following steps:
45
Darwall’s idea thus seems to be that the moment one enters the moral domain, there is no way to (rationally)
escape the demands of morality, which, according to Darwall, include respect for the dignity of all persons. I think
that there is something deeply counterintuitive about the idea that once I have asked a person to remove her foot
from my foot, I am necessarily committed to the equal dignity of all persons. In the text, I will criticize Darwall’s
argument for this conclusion.
CHAPTER 5
176
1) Accepting the second-personal authority of, for instance, the person who stands
on your foot means that you should conceive of the person who stands on your
foot not just as that particular person, but (also) as a free and rational agent, i.e. as
someone who can respond to reasons.46
2) This can only be true if there are interests that a free rational agent has simply
insofar as he or she is a free and rational agent.
3) One of these interests is to not be subject to another person’s arbitrary will, and
this, according to Darwall, supports a (legitimate) demand to not be subject to the
will of another person.
4) Anyone with second-personal competence has the authority to demand that they
will not be subject to another person’s arbitrary will, i.e. that anyone with secondpersonal competence has dignity.
Darwall thus tries to show that having second person capacities or competencies is
sufficient to have equal dignity. He repeats this point in more recent writing:
It is sufficient to be a moral agent in this sense [i.e. as an equal member of the
moral community] that a being has the psychic capacities necessary to enter into
relations of mutual accountability, that is, to take a second-person perspective on
himself and others and regulate his conduct from this point of view (Darwall 2010,
221–22 my emphasis).
Thus, one is not only committed to respecting the persons with whom one actually engages
in second-personal interaction but also any other person who has the competences or
capacities to engage in this type of interaction.47 Darwall concludes that
whenever we take up the second-person standpoint at all and address any claim or
demand whatsoever, we are committed to the shared second-personal authority of
any second-personally competent being (moral agent or person) on which the
ideas of moral obligation and rights depend (Darwall 2010, 224 my emphasis).
46
Darwall writes that “the addressee is ... conceived of as a-person-who-happens-to-stand-in-that-normativerelation” (Darwall 2006, 268).
47
Cf. “second-personal reasons are no less valid for an agent who happens to reject them but who has the capacity
to take up a second-person perspective and accept them” (Darwall 2007, 59 my emphasis).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
177
Darwall thus holds that we have to respect the equal dignity of any person with secondpersonal competencies or capacities, i.e. any person who is able to have reactive attitudes.
The problem with this argument, however, is that I fail to see how this conclusion follows
from an analysis of the second-person standpoint. The reason for this is that there is an
unexplained gap between the following:
1) The claim that insofar as one engages in a second-person relationship one has to
assume that this person is as a free and rational agent.
2) The claim that having second person competencies is sufficient for deserving
moral respect.
Even if it is true that, as Darwall suggests, there are certain interests that persons have
simply by virtue of being a free and rational agent, this does not yet show that I have to
respect these interests independent of my being involved in any second-person relationship
with these persons. For instance, even if I accept that the person who stands on my foot has
an interest in not being subject to my will, at least insofar as I engage in a second-personal
relation with this person, it does not follow that all persons with the same competences can
make the same (legitimate) demand, nor that I consequently have to respect their dignity,
because it is not necessary that I actually have a second-personal relation with them. It does
not necessarily follow from the claim that I have to respect the interest of the person who
stands on my foot as an interest of her as a rational and free agent that I also have to respect
the same interest of all other persons with the same competences. In other words, I fail to
see why having second-personal competencies is a sufficient condition of deserving moral
respect. Or at least I fail to see how this conclusion is supposed to follow from Darwall’s
analysis of the second-person standpoint.
In addition, the reliance on the sufficiency of having certain competences for having
equal moral status seems to be in tension with Darwall’s own argument, or at least signifies
a shift in the argument. Recall that (the assumption of) second person competence was
originally understood as a felicity condition of the second-person standpoint, i.e. if one is
involved in a second-person relationship, one has to presuppose that both the addressor
and the addressee have the competences to respond to reasons. But now this competence is
used to determine the scope of the moral community. The moral importance of these
competences thus no longer seems to follow from the argument, but now seems to precede
the argument. As it stands, Darwall’s argument thus presupposes rather than argues for the
CHAPTER 5
178
conclusion that any person with certain capacities should be respected as an equal moral
person.
Let me briefly summarize my argument in this subsection. Darwall tries to show that a
categorical and universal principle is among the felicity conditions of the second-person
standpoint. I have tried to show that although Darwall might be right that there are certain
normative felicity conditions involved in taking up the second-person standpoint, he fails
to show that these normative felicity conditions entail a categorical and universal principle
of interpersonal morality. The reason for this is that it seems optional when and towards
whom one takes up the second-person standpoint (as Darwall himself also reluctantly
acknowledges), and that although Darwall is right that human beings cannot reject the
second-person standpoint outright, it remains unclear why taking up the second-person
standpoint towards particular others necessarily commits one to respecting the equal
dignity of all persons with certain capacities. In other words, just like the arguments of
Apel, O’Neill and Korsgaard, Darwall’s argument from the second-person standpoint
cannot escape the general dilemma for Kantian constructivism.
5. What is Wrong with Transcendental Arguments from the Second
Person?
In the previous section, I have critically discussed four different arguments from the
second-person standpoint. I should stress again that these discussions were insufficient to
do full justice to these rich and complex accounts. However, I do think that the discussion
in the previous section is sufficient to bring to light a problem that arguments from the
second person have in common, despite their differences.
The problem, as should have become clear by now, is that these arguments are
susceptible to one or both horns of the general dilemma for Kantian constructivism. Either
some form of interaction is truly inescapable, but in that case it does not seem to commit us
to a universal principle of interpersonal morality, or a form of interaction has a (universal)
principle of interpersonal morality among its necessary conditions of possibility – but in
that case the social practice is merely optional and the resulting principles are therefore not
categorically binding.
The discussion has shown that one of the strengths of transcendental arguments from
the second person is at the same time its weakness. The strength of a transcendental
argument from the second person is that it starts from an interpersonal perspective, which
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
179
means that there is no need to bridge the gap between what Korsgaard calls private reasons
and public reasons, or between reasons that have normative force for me and reasons that
also have normative force for or towards others. Its weakness, however, is that the gap
between private and public reasons is simply replaced by different gaps: the gap between
Wittgensteinian public reasons and Kantian public reasons; between reasons that should at
least be communicable to others and reasons that all other agents should adopt as a
principle of action; between recognizing the dignity of the participants in a particular
second person relation and recognizing the dignity of all agents. The weakness is thus that
it simply seems unconvincing that the necessary conditions of the possibility of
argumentation, communication, reasoning, shared action or second person interaction
include a universal principle of interpersonal morality, or at least that the authors discussed
above have failed to convincingly argue for this conclusion.
Instead, what happens is that the reader is forced into a false dilemma. According to
Apel, the choice is either “paranoid-autistic loss of self” (1975, 268) or that one has to
respect the universal language community. According to O’Neill, one is either “left in
solitary and thoughtless silence” (1986, 539) or one has to accept the principle of
followability. According to Korsgaard, one’s life is either “a kind of war, or combat” (2009,
194) or one has to accept the value of the humanity of all agents. These dilemmas neglect
the fact that there might be all kinds of ways of arguing, communicating and reasoning that
do not require (moral respect for) a universalist audience. I might argue, communicate or
reason with person X, and satisfy the relevant second-personal conditions, without being
obligated to have to engage in this way with person Y. In other words, it is unconvincing
that argumentation, communication, reasoning, shared action or second person interaction
completely break down if one does not accept a universal principle of morality. So insofar as
transcendental arguments from the second person aim to justify universal principles of
interpersonal morality on the basis of the necessary conditions of the possibility of some
form of interaction, I am sceptical about their success.
Another thing to note is the role of the appeal to capacities and competences. As I have
shown above, some authors, Apel, O’Neill and Darwall in particular, ultimately refer to
agential or second-personal capacities to justify the claim that all human beings, or at least
all human beings with these capacities, have equal moral standing. It is, however, unclear
how this appeal could work within the structure of a transcendental argument from the
second person. The problem is that it is unclear how a principle which states that all
individuals with agential capacities have equal moral status could follow from an analysis of
CHAPTER 5
180
the necessary conditions of the possibility of argumentation, communication or secondpersonal interaction so that it follows that A is necessarily committed to this principle
insofar as he or she engages in certain forms of interaction. Instead, this principle seems to
be simply presupposed instead of argued for.48
In other words, what is doing the relevant normative work to justify a universal
principle of interpersonal morality is not the analysis of the necessary conditions of the
possibility of some form of interaction but the idea that any individual with certain
capacities deserves moral respect. In this sense, transcendental arguments from the second
person fail to justify a categorical and universal principle of interpersonal morality.
Although this discussion does not show that it is impossible to justify a categorical and
universal principle of interpersonal morality in terms of the necessary preconditions of
interaction, the failure of several influential accounts to provide a convincing argument for
this conclusion gives rise to legitimate scepticism about the prospects of this kind of
transcendental argument in ethics. In addition, it does not seem to be a coincidence that
several authors ultimately rely on the principle that all individuals with certain capacities
deserve equal moral respect.
The question is thus whether transcendental arguments from the first person can
succeed where transcendental arguments from the second person fail. This will be the topic
of the next two chapters.
6. Conclusion
In this chapter, I have analysed four different transcendental arguments from the second
person for categorical and universal interpersonal morality. I have argued that these
arguments, despite their differences, are all susceptible to the general dilemma for Kantian
constructivism: they cannot show that there is an inescapable form of interaction which at
the same time has a universal moral principle among its necessary conditions of possibility.
This shows that, at best, current transcendental arguments from the second person are
incomplete and, at worst, that these arguments simply fail to justify categorical and
universal principles of categorical morality.
48
Not surprisingly, some critics of the authors discussed above therefore claim that these arguments can only lead
to categorical and universal principles of interpersonal morality by relying on an independent, realist value of
certain (agential) capacities (see e.g. Langton 2007; Barry 2013; Gilabert 2005b; 2005a; 2006).
TRANSCENDENTAL ARGUMENTS FROM THE SECOND PERSON
181
This suggests that the viability of transcendental arguments in ethics (at least the extent
to which they aim to justify categorical and universal principles of interpersonal morality)
depends on the success of transcendental arguments from the first person. In the next
chapter, I will therefore analyse the possibility of justifying categorical and universal
morality on the basis of a transcendental argument from the first person.
Chapter 6
Transcendental Arguments from the First
Person I: The Universal Value of Humanity
1. Introduction
In this chapter and the next, I discuss transcendental arguments from the first person for a
categorical and universal principle of interpersonal morality. The main advantage of a
transcendental argument from the first person for interpersonal morality, compared to an
argument from the second person, is that it tries to ground interpersonal moral normativity
in the first-person perspective of an agent, which seems to be less dependent on the actual
involvement of individuals in specific forms of interaction. That is, one’s selfunderstanding as an agent is at least prima facie less optional than one’s engagement in
certain types of interaction, or at least less optional than the types of interaction which have
a (universal) moral principle among their necessary conditions of possibility. A successful
transcendental argument from the first person might thus be able to show that all (rational)
agents necessarily have to accept a universal principle of interpersonal morality,
independent of their contingent desires, but also independent of their contingent
engagement in certain kinds of practices or interaction.
However, transcendental arguments from the first person are themselves far from
uncontroversial. In the previous chapter, I have already briefly discussed two common
objections to arguments from the first person: the objection that the argument is logically
flawed and the objection that it provides the wrong kind of reason for morality. Now that
we have seen that transcendental arguments from the second person fail, we should
evaluate the plausibility of transcendental arguments from the first person and assess to
CHAPTER 6
184
what extent they are susceptible to these two objections.
In the previous chapter, I have criticized Korsgaard’s argument from the publicity of
reasons and her argument from the necessary conditions of the possibility of shared action
as attempts to justify interpersonal morality (what I have referred to as the second step of
her argument). In this chapter, I analyse whether it is possible to provide another argument
from the first person for the claim that humanity has universal value. In other words, I
discuss the possibility that there is another way in which to universalize the first step of
Korsgaard’s argument, i.e. the claim that an agent, A, necessarily has to value his or his or
her (own) humanity. I draw on an early paper by Korsgaard, ‘Kant’s Formula of Humanity’
(1986), and on more recent work by Korsgaard (forthcoming) in which she does not rely on
her argument from the publicity of reasons but in which she argues for the claim that one
must value the humanity of all persons on the basis of a transcendental argument from the
first person, what I will call ‘the argument from the sufficiency of agency’.1
I argue that, just like her argument from the publicity of reasons, Korsgaard’s
transcendental argument from the first person also fails to justify a categorical and
universal principle of interpersonal morality. This failure, however, has less to do with the
argument from the sufficiency of agency or the first-personal nature of her argument and
more to do with the details of the first step of her argument (the argument for the value of
your own humanity). On one interpretation of Korsgaard’s argument for the value of your
(own) humanity, the starting point of the argument is escapable (i.e. it is susceptible to the
escapability objection). On another interpretation of her argument, it is unclear how the
value of humanity could generate any normative constraints on action, including norms of
interpersonal morality (i.e. it is susceptible to the no-normativity objection, i.e. the
objection that no normative conclusions follow from a transcendental argument). In other
words, Korsgaard’s argument is, on both interpretations of it, susceptible to one (but
different) of the horns of the general dilemma for Kantian constructivism.
In the next chapter, I will discuss Alan Gewirth’s transcendental argument from the
first person in Reason and Morality. Gewirth also uses the argument from the sufficiency of
agency to argue for a universal principle of interpersonal morality. I will argue that, in
contrast to Korsgaard, Gewirth does succeed in justifying a universal principle of
interpersonal morality through the argument from the sufficiency of agency and that this
1
I take this label from a similar argument put forward by Alan Gewirth (Gewirth 1978, 110). I will come back to
Gewirth’s argument in the next chapter.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
185
argument is not susceptible to the common objections to transcendental arguments from
the first person.
The structure of the chapter is as follows: first, I introduce Korsgaard’s argument from
the sufficiency of agency (section 2). I then critically evaluate the attempt to justify
interpersonal morality through the argument from the sufficiency of agency (section 3). I
do so by first attempting to clarify what Korsgaard means by ‘the value of humanity’ and
then distinguishing between different arguments that Korsgaard has put forward (section
3.1). I then critically evaluate two different arguments for the value of humanity and argue
that, even if we assume that the value of one’s own humanity can be successfully
universalized through the argument from the sufficiency of agency, both arguments fail to
justify a universal principle of interpersonal morality. The relevant arguments are the
regress argument (section 3.2) and the source of reasons argument (section 3.3). More
specifically, I argue that the regress argument is susceptible to one of the horns of the
general dilemma for Kantian constructivism: it can only justify the (universal) value of
humanity by relying on an escapable starting point. The source of reasons argument, on the
other hand, is susceptible to the other horn of the dilemma: although its starting point
might be inescapable, it fails to lead to any normative conclusion.
2. Korsgaard’s Argument from the Sufficiency of Agency
In the previous chapter, I have discussed Korsgaard’s criticism of transcendental arguments
from the first person on the basis of her idea that one cannot argue from ‘private’ reasons to
‘public’ reasons. Instead, she puts forward the argument from the publicity of reasons and
the argument from shared action. This is the approach that Korsgaard takes in her much
discussed work The Sources of Normativity. However, both in her earlier work (1986) and
in her more recent work (forthcoming), Korsgaard herself hints at the possibility of
providing an argument from the first person for interpersonal morality. Now that we have
seen that Korsgaard’s argument from the publicity of reason fails, it might be worth
spending some time on the pre- and post-The Sources arguments that Korsgaard makes.
In her early paper, ‘Kant’s Formula of Humanity’ (1986), Korsgaard argues that the
claim that you must value your humanity commits you to valuing the humanity of all other
rational agents at the same time. She writes that “if you view yourself as having a valueconferring status [i.e. if your humanity is valuable] in virtue of your power of rational
choice, you must view anyone who has the power of rational choice as having, in virtue of
CHAPTER 6
186
that power, a value-conferring status” (Korsgaard 1986, 196). Korsgaard thus claims that
anyone who has the power of rational choice must value his or her humanity and that this
person must also value the humanity of all other agents (i.e. all other persons with the
power of rational choice). Unfortunately, however, in this paper Korsgaard says very little
about why one must value the humanity of anyone with the power of rational choice.
In a recent, unpublished, paper, Korsgaard (forthcoming) elaborates on her earlier
suggestion. She compares the reason for being committed to the universality of humanity
with extending voting rights to all citizens of a country: if you have voting rights simply by
virtue of being a citizen, you should accept that every citizen has voting rights:
[S]uppose you ask me, “In virtue of what do you have the right to vote here?” and I
reply, “I am a citizen of this nation.” Citizenship, as I understand it, is a form of
normative standing: it gives its possessor certain normative or moral powers. You
might reply, “Well, I am a citizen too, so I have the right to vote here as well.”
Notice that it would not make sense for me to respond, “no, my own citizenship
has that normative implication – but so far as I am concerned, yours does not”
(forthcoming, 28–29).
Korsgaard claims that a similar argument could be made about the universality of the value
of humanity. The only difference is that whereas the person in the quote above has voting
rights by virtue of being a citizen, one’s humanity has value by virtue of one being an agent:
Kant’s argument for the Formula of Humanity, treats humanity, or the power of
rational choice, as if conferred a kind of normative standing on us. When we look
at the argument this way, Kant asks, “in virtue of what do we have the right to treat
our ends as good, that is, to confer normative value on them, and so in effect to
legislate values?” and he answers, “Our humanity.” So the argument assigns us a
normative standing in virtue of our humanity, like the normative standing we have
in virtue of say, being born in a certain country (Korsgaard forthcoming, 29).
Korsgaard concludes that the fact that we have a normative standing by virtue of the power
of rational choice “commits us to assigning the same standing to every other rational being”
(Korsgaard forthcoming, 31).
Below, I will come back to the question of how exactly to understand the notions of
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
187
value, humanity and normative standing, because these notions are crucial to
understanding and evaluating the details of her argument. Here I want to focus on the
structure of the argument that Korsgaard is making in order to universalize the value of
humanity. I take Korsgaard to be making the following argument: if you necessarily have to
value your humanity (i.e. your capacity for setting an end) simply by virtue of being an
agent (i.e. by virtue of having the power of rational choice, i.e by having the power to set an
end), you must also value the humanity of any other agent. The reason for this is that being
an agent is a sufficient condition of having value (just like citizenship is a sufficient
condition of having voting rights). Thus, if being an agent is a sufficient condition of having
value, any agent, and not just the particular agent you are, has value.
This argument from the sufficiency of agency clearly does not rely on the idea of the
publicity of reasons, and so it is quite different from the argument made in The Sources and
which Korsgaard elaborates on in Self-Constitution. The reason why one has to value the
humanity of any agent is not because reasons are intrinsically public or because it is a
necessary condition of the possibility of shared action, but because one’s reasons for valuing
one’s own humanity necessarily imply that one also has to value the humanity of all other
agents. It’s a distinctively first-personal argument for interpersonal morality because it
justifies interpersonal morality from the perspective of the individual agent in terms of
what the agent is necessarily committed to insofar as he or she understands him- or herself
as an agent.2
To me it is a mystery why Korsgaard in both The Sources and Self-Constitution relies on
the argument from the publicity of reasons, whereas in her earlier work and in her
forthcoming work she puts forward the argument from the sufficiency of agency, or in
other words turns to a argument from the first person strategy rather than an argument
from the second person. Korsgaard herself does not explicitly acknowledge these changes in
her work, so one can only guess about her motivations. I am not, however, interested in
exegetical questions about Korsgaard’s work. Instead, the reason why I discuss the
argument from the sufficiency of agency is as follows: now that we have seen that
Korsgaard’s argument from the publicity of reason and other arguments from the second
2
The argument from the sufficiency of agency itself is not a transcendental argument, but it is a crucial element of
a transcendental argument from the first person. The argument from the sufficiency of agency is an instantiation
of a more general logical principle which states that if X (agency) is a sufficient condition of having Y (value),
every X has Y. Gewirth calls this the ‘criterion of relevant similarities’ (Gewirth 1978, 104). I will come back to this
criterion in the next chapter.
188
CHAPTER 6
person fail, there are good reasons to ask the question whether there are transcendental
arguments from the first person for interpersonal morality. Given that Korsgaard herself in
some of her writings uses the argument from the sufficiency of agency to argue for
interpersonal morality (and given that other authors, such as Alan Gewirth, put forward a
similar argument), we should investigate to what extent the argument from the sufficiency
of agency succeeds where the argument from the publicity of reasons fails.3 In other words,
is there an argument from the first person for the universal value of humanity? In the
remainder of this chapter, I will evaluate the success of this argument in the context of
Korsgaard’s argument for the value of humanity. In the next chapter, I will discuss the
argument from the sufficiency of agency in the context of Gewirth’s moral theory.
Importantly, the success of the argument from the sufficiency of agency in Korsgaard’s
theory cannot be assessed independently of the success of the first step of her argument, i.e.
her argument which leads to the conclusion that A has to value his or her own humanity
(see chapter 4 for a discussion of the different steps of Korsgaard’s argument). Obviously, if
Korsgaard’s argument for the claim that A has to value his or her own humanity fails, there
is nothing to universalize. In addition, even if the first phase of Korsgaard’s argument
succeeds, the reason why one necessarily has to value one’s own humanity has implications
for the plausibility of its universalization through the argument from the sufficiency of
agency.4 For instance, if one has to value one’s own humanity because of a particular
3
In The Sources, Korsgaard describes transcendental arguments from the first person for interpersonal morality as
follows: “These neo-Kantian justifications characteristically begin by showing that you are rationally committed to
a certain normative conception of yourself, or to valuing certain features of yourself. They then try to move from
that conclusion to the further conclusion that you must hold the same normative conception of others, or value
the same features in them, on pain of contradiction. Since I regard my humanity as [the] source of value, I must in
the name of consistency regard your humanity that way as well. So I must value the things that you value. Or, to put
it another way since I think my humanity is what makes my desires into normative reasons, I must on pain of
contradiction suppose that the humanity of others makes their desires into normative reasons as well” (Korsgaard
1996, 133 my emphasis). This description of ‘Neo-Kantian justifications’, however, does not seem to do justice to
the argument from the sufficiency of agency. In The Sources, Korsgaard describes transcendental arguments from
the first person as follows: “Since I regard my humanity as [the] source of value, I must in the name of consistency
regard your humanity that way as well” (1996, 133). But this seems to be very uncharitable towards the argument
from the sufficiency of agency. Korsgaard’s argument from the sufficiency of agency is not that I regard my
humanity as the source of all values and that this commits me (in the name of consistency) to the value of the
humanity of others. Instead, the argument is that I necessarily have to value my own humanity, my rational
nature, simply by virtue of being an agent. Given that this is a characteristic I share with other agents, I must also
value their rational nature. This argument sounds much more plausible than Korsgaard’s description of the
argument in The Sources. This does not mean that the argument from the sufficiency of agency does indeed work,
but that there might be more to the argument than Korsgaard suggests in The Sources.
4
This is, for instance, noted by several critics of Gewirth’s reliance on the argument from the sufficiency of agency
(McMahon 1986; Kramer and Simmonds 1996; Chitty 2008). I will come back to these objections to Gewirth’s
argument below.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
189
characteristic, which is not shared by all agents, the argument from the sufficiency of
agency does not apply in the first place, because the argument from the sufficiency of
agency only works if agency, and some particular characteristic of it, is a sufficient reason
for being committed to the value of one’s own humanity.
In what follows, I will therefore critically evaluate the first phase of Korsgaard’s
argument and subsequently raise the question of whether the conclusion of this argument
can be successfully universalized through the argument from the sufficiency of agency. I
will argue that even if we assume that the argument from the sufficiency of agency is valid,
Korsgaard fails to justify any principles of interpersonal morality.
3. The Argument for the (Universal) Value of Humanity
3.1. Which Humanity? What Value? Which Argument?
In chapter 4, I have already briefly outlined the main structure of the first phase of
Korsgaard’s argument. Korsgaard tries to show that any agent is necessarily committed to
valuing his or her humanity simply by virtue of understanding him- or herself as a
reflective being that needs reasons for action. This argument can be roughly summarized as
follows:
1) Agency (X) is inescapable for A.
2) Valuing your own humanity (Y) is a necessary condition of the possibility of
agency (X).
3) Therefore, A must value his or her (own) humanity (Y).
In what follows, I will discuss this argument in more detail and assess to what extent it is
convincing.
Before discussing the argument, however, let me briefly say a few words about what it
means to say that one must ‘value humanity’. This is important because both the concept of
value and the concept of humanity could be understood in many different ways.
First, what does it mean to say that one must value (your own) humanity?
Unfortunately, Korsgaard is not very clear about how she understands this concept. The
main problem is that she interchangeably uses humanity as some sort of technical term to
refer to certain rational capacities and as some sort of thick moral concept which is
CHAPTER 6
190
supposed to be close to our intuitive idea of humanity, including, for instance, a respect for
human rights.
In ‘Kant’s Formula of Humanity’ (1986), Korsgaard defines humanity in terms of “the
capacity for setting an end” (1986, 187), which implies that valuing humanity means
‘valuing the capacity for setting an end’. In addition, she suggests that the value of humanity
is simply a redescription of Kant’s formula of humanity of the categorical imperative, i.e.
the formula of treating “rational nature or humanity as an end in itself” (Korsgaard 1986,
183). 5 Throughout The Sources, Korsgaard refers interchangeably to the “value of
humanity”, our value or our identity as “human beings” (Korsgaard 1996, 118, 121), and
our “moral identity” (Korsgaard 1996, 121). In addition, she claims that “to value yourself
just as a human being is to have moral identity, as the Enlightenment understood it. So this
puts you in moral territory” (Korsgaard 1996, 121). Finally, and most substantively,
Korsgaard claims that “it follows from this argument [the argument for the value of
humanity] that human beings are valuable. Enlightenment morality is true” (Korsgaard
1996, 123). It is questionable whether these claims are really all equivalent. Showing that we
must value the capacity for setting an end or showing that we must value our identity as
human beings arguably are not necessarily the same as showing that enlightenment
morality is true.
A lot here depends on how exactly one understands this capacity for setting ends. I will
come back to this below when I discuss Korsgaard’s argument for the value of humanity in
more detail. The reason for this is that depending on how exactly one understands
Korsgaard’s argument, the meaning of ‘humanity’ might change (after all, the conclusion of
her argument will contain nothing more than can be found in the premises of the
argument). For now, I will understand humanity as roughly referring to our rational
capacity to set ends. Although this needs to be specified in order to really know what the
argument proves, this abstract definition should be sufficient to be able to proceed.
Second, what does it mean to say that one must value humanity? Just like ‘humanity’,
‘value’ can be understood in different ways. Unfortunately, Korsgaard is also not entirely
clear about what she means by value. She is, however, very explicit about what she does not
mean when she talks about the value of humanity. Korsgaard rejects realism regarding
value, according to which “to value something is to respond appropriately to the fact that it
5
For a criticism of Korsgaard’s value-based reading of the formula of humanity of the categorical imperative, see
Nyholm (2016)
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
191
has value” (Korsgaard forthcoming, 4). Instead, Korsgaard claims that something has value
only if we (have to) value it. So what does it mean to say that something has value on this
constructivist view?
In the prologue of The Sources, Korsgaard describes the general idea of values as
follows: “It is the most striking fact about human life that we have values. We think of ways
that things could be better, more perfect, and so of course different, than they are; and of
ways that we ourselves could be better, more perfect, and so of course different, than we
are” (Korsgaard 1996, 1). For instance, if I value equality, I might think the world could or
should be organized much better than it currently is.
Korsgaard, however, seems to understand the value of humanity in different terms. For
Korsgaard the value of humanity refers, first and foremost, to a normative status or moral
standing (Korsgaard 1996, 145; Korsgaard forthcoming, 25–32). She writes, for instance, “I
force you to acknowledge the value of my humanity, and I obligate you to act in a way that
respects it” (Korsgaard 1996, 143). And more recently, she writes that to say that humanity
has value means “that we should respect the normative standing people have in virtue of
their humanity” (Korsgaard forthcoming, 28). Korsgaard contrasts this idea of the value of
humanity as conferring status or a normative standing with a view according to which we
value a certain property, such as our rational nature (Korsgaard forthcoming, 26). On the
property interpretation of value, “we care about having that property, or think it worth
preserving, or wish to develop it in appropriate ways, and so forth” (Korsgaard
forthcoming, 26).
To say that humanity has value, on the status interpretation of value which Korsgaard
accepts, is thus to say that we should respect the normative standing of people (compare
this with saying that inequality has value or with the property interpretation of value). The
value of humanity is thus not something that should be promoted or brought about, but
something which deserves respect. On this view, the value of humanity does not directly
entail norms or duties, but only indirectly insofar as there might be things one should or
not should do to oneself and to others in terms of respecting their and one’s own normative
standing. Writing in the context of duties to oneself, Korsgaard writes that “duties are
expressions of respect, not for the property of rationality, but for the legislative standing
that it confers upon us” (Korsgaard forthcoming, 31–32).
The relation between agency, value, normative standing and moral norms is thus
highly complex, and this should be taken in mind when evaluating Korsgaard’s argument.
Much more can and needs to be said about the notion of value in Korsgaard’s work. For
CHAPTER 6
192
now, I mainly want to stress that Korsgaard uses a particular notion of value which does
not necessarily correspond to the common usage of this notion (and this also leads to much
confusion about her argument). I will come back to the notion of value, as well as to the
notion of humanity, when I discuss Korsgaard’s argument for the value of humanity. Again,
how exactly the concept of value should be understood also depends on how one
understands the argument for the value of humanity.
With these preliminary remarks in mind, let me now turn to Korsgaard’s argument for
the claim that A must value his or her (own) humanity. In The Sources, this argument is put
forward by Korsgaard as a “fancy new model of an argument that first appeared in a much
simpler form, Kant’s argument for his Formula of Humanity” (1996, 122). The argument
tries to show that the value of humanity is a necessary condition of having a reason for
action or for acting for any end (which I take to be roughly equivalent), i.e. that in acting we
are “taking things to be important … to us” (1996, 122).
The first problem we face is that Korsgaard’s argument is open to different
interpretations, even if we have a better grasp of the respective notions of humanity and
value. So before critically analysing Korsgaard’s argument for humanity, it’s necessary to
indicate which argument I will focus on. Following Robert Stern (2011), we can distinguish
between two different transcendental arguments that Korsgaard gives for the value of
humanity, both of which I discuss in more detail below.6
The first argument is often referred to as the ‘regress of identities argument’ (Stern
2011, 86; for interpretations of Korsgaard’s argument as a regress argument see e.g. Street
2012, 49; Kerstein 2001, 42–52; Sussman 2003) (referred to from now on as the regress
argument). The regress argument states that we can or should always ask ‘why-questions’
about our actions and also about our reasons for action (‘Why do I want to drink coffee?’
‘Because I want to be more focused.’ ‘Why do I want to be more focused?’ ‘Because I want
to finish this chapter.’ ‘Why do I want to finish this chapter?’ etc.). Korsgaard claims that
6
A third candidate is Korsgaard’s reconstruction of Kant’s argument for the value of humanity in her paper
‘Kant’s Formula of Humanity’ (1986). In this argument, the value of humanity is understood to be a necessary
condition of the possibility of acting for an objectively good end (Korsgaard 1986, 190). Korsgaard writes that “if
humanity is not regarded and treated as unconditionally good then nothing else can be objectively good”
(Korsgaard 1986, 198). This argument presupposes that there is such a thing as a categorical imperative. This
means that the argument is conditional: if there is a categorical imperative, then one must value humanity.
Korsgaard describes this argument and the relation between the categorical imperative, the objective good and the
value of humanity as follows: “having established that if there is a categorical imperative there must be something
that is unconditionally valuable, Kant proceeds to argue that it must be humanity” (Korsgaard 1986, 194). Because
the argument is contingent on already accepting the categorical imperative, I will not discuss this argument further
and instead will focus on two more ambitious arguments, which do not assume the categorical imperative.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
193
this regress of why-questions can only be brought to an end by the value of humanity.
I will call the second interpretation ‘the source of reasons argument’. In this argument,
the value of humanity is not justified because it would bring an end to a regress of ‘whyquestions’, but because, according to Korsgaard, humanity is “the source of all reasons and
values” (Korsgaard 1996, 122). This argument is harder to understand than the regress of
reasons argument, but the general idea, which I elaborate on below, is that one can only
think that a particular end is valuable if one values oneself as the source of all value. The
argument presupposes that realism about values should be rejected and subsequently
locates the source of value in the agent him- or herself.
These brief descriptions of the two different arguments for the value of humanity are
meant to illustrate the complexity and ambiguity of Korsgaard’s theory, especially because
Korsgaard herself considers these arguments to be identical (Korsgaard 2009, 76).7 In what
follows, I will critically discuss both the regress of identities argument and the source of
reasons argument in more detail. In my discussion of Korsgaard’s arguments, I will draw
on Robert Stern’s (2011) recent reconstruction of these two arguments. The reason for this
is twofold. First, as I have already noted above, Korsgaard’s own argument is somewhat
ambiguous and Stern has put forward the most systematic reconstruction of Korsgaard’s
arguments that I am aware of. Second, Stern is one of the few people who actually thinks
that (one of) Korsgaard’s argument(s) can be made to work. He writes the following:
“rather to my surprise, and rather against the run of the critical literature on Korsgaard’s
book, I will suggest that in one of its forms, the argument can be made to work” (Stern
2011, 74).8 More specifically, Stern claims that whereas the regress of identities argument
fails, the source of reasons argument can work. I will now discuss both arguments in turn.
7
One might argue that Korsgaard’s argument about self-constitution in Self-Constitution is different from the
arguments I mention below. The relation between The Sources and Self-Constitution is complicated, and in this
section I only focus on Korsgaard’s argument before and in The Sources. Besides reasons of space there are two
additional reasons for this restriction. First, in Self-Constitution, Korsgaard refers back to and even summarizes
her transcendental argument in The Sources (Korsgaard 2009, 20–25). Second, she remarks, “I am always making
the same argument” (Korsgaard 2009, 76).
8
Or at least Stern claims that the argument works on its own terms. According to Stern, Korsgaard’s argument
hinges on the rejection of realism, whereas Stern, in other writings, defends realism (See e.g. Stern 2012a; Stern
2013b). I come back to this point below. Furthermore, Stern does not discuss the universalization step of
Korsgaard’s argument (Stern 2011, 77–78). My discussion of the source of reasons argument could be understood
as a discussion of the relation between the argument for humanity (what I call the first step of the argument) and
the argument for the universality of humanity (what I call the second step of the argument.)
CHAPTER 6
194
3.2. The Regress of Identities Argument
The regress of identities argument goes as follows: Korsgaard starts from the now familiar
claim that we have to act on the basis of reasons:
[reflective consciousness] sets us a problem no other animal has. It is the problem
of the normative. For our capacity to turn our attention on to our own mental
activities is also a capacity to distance ourselves from them, and to call them into
question. I perceive, and I find myself with a powerful impulse to believe. But I
back up and bring that impulse into view and then I have a certain distance. Now
the impulse doesn't dominate me and now I have a problem. Shall I believe? Is this
perception really a reason to believe? I desire and I find myself with a powerful
impulse to act. But I back up and bring that impulse into view and then I have a
certain distance. Now the impulse doesn't dominate me and now I have a problem.
Shall I act? Is this desire really a reason to act? The reflective mind cannot settle for
perception and desire, not just as such. It needs a reason. Otherwise, at least as
long as it reflects, it cannot commit itself or go forward (Korsgaard 1996, 93).
For the sake of argument, I take the inescapability of agency, understood in terms of
needing a reason for action, to be an uncontroversial starting point for a transcendental
argument:
1) Agency is inescapable for A.
Second, Korsgaard introduces the idea of practical identities. A practical identity is “a
description under which you value yourself, a description under which you find your life to
be worth living and your actions to be worth undertaking” (Korsgaard 1996, 101).
Examples of practical identities are one’s identity of being a student or a mother. According
to Korsgaard, reasons spring from these practical identities: “we endorse or reject our
impulses by determining whether they are consistent with the ways in which we identify
ourselves” (Korsgaard 1996, 122). For instance, if one understands oneself as a student one
has a reason to go to lectures, to read books and to respect standards of academic integrity
(or to be loud and to drink too much). This gives us the second premise:
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
195
2) Reasons spring from one’s practical identities.
Korsgaard stresses that most, if not all, of our practical identities are essentially contingent,
even those identities we are born into (Korsgaard 1996, 120; Korsgaard 2009, 23). One
might stop caring about being a student or even about being a mother, and different
practical identities might conflict with each other. Crucially, Korsgaard argues for the
contingency of our particular practical identities by relying on the idea that we have a
reflective consciousness so that we can always ask why we should accept a particular
practical identity (just like we can always ask whether a particular desire is really a reason to
act). She writes that a human being is someone “who arrives at his reasons through
reflection” (Korsgaard 1996, 257) and, crucially, that “we should never stop reflecting until
we have reached a satisfactory answer; one that admits of no further questioning ... we
should seek the unconditioned” (Korsgaard 1996, 258). So although Korsgaard might grant
that as a matter of fact not everyone actually questions his or her practical identities, our
reflective self-consciousness means that we always “may” (Korsgaard 1996, 122) or even
“should” (Korsgaard 1996, 258) do so (I come back to these two different claims – may
versus should – below).
In a recent summary of her practical identity argument in Self-Constitution (2009),
Korsgaard describes the contingency of practical identities as follows:
[T]he sorts of [practical] identities I am talking about remain contingent in this
sense: whether you treat them as a source of reasons and obligations is up to you.
If you continue to endorse the reasons the identity presents to you, and observe
the obligations it imposes on you, then it’s you. Leaving morality aside for the
moment—because there may be moral reasons for not doing the things I am about
to describe—you can walk out even on a factually grounded identity like being a
certain person’s child or a certain nation’s citizen, dismissing the reasons and
obligations that it gives rise to, because you just don’t identify yourself with that
role (2009, 23).
Korsgaard’s point is thus that although reasons spring from our practical identities, we may
or even should always ask whether we have reason to endorse these practical identities.
This, then, is the third premise of the argument:
CHAPTER 6
196
3) Practical identities are contingent; we may or should always ask for a reason to
adopt a practical identity (regress of reasons).
From here, the argument, as I understand it, proceeds as follows. Fourth, Korsgaard rejects
that the reasons to adopt a practical identity can be found in realist, i.e. mind-independent,
moral facts. Describing Kant’s argument for the value of humanity, which is supposed to
make the same point, and of which Korsgaard’s argument is supposed to be “just a fancy
new model” (Korsgaard 1996, 122), she writes: “[Kant] asked what it is that makes these
objects [i.e. the end of actions] good, and, rejecting one form of realism, he decided that the
goodness was not in the objects themselves” (Korsgaard 1996, 122). So
4) There are no moral facts which can stop the regress of reasons.
Fifth, Korsgaard proposes that only the value of humanity can bring the regress of whyquestions to an end. The reason for this is as follows: “what is not contingent is that you
must be governed by some conception of your identity” (Korsgaard 1996, 120). After all, if
you have to act on a reason, and if reasons spring from our practical identities, it is
necessary to have a practical identity, even if our particular practical identities are
contingent.
5) One can only have a reason for action if one adopts a practical identity; adopting a
practical identity is a necessary condition of the possibility of having a reason for
action.
Finally, Korsgaard claims that the necessity of having a practical identity commits one to
valuing one’s own humanity:
But this reason for conforming to your particular practical identities is not a
reason that springs from one of those particular practical identities. It is a reason
that springs from your humanity itself, from your identity simply as a human
being, a reflective animal who needs reasons to act and to live. And so it is a reason
you have only if you treat your humanity as a practical, normative, form of
identity, that is, if you value yourself as a human being (Korsgaard 1996, 121).
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
197
Thus, according to Korsgaard, you can only have a reason to adopt a practical identity if
you value yourself as a human being, i.e. if you value your (own) humanity:
6) You must value your (own) humanity if you are to have a reason for action.
This concludes the brief reconstruction of Korsgaard’s regress argument for the value of
your own humanity. According to Korsgaard, one has to act for a reason, but at the same
time our reflective self-consciousness seems to lead to an infinite regress because one could
always ask a why-question about any reason for action, and because reasons for actions
spring from one’s practical identities one could also always ask a why-question about these
practical identities. Because it is necessary to act for a reason, and because Korsgaard rejects
moral realism, she concludes that the only thing that can stop the regress is the idea that
one needs a practical identity (whatever it might be) because reasons for actions spring
from practical identities. This, according to Korsgaard, amounts to saying that one must
value one’s humanity. Let me now turn to Stern’s criticism of this argument, which applies
to Korsgaard’s claim that our particular practical identities are necessarily contingent.
Stern objects to Korsgaard’s claim that we should always doubt our particular practical
identities, and he therefore concludes that certain particular practical identities could also
stop the regress of why-questions (Stern 2011, 87–88).9 If this is correct, it is not necessary
to value one’s humanity because one could instead just value a particular identity in order
to stop the regress. He puts forward the following example to make his point:
[S]uppose that I am a loving son. I am certainly conscious that I might not have
been, had I been raised in a different way or in a different time or place, so it is a
contingent identity in that sense. But suppose someone offers me an alternative
identity. Unless they can supply me with reasons to think being a loving son is
wrong or mistaken in some way, why should I be brought to doubt my identity?
But it might seem that the only reasons that would count as reasons to give up that
9
Kerstein objects to Korsgaard for similar, but slightly different, reasons. He also claims that Korsgaard has not
given the reader any argument for why one necessarily has to accept that one has a reason to have a practical
identity to account for the normativity of one’s particular practical identities. His objection, however, is that he
fails to see how the claim that one necessarily has a practical identity can generate a reason to choose one
particular practical identity over any other possible practical identity (Kerstein 2001, 48–51).
CHAPTER 6
198
identity are ones that would only do so once that identity has been given up (Stern
2011, 87).
Stern’s suggestion is thus that there might be particular practical identities that are
fundamental to one’s identity. In this case, Stern argues that it is not obvious that one
should doubt this identity. And he adds that “it seems that you would need to give me some
internal grounds for giving up my identity (for example, that my loving regard for my
mother is making her life worse and not better), which even in the case of an identity which
is contingent may not be forthcoming” (Stern 2011, 87). I think that Stern is right to point
to a gap in Korsgaard’s argument. The gap is that Korsgaard does not argue for the claim
that we should always question our identities (premise 3). Stern’s example forces Korsgaard
to argue for this claim because, indeed, it is not obvious that the loving son should (or even
could) question his identity in the absence of any good reason for doing so without losing
his very identity.
I think the problem is in fact more serious than Stern recognizes. I think that the
problem is not merely that Korsgaard has “not done enough to show that it is only the
identity of humanity that is invulnerable to the regress issue” (Stern 2011, 88), but that
Korsgaard cannot, on her own premises, show that one should always question one’s
particular practical identities. When Korsgaard discusses the contingency of our practical
identities, she shifts between claiming that we may always doubt our practical identities
identity (Korsgaard 1996, 122) to the much stronger claim that we should do so and that we
therefore need a reason to adopt a specific practical (Korsgaard 1996, 258). Now, she clearly
needs the latter claim to block the possibility that a particular practical identity can block
the regress of why-question. After all, if we only ‘may’ or ‘could’ doubt some particular
practical identity, it is not clear that we necessarily need a reason to adopt a particular
practical identity and that we therefore necessarily have to value our (own) humanity. In
that case, one only has to value one’s own humanity insofar as one actually doubts all of the
particular practical identities one has.
The problem, however, is that only the weaker claim, i.e. the claim that we may doubt
our practical identities, follows from Korsgaard’s argument. The claim that one may doubt
one’s practical identities follows from Korsgaard’s argument because if we assume that we
have a reflective consciousness, this implies that we could question our practical identities,
just as we could raise the question of whether a desire really is a reason to act. It does not
follow, however, from the fact that we could always question our practical identities, that
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
199
we should (always) do so. The imperative to never stop reflecting until one has reached
bedrock simply does not follow merely from the reflective nature of our consciousness. In
fact, the claim that one should always question one’s practical identities seems to
presuppose an unvindicated and controversial value commitment. It is unvindicated
because it is simply presupposed instead of argued for. It is controversial because it seems
implausible that this kind hyper-reflective life is the best or only way to live one’s life (cf.
Cassam 2014, 82). Thus, insofar as Korsgaard aims to argue for the value of humanity
simply on the basis of the idea that we need reasons to act, it fails. Only the hyper-reflective
agent, who never stops reflecting, would have to value his or her humanity. Korsgaard,
however, does not have any resources to claim that human life should be like the life of a
hyper-reflective agent. In other words, the starting point of this interpretation of
Korsgaard’s argument is escapable because it only applies to hyper-reflective agents and not
to all agents.
Let me end this discussion of the regress argument by briefly relating it to the argument
from the sufficiency of agency. The argument from the sufficiency of agency states that if
one has to value one’s humanity simply by virtue of being an agent, one also has to value
the humanity in all other agents. However, applying the argument from the sufficiency of
agency to the conclusion of the regress argument obviously does not lead to any categorical
and universal principles of interpersonal morality. The problem is that if my criticism is
correct, the regress argument does not show that I have to value humanity simply by virtue
of being an agent, but only by being a hyper-reflective agent. The most the argument could
show is that insofar as I am a hyper-reflective agent I do not just have to value my own
humanity, but also the humanity of all other hyper-reflective agents. In other words, even if
the argument from the sufficiency of agency is valid, the value of your humanity (on the
regress interpretation of the argument) cannot lead to universal and categorical principles
of interpersonal morality.
3.3. The Source of Reasons Argument
The failure of the regress argument is often taken as an indication that Korsgaard fails to
justify the value of (your own) humanity (Kerstein 2001; Street 2012). Stern, however, puts
forward a reconstruction of a different line of argument from The Sources, which, according
CHAPTER 6
200
to Stern, is valid.10 In this section, I discuss Stern’s reconstruction of Korsgaard’s argument,
which I call ‘the source of reasons argument’, and argue that even if the argument is valid,
its universalization cannot lead to any principle of interpersonal morality. The reason for
this is that, on this interpretation of the argument, it cannot lead to any normative
conclusion. In other words, I argue that the source of reasons argument is susceptible to a
variant of the no-normativity objection.
Let me start by laying out the argument. Stern takes his cue from Korsgaard’s
description of the argument for the value of humanity which she finds in Kant:
[Kant] started from the fact that when we make a choice we must regard its object
as good. His point is the one I have been making—that being human we must
endorse our impulses before we can act on them. He asked what it is that makes
these objects good, and, rejecting one form of realism, he decided that the
goodness was not in the objects themselves. Were it not for our desires and
inclinations—and for the various physiological, psychological, and social
conditions which gave rise to those desires and inclinations—we would not find
their objects good. Kant saw that we take things to be important because they are
important to us—and he concluded that we must therefore take ourselves to be
important. In this way, the value of humanity itself is implicit in every human
choice. If complete normative scepticism is to be avoided—if there is such a thing
as a reason for action—then humanity, as the source of all reasons and values,
must be valued for its own sake (Korsgaard 1996, 122).11
Stern takes the central idea of this argument to be the following: “as long as we think we can
act for reasons based on the value of things, but at the same time reject any realism about
that value applying to things independently of us, then we must be treated as the source of
value and in a way that makes rational choice possible” (Stern 2011, 90). The idea is thus
that if one acts, one must think that the end of one’s action is good. If one subsequently
10
I do not discuss the exegetical question of whether Stern’s reconstruction of Korsgaard’s argument is plausible in
light of the original text. Instead, I discuss Stern’s reconstruction of Korsgaard’s argument as the strongest possible
Korsgaardian argument for the value of humanity.
11
Korsgaard notes that this is meant as a summary of her interpretation of Kant’s argument in Kant’s Formula of
Humanity (Korsgaard 1996, 122f38). I think this is slightly misleading insofar as the account of Kant’s argument in
The Sources is more ambitious than the account in Kant’s Formula of Humanity. The former account neither starts
from an idea of the objective good nor from the categorical imperative. Instead, this argument explores the
necessary conditions of the possibility of having a reason for action.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
201
rejects the idea that the end of one’s action is good in itself (i.e. if one rejects moral realism),
the end of one’s action can only be good because it is good for oneself. But if the end of
one’s action is only good if it is good for oneself, one must regard oneself, i.e. the life of
someone who acts for reasons, as valuable because otherwise one’s end cannot be good.
Why do you have to see yourself as valuable in order to act for a reason? Stern puts
forward the following reconstruction of Korsgaard’s argument along the general lines of
Korsgaard’s interpretation of Kant’s argument for the value of humanity.
1.
To rationally choose to φ, you must take it that φ-ing is the rational thing to do.
2.
Since X [i.e. the object of your action] in itself gives you no reason to φ, you can
take it that φ-ing is the rational thing to do only if you regard your practical
identity as making it rational to φ.
3.
You cannot regard your practical identity as making φ-ing the rational thing to do
unless you can see some value in that practical identity.
4.
You cannot see any value in any particular practical identity as such, but can
regard it as valuable only because of the contribution it makes to giving you
reasons and values by which to live.
5.
You cannot see having a practical identity as valuable in this way unless you think
your having a life containing reasons and values is important.
6.
You cannot regard it as important that your life contain reasons and values unless
you regard your leading a rationally structured life as valuable.
7.
You cannot regard your leading a rationally structured life as valuable unless you
value yourself qua rational agent.
8.
Therefore, you must value yourself qua rational agent, if you are to make any
rational choice (Stern 2011, 90).
Let me unpack this argument. The first three steps are roughly the same as the first steps of
the regress argument: we need a reason for action (we need to rationally choose to φ) and
our practical identities can provide us with reasons for action. The main difference between
the regress argument and the source of normativity argument has to do with the
justification for premise 4: the idea that practical identities are not valuable as such. The
reason for this is that, according to Stern, Korsgaard does not necessarily have to rely on
something like the regress argument, i.e. the idea that we should always question our
particular practical identities, to reach the conclusions that particular practical identities are
CHAPTER 6
202
not valuable as such. Instead, Stern states that
to see value in any particular identity as such is to be committed to realism, to
thinking that being a father, an Englishman, a university lecture or whatever
matters as such; or (in a way that is in the end equally realist), it matters because of
the intrinsically valuable things it leads you do to. But, as we have seen, Korsgaard
also takes such realist positions to be problematic, so can perhaps use such
arguments [i.e. the argument that practical identities are not valuable as such]
here, without appealing to the regress considerations at all (2011, 91–92).
So, according to Stern, Korsgaard can justify the idea that a particular practical identity is
not valuable as such by relying on her rejection of moral realism. The reason for this is that
practical identities could only be valuable as such if one commits oneself to realism. Thus,
Stern’s point is that if one rejects realism, particular practical identities cannot be valuable
as such. Assuming for the sake of argument that moral realism is indeed problematic (I
come back to this assumption below), practical identities cannot be valuable as such.
Although practical identities are not valuable as such, we need a practical identity in
order to have a reason for action (premise 2), so there must be another source for the value
of our particular practical identities (premise 3). According to Stern’s Korsgaard, if one
needs a practical identity to act on reasons, and if particular practical identities are not
valuable as such, the only value of particular practical identities is that they make it possible
for an agent to act on the basis of reasons (the second part of step 4): “such identities have
the general capacity of enabling the agent to live a life containing reasons: because I have
whatever particular practical identities ... I can then find things to be valuable and act
rationally accordingly, in a way that gives me unity as a subject” (Stern 2011, 92).
The argument from step 5 to step 8 of the argument subsequently proceeds as follows:
But then (step 5), to think that this makes having some sort of particular practical
identity important, you must think that it matters that your life have the sort of
rational structure that having such identities provides; but (step 6), to see that as
mattering, you must see value in your leading a rationally structured life. And
then, finally, to see value in your leading such a life, you must see your rational
nature as valuable, which is to value your humanity (Stern 2011, 92).
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
203
This is a very condensed argument, but roughly the argument is that if practical identities
can only be valuable insofar as they make it possible to live a life containing reasons, you
must think that it is important that your life has a rational structure, and this is
subsequently taken to mean that you must value your rational nature (step 7) and your
humanity (step 8).
I think that the main conclusion of the argument is reached in step 5: that you must
think that your having a life containing reasons and values is important for your practical
identities to be valuable. Since Stern, following Korsgaard herself, does not further define
‘rational nature’ and/or ‘humanity’ except by reference to each other and to the idea of
leading a life which contains reasons, I understand the remaining steps of the argument
(the claim that you must regard your leading a rationally structured life as valuable (step 6)
and that you must value your rational nature (step 7) and ultimately your humanity (step
8)) as saying roughly the same thing in different words. My understanding is therefore that
the concept of rational nature and humanity does not mean anything over and above living
a life containing reasons, i.e. leading a rationally structured life (I come back to this below,
and I also come back to the question of what it means to ‘value’ your own ‘humanity’ within
the terms of this argument).
How successful is the source of reasons argument? I think that Stern’s reconstruction of
Korsgaard’s argument can answer some of the recurring objections to Korsgaard’s
argument for the value of humanity. For instance, one recurring objection is that even if
one accepts that humanity is the source of reason and/or value, this does not necessarily
imply that we should value it. Rae Langton criticizes Korsgaard by claiming that
we have no more antecedent reason to expect the creators of goodness to be good
than to expect painters of the blue to be blue, or the creators of babies to be babies.
In general we don’t think the source of something valuable must itself be valuable.
War can produce good poets, chicken manure can produce good roses, and in
general the sources of good things can be bad (Langton 2007, 156–57).
Stern’s reconstruction allows us to respond to this objection. Korsgaard’s point, on Stern’s
reconstruction, is not simply that the source of something valuable must itself be valuable,
but that practical identities can only be valuable insofar as one values living a life containing
reasons, which in turn means that one must value one’s humanity, where humanity is
understood simply in terms of leading such a life. Humanity should be valued as the source
CHAPTER 6
204
of all reasons not simply because it is the source of value, but because otherwise one could
not act for reasons or one could not have any values.12 This is certainly different in the
examples Langton mentions. Obviously there can be good poets without war.
There are, however, other problems with the source of reasons argument. First, as Stern
(2011, 93) notes, the argument is only valid if one rejects moral realism. In The Sources,
Korsgaard acknowledges that her argument presupposes that realism does not work when
she claims that “[Kant’s] point is the one I have been making - that being human we must
endorse our impulses before we can act on them. He asked what it is that makes these
objects good, and, rejecting one form of realism, he decided that the goodness was not in
the objects themselves” (Korsgaard 1996, 122). In other writings, however, Korsgaard
suggests that she wants to remain neutral about moral realism. She states that
if we could give an otherwise adequate account of practical reason without taking
on the commitments of substantive realism about reasons, we would not need to
argue against substantive realism over and above doing that. This is because the
only reason for accepting the commitments of substantive realism is the worry
that we cannot give an adequate account of practical reason without them
(Korsgaard 2007, 12f5).
The current discussion suggests that Korsgaard cannot stay neutral about moral realism
because her argument for the value of humanity presupposes the rejection of moral realism.
In his discussion of the source of reasons argument, Stern assumes, for the sake of
argument, that realism should indeed be rejected and concludes that, on its own terms, the
source of reasons argument is valid. In other writings, Stern defends moral realism against
Kantian constructivism (Stern 2012a; Stern 2013b). The observation that Korsgaard has not
provided a convincing argument against moral realism is a recurring objection to her view
(See e.g. Regan 2002, 272; Darwall 2006, 231). Whether or not one has to accept
Korsgaard’s argument thus ultimately depends on whether or not one thinks that moral
realism should be rejected.
In general, it seems that a Kantian constructivist would be well advised to try to provide
12
Langton ultimately rejects Korsgaard’s constructivism and she thinks that Korsgaard has not provided a
sufficient argument against moral realism (Langton 2007, 182f39). Langton could thus deny that one’s humanity is
the only available source of reasons. I come back to the role of the rejection of realism in Korsgaard’s argument
below.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
205
an argument for interpersonal morality which does not presuppose the failure of moral
realism but which ‘merely’ tries to provide an alternative non-realist justification of
objective morality. An argument that presupposes the failure of moral realism is in
principle more controversial than an argument which tries to show that one has to accept
certain objective moral principles without relying on moral facts. I think Gewirth’s
argument, which I will discuss in the next chapter, succeeds in doing so. This alone would
be a reason to prefer Gewirth’s argument over Korsgaard’s argument.
However, I think that there is a second, more fundamental, problem with the source of
reasons argument, independent of the issue of realism and even independent of the validity
of the argument as a whole. In what follows, I therefore assume that realism should indeed
be rejected, and I will also assume that Stern’s reconstructed argument is valid. My
objection is this: even if the argument for the value of humanity, as Stern reconstructs it, is
valid, it is unclear whether the argument could lead to any normative conclusion. The
reason for this is that it is unclear how, at least on this specific argument, acknowledging
the value of agency translates into any normative requirements for action. This implies that
even if the conclusion of this argument can be successfully universalized through the
argument from the sufficiency of agency, so that one has to value the humanity of all other
agents as well, it would fail to justify any normative principle of interpersonal morality.
This objection might sound counterintuitive at first. After all, is it not obvious that one
could do all kinds of things to express respect for the value of one’s own and someone else’s
humanity? And is there not also a plurality of ways in which one could fail to express this
kind of respect? Torturing other people is clearly inhumane. The same goes for slavery,
living in severe poverty, human trafficking and so on. In other words, is it not obvious that
the value of humanity translates into certain normative requirements?
Although I agree that on our common-sense understanding of the ‘value’ of humanity,
activities like torture or slavery very obviously fail to express respect for the value of
humanity. But I am not so sure that Korsgaard’s argument justifies this particular
understanding of the value of humanity. After all, as I have already noted above, Korsgaard
introduces the value of humanity as a technical term in her argument. This means we can
only assess the success of the argument for the value of humanity if we are clear about what
it means to ‘value’ one’s (own) ‘humanity’ in terms of Korsgaard’s argument. Now that we
have discussed the source of reasons argument, we can turn to this question: what does it
mean, exactly, to say that one must value one’s (own) humanity in the context of Stern’s
reconstruction of Korsgaard’s source of reason argument?
CHAPTER 6
206
Before turning to the ‘valuing’ in valuing one’s (own) humanity, let me start with the
concept of humanity. On Stern’s reconstruction of Korsgaard’s argument, ‘humanity’ refers
to ‘rational nature’, where rational nature is understood as a life containing reasons and
values. As mentioned, Stern does not define humanity except by reference to the idea of
leading a rationally structured life. In addition, if one looks at the argument, humanity
could not possibly mean anything more than ‘one’s rational nature’. Practical identities
provide reasons, and practical identities can only be valuable to the extent to which one
thinks it is valuable that they provide reasons. That is, practical identities are important
insofar as they contribute to a rationally structured life, which means that they can only be
important insofar as one thinks that leading a rationally structured life is important. On
this definition of humanity, the absence of humanity is not an inhumane life, as we would
normally perhaps think of it, but means living a life without any reasons and values.
Perhaps such a life would be the life of certain non-human animals, or robots or inanimate
things. In other words, the absence of humanity means that one would be a mere object and
not an agent; one would be “the mere undergoer of ... experiences” (Korsgaard 1989, 120).
Humanity is thus another word for agency or an agential life. Crucially, in this argument
humanity is understood as a technical term, which, again, should not be confused with a
more encompassing everyday notion of humanity. This should make us cautious, though,
when confronted with conclusions about ‘respecting’ or ‘valuing’ our humanity, because
they will not be conclusions about, for instance, basic human rights but rather claims about
respecting or valuing our leading a life containing reasons (and it is at least not obvious that
these are necesssarily the same things).
So now we turn to the following question: what does it mean to value humanity on this
reconstruction of the argument? I think this is less clear on Stern’s reconstruction of
Korsgaard’s argument. Stern interchangeably speaks about the need “to value your
humanity” (Stern 2011, 92) and about thinking that “having a life containing reasons and
values is important” (Stern 2011, 90 my emphasis). In addition, he claims that denying the
value of humanity means thinking that “you and your life were utterly worthless, pointless,
meaningless—that in your eyes, you were valueless” (Stern 2011, 89). In addition, he
interchangeably talks about ‘thinking’, ‘seeing’ and ‘regarding’ yourself as valuable (Stern
2011, 89–90).
The value of one’s humanity, then, is to think or see that it is important to have or that
there is a value to having a life containing reasons and values. However, if this is what it
means to value one’s humanity, it is unclear how this value could generate any normative
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
207
constraints on action. Either one is an agent and therefore one necessarily has to think that
it is important to have a life containing reasons (because otherwise one could not have a
reason for action) or one does not think that this is important, but in that case one would
not be an agent in the first place. In other words, if valuing your humanity (on Korsgaard’s
technical definition of humanity) is a necessary condition of the possibility of acting for a
reason, there can be, according to this argument, no agents who do not value their
humanity, but who should value their humanity. 13 There is no reason to value one’s
humanity if one is not an agent, but if one is an agent one necessarily has to think that
having a life containing reasons is important because otherwise one could not have a reason
in the first place (again, assuming that the argument is valid).14
The same point could be made in a slightly different way. Recall that the starting point
of Korsgaard’s argument is that we inescapably act for reasons, i.e. that we inescapably
understand ourselves as agents. The conclusion of the argument is that we should value our
humanity, i.e. that we should value having a life containing reasons and values. However,
for the value of humanity to lead to any normative conclusions there should be ways in
which we could fail to understand ourselves as agents. As Korsgaard herself acknowledges,
“[T]here is no normativity if you cannot be wrong” (Korsgaard 1996, 161). But if one
cannot not understand oneself as an agent (because, according to Korsgaard, one
inescapably understands oneself as an agent), one could not fail to value one’s humanity,
given that, according to the argument, valuing one’s humanity is a necessary condition of
the possibility of agency. And if one does not understand oneself as an agent, one does not
have a reason to value one’s own humanity. The value of humanity, in other words, does
not provide a source of (moral) normativity, because, according to this argument,
understanding oneself as an agent and valuing one’s humanity cannot be seperated because
valuing one’s agency means nothing over and above thinking it important to be an agent,
and one can only, according to this argument, understand oneself as an agent if one actually
thinks that it is important to be an agent.
But what about the interpersonal case? Could one not fail to value the humanity of
other persons? If so, the value of humanity could lead to normative constraints in the
13
On this reconstruction, Korsgaard’s argument might also be susceptible to the problem of misconception (see
chapter 4): couldn’t we imagine someone who acts on reasons but who does not actually value leading a life with
reasons (cf. Street’s example of a social insect)?
14
Cf. Beyleveld, who writes that “my having this power [i.e. the power for rational choice] is not therefore an end
that I ought to pursue for its own sake, because I necessarily have it when I set any end” (Beyleveld 2015, 595).
CHAPTER 6
208
interpersonal domain (even if it does not lead to any intrapersonal normative claims).
Assume, for the sake of argument, that the argument from the sufficiency of agency
succeeds and that one must not only value one’s own humanity but also the humanity of all
other agents, i.e. that one must value that others live a life containing reasons. What are the
normative implications of the universal value of humanity?
I am not sure about the interpersonal case, and again a lot hinges on how exactly one
understands ‘agency’, value’ and ‘humanity’. Let me start by saying that, in a trivial sense,
the universal value of humanity does lead to a normative requirement. After all, one could
fail to think that others have value without ceasing to be an agent oneself. Valuing the
agency of others is not something that I must do in order to be able to act on reasons, but
something I must accept on pain of consistency, i.e. on pain of denying that my humanity is
valuable by virtue of being an agent. Valuing the humanity of other agents is thus
something one must do even if one currently does not. This normative requirement is,
however, trivial as long as it remains unclear whether acknowledging the value of agency
actually generates any normative requirements on actions, i.e. whether it tells us how (not)
to treat other agents (instead of just saying that one must think that the other is valuable,
whatever that might mean). Whether or not acknowledging the value of other agents
actually generates any constraints on action thus depends on how, on Korsgaard’s account,
values translate into norms of action.15
Earlier in this chapter, I have already noted that in recent work, Korsgaard proposes to
understand the value of humanity first and foremost as referring to a normative status or
standing and that norms should subsequently be understood as expressing respect for this
standing or status. She writes that “duties are expressions of respect, not for the property of
rationality, but for the legislative standing that it confers upon us” (Korsgaard forthcoming,
31–32). So this leads to the following question: what would it mean to express respect for
the status that others have by virtue of leading a life containing reasons?
Again, given that Korsgaard claims that leading a life containing reasons is inescapable
not just for A but for any agent, there seems to be little, if nothing, one could do to
undermine this status. Even in the most deprived situations it seems to be at least possible
to “back up and bring [an] impulse into view” (Korsgaard 1996, 93). Even ending
15
To a certain extent, the discussion that follows anticipates the discussion of the emptiness objection in the final
two chapters of this thesis. My objection in this section is, I think slightly different from the emptiness objection,
because my objection is not that the value of humanity is empty (because it is not: it is to value a life containing
reasons). The objection is that the value of humanity is not, or does not generate, a normative principle of action.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON I
209
someone’s life does not necessarily seem to be incompatible with valuing the humanity of
another person. After all, the moment the other person is still alive he or she is, on
Korsgaard’s view, still an agent (which means that it is hard to see how one could possibly
fail to express respect for his or her agency), whereas the moment the other person is dead
there is no reason to value his or her agency in the first place (because he or she is no longer
an agent). Killing another person would only be problematic if the argument requires
respect for a continued existence of the agent. But, at least as it stands, this is not what the
argument argues for, and it is unclear how it could follow from the premises in Stern’s
reconstruction of Korsgaard’s argument. The problem here thus seems to be very similar to
the intrapersonal case: if agency is inescapable (not just for you but also for others) then it is
hard to see how one could fail to act in a way that respects the value of the agency of other
persons.
Note that my point is not that it would be impossible for Korsgaard to revise or
reconstruct her argument in such a way as to show that certain normative constraints can
follow from this argument – for instance, if it is our continued existence as an agent that is
somehow inescapable and which deserves moral respect, the argument might lead to
certain (minimal) norms of interpersonal morality (but even in that case it would still be a
far cry from justifying enlightenment morality).16 The point is rather that Korsgaard has
not given us a convincing account of how exactly the argument is supposed to lead to any
normative requirement.
What I hope to have shown is that the underlying problem with Korsgaard’s argument
is that what it means to value humanity depends both on how one understands value and
humanity and on how one reconstructs Korsgaard’s argument for the value of humanity.
Both value and humanity are introduced as technical terms and should not be confused
with the everyday understanding of these concepts. I am afraid that it is only if we lose sight
of the crucial distinction between the concepts of humanity and of value as the technical
terms in Korsgaard’s argument and the everyday understanding of these concepts that have
the moral–political depth that Korsgaard wants them to have that we can conclude that
justifying the value of humanity shows that “Enlightenment morality is true” (Korsgaard
1996, 123). At best, Korsgaard’s argument for the value of humanity as articulating the
source of (interpersonal) moral normativity is incomplete. At worst, the argument fails
16
In fact, Korsgaard has written quite extensively on the diachronic nature of agency in other work. See in
particular Korsgaard (1989).
CHAPTER 6
210
because the idea of the value of humanity is too trivial to generate any norms of action.
I will not try to reconstruct, specify or revise Korsgaard’s argument in such a way as to
create normative space between inescapably understanding oneself as an agent and
necessarily valuing one’s own and other agents’ humanity. Instead, in the next chapter, I
will address Gewirth’s transcendental argument from the first person. As will become clear
in the next chapter, the starting point of Gewirth’s argument is roughly similar to
Korsgaard’s argument, and Gewirth also makes use of the argument from the sufficiency of
agency. In contrast to Korsgaard, however, Gewirth succeeds in justifying categorical
normative commitments on the basis of agential self-understanding. Or so I will argue.
Although I will not discuss the (exegetical) question of the extent to which Korsgaard’s and
Gewirth’s theories are ultimately compatible, I think that given the failure of Korsgaard’s
argument to generate normative conclusions, and given the structural similarities between
their arguments, Gewirth might provide a way to save a roughly Korsgaardian
transcendental argument from the objections that I have presented in this chapter.
4. Conclusion
In this chapter, I have argued that Korsgaard’s argument from the first person for
interpersonal morality, through the argument from the sufficiency of agency, fails. This
failure, I should stress again, has less to do with the argument from the sufficiency of
agency, and more to do with the details of Korsgaard’s argument for the value of humanity.
The problem with Korsgaard’s argument is that the first phase of the argument, the
argument for the claim that A must value his or her (own) humanity, either relies on an
escapable starting point (the regress of reasons argument) or does not lead to a non-trivial
normative conclusion, and that therefore the value of humanity does not entail a normative
principle of interpersonal morality even if it can be successfully universalized through the
argument for the sufficiency of agency (the source of reasons argument).
The discussion of Korsgaard has not so much shown that transcendental arguments
from the first person cannot work, but rather that the success of the argument from the
sufficiency of agency depends on what I have called the first step of a transcendental
argument from the first person, and there are certain problems with Korsgaard’s specific
argument for her claim that A must value his or her (own) humanity.
In the next chapter, I will discuss a different transcendental argument from the first
person: Alan Gewirth’s argument for the principle of generic consistency.
Chapter 7
Transcendental Arguments from the First
Person II: The Principle of Generic
Consistency
1. Introduction
In Reason and Morality, Alan Gewirth puts forward one of the most comprehensive
transcendental arguments from the first person to date. According to Gewirth,
interpersonal morality is grounded in the self-understanding of the individual agent, where
agency is understood in terms of voluntarily trying to achieve an end (‘purposive agency’).
Gewirth aims to show that any person who understands him- or herself as a purposive
agent is necessarily committed to a supreme principle of interpersonal morality whose
status and function is structurally similar to the status and function of the categorical
imperative in Kant’s ethics. Gewirth calls this moral principle ‘the principle of generic
consistency’, which roughly states that each agent has a claim right to the necessary social
and material preconditions of his or her agency – or what Gewirth calls the ‘generic features
of agency’ (1978, 135). Gewirth claims that any agent necessarily has to accept this principle
of interpersonal morality on pain of contradicting that one understands oneself as an agent.
Gewirth’s argument has been the object of extensive (mostly critical) discussion. In
recent debates on Kantian constructivism and transcendental arguments in meta-ethics,
however, Gewirth’s transcendental argument is hardly if ever discussed, and if his work is
mentioned it is only in passing or in footnotes (see e.g. Enoch 2006, 189f42), even by fellow
Kantian constructivists (see for instance Korsgaard 1996, 133; Darwall 2006, 30). Recent
CHAPTER 7
212
overview papers on constitutivism (Tubert 2010) and (Kantian) constructivism (Street
2010; Bagnoli 2014; Schafer 2015a; Schafer 2015b) do not even mention Gewirth.1
The goal of this chapter is limited. I do not aim to contribute anything substantive to
the existing literature on Gewirth, 2 nor to investigate why Gewirth has been largely
overlooked in recent debates in meta-ethics. Instead, the aim of this chapter is to discuss
Gewirth’s argument for the principle of generic consistency as an example of a
transcendental argument from the first person and to explain why I think that, at least on
one interpretation of the argument, Gewirth succeeds in justifying a categorical and
universal principle of interpersonal morality. I focus specifically on Gewirth’s argument for
interpersonal morality, the argument from the sufficiency of agency. I will come back to the
substantive implications of Gewirth’s theory in the chapters that follow.
In my discussion of Gewirth, I draw on the recent elaboration and clarification of
Gewirth’s argument by Deryck Beyleveld (Beyleveld 1991; Beyleveld and Bos 2009;
Beyleveld 2013; Beyleveld 2015). In his more recent work (beginning with Beyleveld and
Bos 2009), Beyleveld has stressed the foundational role of the instrumental principle in
Gewirth’s argument. I follow Beyleveld’s reconstruction of Gewirth’s argument because I
take this to be the strongest possible reconstruction of Gewirth’s argument (as I will try to
show below). I do not pursue the question of the extent to which Beyleveld’s Gewirth and
the Gewirth of Reason and Morality are necessarily the same.3
1
A notable exception is Stern (2013a), but he only mentions Gewirth and subsequently only engages with
Korsgaard’s transcendental argument.
2
One the most controversial elements of Gewirth’s project is his argument for interpersonal morality through the
argument from the sufficiency of agency (MacIntyre 1985, 2nd:80; Williams 1986; McMahon 1986; Kramer and
Simmonds 1996; Illies 2003; Chitty 2008; for a collection of papers on Gewirth’s work see Regis 1984; Boylan 1999;
Bauhn 2016). Gewirth has responded to many of his critics (see e.g. Gewirth 1988; Gewirth 1997; Gewirth 1996).
In addition, Deryck Beyleveld has taken up the project of clarifying Gewirth’s argument and of defending it against
critics (Beyleveld 1991; for some recent articles on Gewirth see Beyleveld and Bos 2009; Beyleveld 2013; Beyleveld
2015). In addition, Beyleveld has worked on the application of Gewirth’s theory to law and bioethics (see e.g.
Beyleveld and Brownsword 2001; Beyleveld and Brownsword 2007).
3
As I will briefly discuss in section 2, the main difference between Beyleveld’s Gewirth and the Gewirth of Reason
and Morality is that whereas Gewirth claims that the instrumental principle is part of what it means to be rational
and is itself not extensively argued for (Gewirth 1978, 22–23, 46), Beyleveld, at least in his recent work, claims that
it is dialectically necessary to accept the instrumental principle (Beyleveld and Bos 2009). Beyleveld presents this as
a “clarification” of Gewirth’s argument (Beyleveld 2013, 209f15). Because I am interested in the possibility of
providing an argument from the first person, I will not pursue the question of whether this is a clarification of
Gewirth’s argument or a (promising) departure from Gewirth’s position. Note, however, that Gewirth’s remarks
that a commitment to the standards of reason are “not a mere optional or parochial ‘commitment’” (1978, 22),
that “any attempt at … [a] justification [of the standards of reasoning] must make use of deduction or induction
or both” (1978, 22) and that “any attack on reason or any claim to supersede it by some other human power or
criterion must rely on reason to justify its claim” (1978, 23) at least hint at a transcendental justification of
rationality, including instrumental rationality.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
213
The structure of this chapter is as follows. First, I briefly introduce Gewirth’s
transcendental argument from the first person and his specific method: the dialectically
necessary method (section 2). Subsequently, I discuss the two stages of his argument: the
first stage argues for the claim that an agent necessarily has to claim a right to the necessary
preconditions of his or her agency (section 3). The second stage argues that this right
should be universalized through the argument from the sufficiency of agency (section 4). In
section 5, I briefly revisit the logical objection to transcendental arguments from the first
person and argue that the logical objection fails. In section 6, I briefly discuss the wrong
kind of reason objection and argue that there is no reason to think that Gewirth provides
the wrong kind of reason for interpersonal morality.
2. Gewirth’s Argument from Purposive Agency
Just like Korsgaard’s argument, Gewirth’s transcendental argument starts from A’s selfunderstanding as a rational agent. Gewirth defines an agent as someone who understands
him- or herself as acting purposively (acting so as to achieve a purpose) and voluntarily
(chosen in an unforced way) (Gewirth 1978, 31). In addition, an agent is defined as
(minimally) rational if he or she accepts the “canons of deductive and inductive logic”
(Gewirth 1978, 22), which include, most importantly, the principle of non-contradiction
and the principle of instrumental rationality (Gewirth 1978, 46). Much hinges on the
question of what exactly it means to act purposively and voluntarily (and I will come back
to this question in the next chapter), but for the moment let us work with an intuitive
understanding of agency as involving typical human activities, such as making coffee,
eating and working.
Again, just like in Korsgaard’s argument, agency is not understood from the thirdperson perspective, but as an inescapable aspect of our self-understanding. Agency is thus
understood from the first-person perspective. Gewirth’s point is not so much that we
inescapably are agents, but that we inescapably understand ourselves as acting to achieve
certain purposes. Gewirth describes this starting point as being “internal” and “practical”
(Gewirth 1978, 160). Beyleveld therefore concludes that the starting point is
phenomenological (cf. Beyleveld 1991, 69).
Gewirth puts forward two arguments for the claim that agential self-understanding is
inescapable in the sense that is relevant in the context of justifying morality (see also
chapter 4). First, Gewirth claims that agency is the inescapable subject matter of morality
CHAPTER 7
214
and practical reasoning, insofar as all practical precepts, including moral precepts, have
action as their object, independent of the content of these precepts (Gewirth 1978, 25).
Beyleveld therefore writes that “the properties attributed to an agent are those that a being
must have in order to be an intelligible subject of practical precepting (a being that can ask
what it has reason to do)” (Beyleveld 2013, 212f26). Practical precepts, including moral
principles, tell us how to act, and in this sense all practical precepts have action as their
subject matter. Action is therefore inescapable from the perspective of practical reasoning,
including moral reasoning.
Second, Gewirth acknowledges that the observation that action is the subject matter of
morality will not satisfy the amoralist (Gewirth 1978, 25). The amoralist, after all, denies
that there are any justified moral principles, and she will therefore not be impressed by the
claim that all moral theories will have to acknowledge the centrality of action. Gewirth
therefore claims that purposive agency is not just the relevant aspect of our selfunderstanding in the context of morality, and adds that agency is presupposed by all
human practices in the sense that insofar as one engages in any human practice one
understands oneself as an agent (Gewirth 1978, 29). Agency is thus not just a contingent
starting point but can only be ‘escaped’ by either denying morality and practical reasoning
or, more fundamentally, “by disavowing and refraining from all intentional action and
from the whole sphere of practice” (Gewirth 1978, 29). In other words, just like Korsgaard,
Gewirth claims that agency is inescapable, at least from our first-person perspective.
How does Gewirth argue from purposive agency as the inescapable feature of our selfunderstanding to intersubjective morality? Gewirth employs a specific type of
transcendental argument, which he calls the ‘dialectically necessary method’ (Gewirth 1978,
42). The method is ‘dialectical’ insofar as it aims to show that “certain judgments, in a
certain sequence of argument, are necessarily attributable to every rational agent” (Gewirth
1978, 42). The method thus starts “from the standpoint of the agent” and subsequently
presents certain judgements that agents necessarily have to make “on the basis of what is
necessarily involved in their actions” (Gewirth 1978, 44).
Gewirth contrasts the dialectical method with an ‘assertoric’ method, which makes
statements independent of the standpoint of an agent. Gewirth describes the difference
between a dialectical and an assertoric method by saying that
it is one thing to say assertorically that X is good; it is another thing to say
dialectically that X is good from the standpoint of some person, or that some
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
215
person thinks or says ‘X is good.’ Where the assertoric statement is about X, the
dialectical statement is about some person’s judgement or statement about X
(Gewirth 1978, 44).
An assertoric statement is thus a third-personal statement about some X, understood
independent of the standpoint of an agent. A dialectical statement is a statement about X
from the perspective of the agent.
The method is a method of dialectical necessity because it presents judgements that
every agent necessarily has to make insofar as he or she understands him- or herself as an
agent. This contrasts with a dialectically contingent method, which “begins from singular
or general statements or judgements that reflect the variable beliefs, interests, or ideals of
some person or group” (Gewirth 1978, 43). Dialectically contingent judgements only have
to be accepted by the person who accepts certain particular statements or judgements,
whereas dialectically necessary judgements have to be accepted by any purposive agent.
Importantly, Gewirth stresses that although the method of dialectical necessity is
supposed to justify necessary judgements, these judgements should only necessarily be
accepted by human purposive agents given the conditions they find themselves in. He
writes that
there is no commitment … to the idea that necessary propositions are true “in all
possible worlds,” in the sense in which the “possible” reflect the infinity of logical
possibilities. The dialectical necessities that derive from what every agent must say
or think because of the nature of action reflect the necessities of this existing
world, including the limits set by its own structure and potentialities (Gewirth
1978, 45).
In line with this emphasis on human agency, Beyleveld and Brownsword claim that “having
the power of agency is necessary, but it is not sufficient to generate a need for the objects of
rights” (Beyleveld and Brownsword 2001, 115). They write:
“I” (the agent who is the subject of the argument) am required to claim that the
sufficient reason why I have the generic rights is that I am an agent. It is
important, however, to appreciate that it is implicit that agents are vulnerable.
While it is true that “I” (as such a vulnerable agent) must claim that my being an
CHAPTER 7
216
agent is the sufficient reason why I have the generic rights, the argument cannot be
applied to non-vulnerable agents (if such can exist) (Beyleveld and Brownsword
2001, 115).
According to Beyleveld and Brownsword, the vulnerability of human agents consists of two
elements. First, the capacity to act against the moral law, for otherwise the moral law would
have no application. Second, the possibility of being harmed by other agents, for otherwise
there would be nothing that could deserve moral protection (Beyleveld and Brownsword
2001, 114). As we will see, Gewirth’s argument for the principle of generic consistency relies
on these kinds of general anthropological considerations, and its claim to necessity should
thus be understood in the context of the phenomenology of human agents.4
A dialectically necessary method is a form of transcendental argumentation insofar as it
starts from our first-person perspective and insofar as it makes claims about what every
agent, from his or her first-person perspective as an agent, necessarily has to accept insofar
as he or she understands him- or herself as an agent. Gewirth’s aim is to show that there are
certain judgements that an agent cannot deny on pain of self-contradiction, i.e. on pain of
denying that he or she is a purposive agent, which, according to Gewirth, he or she cannot
deny at least insofar as the claim ‘I am an agent’ is understood dialectically, i.e. from the
first-person perspective, rather than assertorically, i.e. from the third-person perspective.
There are different ways in which one could reconstruct Gewirth’s argument for the
principle of generic consistency. I think that one of the most important, and also the most
distinguishing, features of Gewirth’s argument is the importance of the principle of
instrumental reasoning (referred to from now on as the instrumental principle) for the
justification of interpersonal morality.5 Whereas both Gewirth’s definition of agency and
his dialectically necessary method are, I believe, largely compatible with, for instance,
Korsgaard’s transcendental argument, it’s the emphasis on the instrumental principle in the
justification of morality that distinguishes Gewirth’s argument from Korsgaard’s and other
arguments, and, as I have argued in chapter 4, this is also the interpretation of
transcendental arguments that is able to explain how normative conclusions can follow
4
In the chapters that follow, I will discuss the importance of anthropological considerations for the application of
a principle of morality.
5
Beyleveld stresses the importance of the instrumental principle in his influential reconstruction of Gewirth’s
argument (Beyleveld 1991, 13–46). In recent articles, Beyleveld claims that the instrumental principle plays a
foundational role in Gewirth’s argument (See in particular Beyleveld and Bos 2009; Beyleveld 2013; Beyleveld
2015).
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
217
from a transcendental argument (see the discussion of Street’s no-normativity objection in
chapter 4).6
The instrumental principle states that if one pursues X and if Y is a necessary means to
achieve X, one ought either to do Y or give up the pursuit of X. As I have already noted
above, in Reason and Morality Gewirth includes the instrumental principle in his definition
of rational agency, and he does not, or at least not explicitly, give a further justification of
the instrumental principle. Beyleveld, on the other hand, claims that it is dialectically
necessary for an agent to accept the instrumental principle (see in particular Beyleveld and
Bos 2009, 16–18).
In a recent paper, Beyleveld and Bos argue for the dialectical necessity of the
instrumental principle as follows:
What defines me as an agent is that in doing something X for E, I am (voluntarily)
trying thereby to achieve E. But, to try to achieve a purpose by doing something is
(by definition) to be committed to doing what is necessary and sufficient to
achieve the purpose. If I do not do what I believe to be necessary and sufficient to
achieve E then I am not genuinely or sincerely trying to achieve E. So, if I believe
that my doing X is necessary to achieve E, I must (it is necessary for me to) do X or
deny that I am trying to achieve E. Hence, I ought (on pain of contradicting that I
am trying to achieve E) to do X where I believe that my doing X is necessary to
achieve E. So, where I believe that my doing X is necessary to achieve E, I ought to
do X or give up E as my purpose. Since doing something (voluntarily) in order to
achieve a purpose defines me as an agent, it follows that where I believe that my
doing X is necessary to achieve E, I ought (on pain of contradicting that I am an
agent) to do X or give up E as my purpose (Beyleveld and Bos 2009, 18).
Beyleveld and Bos thus claim that taking the means to one’s end (or giving up the end) is
simply what it means to be an agent and that this cannot therefore be denied without
contradicting that one is an agent.
James Skidmore makes a similar point when he discusses the possibility of providing a
transcendental argument for the instrumental principle. He claims that an agent who
6
For a brief comparison of the role of the instrumental principle in Gewirth and Korsgaard see Beyleveld (2015,
595–97). For a comparison between Gewirth and Kant on this issue see (Beyleveld 2016).
CHAPTER 7
218
denies the instrumental principle would be unintelligible (Skidmore 2002, 138). Skidmore
illustrates this point by imagining a conversation between a ‘questioner’ and a ‘sceptic’
regarding the instrumental principle. The conversation takes place on a Friday afternoon,
and the sceptic has just told the questioner that he is planning to go the beach for the
weekend. Skidmore imagines that the following conversation unfolds after the sceptic has
told the questioner about his plans for the weekend:
Questioner: Sounds nice. How are you going to get there on such short notice?
Skeptic: I don’t know. What difference does that make?
Questioner: You know, you might be able to catch the bus there. I think the last
one leaves in an hour.
Skeptic: Yes, I know. But I hate riding the bus.
Questioner: So you have some other way to get there?
Skeptic: No.
Questioner: But you said you intended to spend the weekend at the ocean.
Skeptic: I do. They say the weather will be beautiful.
Questioner: And the only way you can get there is by taking the bus?
Skeptic: Yes, as far as I know, but what’s your point? I’m not taking the bus. Ah,
but how wonderful it will be to spend a couple of days relaxing on the beach . . .
(Skidmore 2002, 129).
Skidmore’s point is that the sceptic becomes unintelligible as an agent, i.e. as someone who
pursues an end, and that the sceptic therefore has to accept the instrumental principle.
Although much more needs to be said about the justification of the instrumental
principle, I will assume that any purposive agent should indeed accept the instrumental
principle. After all, it’s not the instrumental principle that’s typically regarded as
controversial, but rather the question of whether this strategy could lead to moral
conclusions (see e.g. Skidmore 2002) (see also chapter 4).
With these three basic elements on the table – the idea of human, purposive agency,
the method of dialectical necessity and the instrumental principle (as a principle that
should be dialectically accepted by any agent) – we can now turn to Gewirth’s justification
of interpersonal morality.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
219
3. Part 1: Rights to the Generic Features of Agency
I will discuss Gewirth’s argument for the principle of generic consistency in two steps
(analogous to the two steps in Korsgaard’s argument). In this section, I discuss the
argument which leads up to the justification of categorical normative commitments of the
individual agent, i.e. the judgement that an agent ought to claim a right to the generic
features of agency (this corresponds with Korsgaard’s argument for the value of our own
humanity). 7 In the next section, I will discuss Gewirth’s argument for interpersonal
morality, which ultimately results in the principle of generic consistency (this roughly
corresponds to Korsgaard’s argument for the universal value of humanity).
As I have already noted above, the starting point of Gewirth’s argument is the selfunderstanding of a person as a purposive agent. This leads to premise 1:
1) It is phenomenologically inescapable that I understand myself as a purposive
agent.
The argument for the principle of generic consistency basically proceeds by applying the
instrumental principle to the standpoint of the purposive agent, while keeping in mind that
Gewirth uses a dialectically necessary method and not a dialectically contingent method
and that it is dialectically necessary to accept the principle of instrumental rationality. A
dialectically contingent application of the instrumental principle would go something like
this: assume that I want to drink a coffee and also assume that the only way in which I can
drink a coffee is by ordering a coffee from the coffee bar. In that case, I ought to order a
coffee (or give up the end of drinking a coffee). It is thus dialectically contingent that I
ought to order a coffee (or give up the end), even if it is dialectically necessary that I accept
the principle of instrumental rationality.
A dialectically necessary application of the instrumental principle, on the other hand,
7
Beyleveld divides Gewirth’s argument into three stages (Beyleveld 1991, 14). The first stage leads to the
conclusion that it is dialectically necessary for the agent to claim that his or her freedom and well-being are
necessary goods. The second step leads to the conclusion that it is dialectically necessary for the agent to claim that
he or she has a claim right to his or her freedom and well-being. The final step leads to the principle of generic
consistency. My reason for discussing the argument in two stages (by taking Beyleveld’s stage 1 and 2 together) is
to make a comparison between Korsgaard and Gewirth easier. In addition, I want to focus mainly on what
Beyleveld calls stages 2 and 3 of the argument because they seem to be most controversial parts of the argument. In
Beyleveld and Bos (2009, 4–5) and Beyleveld (2015, 574–75), Beyleveld also divides the argument into these two
phases.
CHAPTER 7
220
does not proceed from a particular end but from the judgements that an agent necessarily
has to accept insofar as he or she acts on any end whatsoever. These judgements are
categorically instrumental judgements.8 According to Gewirth, there are certain means that
are necessary to engage in any action whatsoever. These so-called generic features of action
do not express any contingent features of action, but should be understood as the “invariant
features that pertain generically to all actions” (Gewirth 1978, 25).
The basic idea behind the ‘generic features of agency’ is that in order to achieve a
purpose there are not only specific means that I need for this specific purpose (e.g. the
existence of a coffee bar in order to be able to buy an espresso) but also general means
which are necessary to engage in any acting whatsoever. For instance, without being alive I
could not go out to buy an espresso, neither could I work on my thesis, talk to friends or
engage in whatever other activity one might think of. The idea is thus that the generic
features of agency are necessary for agency as such, independent of the specific content of
one’s purposes.
According to Gewirth, purposive agency has two generic features which parallel the
two features of action. The first generic feature of agency is freedom, which concerns the
ability to control one’s behaviour by one’s unforced choice (what Gewirth calls occurrent
freedom) and the longer-range ability to exercise such control (dispositional freedom)
(Gewirth 1978, 52). The second generic feature is well-being, which concerns the abilities
and (material and social) conditions of achieving one’s purposes. Well-being is further
subdivided into three levels: basic goods are the necessary preconditions of action; nonsubstractive goods are the goods that are necessary for the maintenance of agential
capabilities; and additive goods are the goods that are necessary for the improvement of
agential capabilities. I will come back to the question of how exactly to understand freedom
and well-being in chapter 9, because this has important implications for the substantive
conclusions that follow from the argument. For now I will just assume that there are at least
some necessary means of action (whatever they might be).
Applying the principle of instrumental rationality to the first-person perspective of the
agent, in combination with the idea of the generic features of agency, leads to the second
premise of the argument:
8
The terminology of ‘categorical instrumentality’ is introduced by (Beyleveld 1991, 30). As far as I am aware, this
terminology is not used by Gewirth himself, but I think this terminology is illuminating because it highlights the
importance of instrumental rationality for the argument.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
221
2) It is dialectically necessary that I ought to want to have the generic features of
agency.
If one ought to take the necessary means to one’s contingent ends, one also ought to take
the necessary means to any end whatsoever. This judgement is dialectically necessary
because it should be accepted simply by virtue of understanding oneself as an agent,
independent of one’s contingent ends.
Subsequently, Gewirth argues that if it is dialectically necessary that you ought to want
to have the generic features of agency, you ought also to want that others not interfere with
your generic features of agency. Gewirth writes that
since the agent holds that freedom and well-being are necessary goods for all his
actions, he also holds that it is necessary that he at least not be interfered with by
other persons in having freedom and well-being. For if he were thus interfered
with, he could not have what is required for him to act (Gewirth 1978, 78–79).9
Non-interference is thus categorically instrumental for having the generic features of
agency, while the generic features of agency are categorically instrumental for achieveing an
end (Gewirth 1978, 90–91; Beyleveld 2013, 215).
The idea that non-interference is categorically instrumental for having the generic
features of agency becomes clear by considering the idea that dialectically necessary
judgements “reflect the necessities of this existing world” (Gewirth 1978, 45) and the idea
that the argument therefore relies on certain general anthropological assumptions.
Although Gewirth does not make these anthropological assumptions explicit, I think they
help to illuminate the method of dialectical necessity and the dialectical necessity of
accepting certain specific judgements, such as the judgement that one ought to want others
not to interfere with one’s generic features of agency (cf. Steigleder 1999).
For one thing, the claim that non-interference is categorically instrumental for having
the generic features of agency presupposes that a plurality of agents exists. If there were no
9
Gewirth goes on to argue that the rights claim also implies that other agents have to support one’s freedom and
well-being to the extent to which this would undermine one’s own freedom and well-being with a similar degree of
necessity of action. In the remainder of this chapter, I will leave out this complication and instead focus only on
non-interference. The reason for this is that if the argument cannot even establish negative duties, it will also not
establish positive duties. The introduction of positive duties thus introduces further complications not essential for
the purpose of my discussion here.
CHAPTER 7
222
other agents, it would not make sense to ought to want others to not interfere with one’s
generic features of agency. More substantially, it also presupposes that one is a vulnerable
and dependent being. For instance, imagine that I am an invulnerable, self-sufficient being
which cannot in any way be harmed or supported with respect to the necessary
preconditions of my agency. In this case, it would not follow that I ought to want others not
to interfere with my generic features of agency, because I could not be interfered with in the
first place.
In addition, it has to be presupposed that other agents have the capacities or powers to
hurt me and to support me with respect to the generic features of agency. Given that ‘ought
implies can’, it has to be presupposed that other agents have these capacities in order for it
to be necessary for me to want others not to interfere with my generic features of agency. It
is not dialectically necessary for me to claim that the sea ought not to interfere with me,
even though the sea can definitely undermine (or increase!) my well-being, because the sea
simply isn’t the sort of thing that can interfere with me in this way. Given these
uncontroversial anthropological assumptions and given that I ought to want to have the
generic features of agency, it follows that it is dialectically necessary that I ought to want
others not to interfere with my generic features of action.10 This is the third premise:
3) It is dialectically necessary that I hold that others ought not to interfere with my
generic features of agency.11
It is crucial to note that this ought statement is made from the internal perspective of the
agent. At this point of the argument, no claims are made about the reasons for action of
other agents, i.e. that others also have to accept that they ought not to interfere with my
10
I think that these assumptions are uncontroversial for the following two reasons. First, although some of the
assumptions might seem controversial when they are understood as assertoric statements (such as the existence of
other minds), they are not controversial from the perspective of the practical self-understanding of the agent.
Second, and more importantly, all the assumptions I have mentioned above seem to be assumptions that need to
be in place for morality in general to make sense, and not just for this specific type of moral theory. That is, if a
plurality of vulnerable and mutually dependent beings did not exist, morality would not make sense in the first
place. I will come back to this in the next chapter.
11
Beyleveld adds the following condition: “unless (and only unless) I am willing to suffer generic damage to my
ability to act” (Beyleveld 2013, 214). His reason for adding this condition is “because not all [generic conditions of
agency] are necessary for the very possibility of action” (Beyleveld 2013, 214f30). I will neglect this complication in
the current discussion and will come back to it in chapter 9 when I discuss Gewirth’s idea about the generic
features of agency in more detail.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
223
generic features of agency. The claim here is thus only that it is dialectically necessary for
me to accept that others ought not to interfere with the generic features of agency. It is a
further question whether other agents should also accept this claim. Gewirth therefore
stresses that “it must be kept in mind that the agent here proceeds from within his own
standpoint of purposive agency, including the entitlements called for by that standpoint”
(Gewirth 1978, 80). According to Gewirth, premise 3 leads to the following premise:
4) It is dialectically necessary that I claim a right to the generic features of agency.
The fourth premise might sound puzzling at first, because how could one claim a right to
something that is just categorically instrumentally necessary for you to accept? It is less
puzzling, however, considering that Gewirth takes this rights claim to be simply equivalent
to premise 3, i.e. that it is dialectically necessary that I hold that others ought not to
interfere with my generic features of agency. Gewirth writes that saying that other persons
should refrain from interfering with another person’s necessary goods “is equivalent to
saying that [s]he has a right to them” (Gewirth 1978, 80). This equivalence is justified
“because the agent holds that other persons owe him this strict duty of at least noninterference” (Gewirth 1978, 80). Gewirth claims that if the agent denies that he or she has a
claim right to freedom and well-being, i.e. if he or she would not object to others interfering
with his or her freedom and well-being, he or she cannot genuinely hold that his or her
freedom and well-being are necessary goods. He states that “[by denying that he or she has
a claim right to freedom and well-being] the agent would have to accept (3) ‘It is not the
case that all other persons ought at least to refrain from interfering with my freedom and
well-being.’ But how can any agent accept (3) and also accept (4) ‘My freedom and wellbeing are necessary goods’?” (Gewirth 1978, 80). Denying this right therefore constitutes a
self-contradiction. Claiming a right to the generic features of agency thus simply means
claiming that others ought not to interfere with your generic features of agency. These
rights are therefore neither moral nor legal rights, but should be understood as categorically
instrumental rights, and ultimately as dialectically necessary rights 12 because it is
12
Cf. “ultimately the generic rights are dialectically necessary rights not prudential ones” (Beyleveld and Bos 2009,
16).
CHAPTER 7
224
dialectically necessary to accept the principle of instrumental rationality.13
Again, it is important to stress that these claims are all made from the perspective of the
agent (Gewirth 1978, 80; Gewirth 1988, 251; Gewirth 1997, 173; See also Beyleveld 1991,
25). The agent necessarily has to claim a right to the generic features of agency from his or
her own perspective as a purposive agent (this is what it means to use the method of
dialectical necessity), but it has yet to be shown that other agents ought to respect this right
or that the agent ought to respect the rights of other agents. This has to be shown in the
second step of the argument.
This concludes the first step of Gewirth’s argument, the argument for categorical
normative commitments. It might be helpful to briefly compare this argument with
Korsgaard’s argument for the value of your own humanity. In the previous chapter, I have
argued that on the most charitable reconstruction of Korsgaard’s argument, the argument
from the source of reasons, Korsgaard fails to justify any normative commitments, or at
least she fails to explain how the value of humanity translates into norms of actions,
especially in light of the fact that, according to Korsgaard, we cannot not act on the basis of
reasons (which is what it means to be part of ‘humanity’, at least on Korsgaard’s argument).
It should have become clear that Gewirth’s argument does not have these problems.
Gewirth’s argument works by exploring the implications of the principle of instrumental
reasoning, understood from the first-person perspective of a purposive agent. Although one
cannot not be an agent, one can obviously fail to claim a right to the necessary means to
one’s end, and one can fail to act in such a way as to respect this right (which becomes clear
in the context of interpersonal morality) – especially when considered that, according to
Gewirth, different generic features of agency have different degrees of necessity for agency
(Gewirth 1978, 52) (I will come back to this in chapter 9). And insofar as there are not only
means that are contingent on one’s particular end but also generic means (the generic
features of agency) which are needed to pursue any end whatsoever, one categorically ought
to claim a right to these generic means to agency. I therefore think Gewirth succeeds in
justifying categorical normativity where Korsgaard fails.
13
Gewirth refers to these oughts and rights as ‘prudential oughts’ and ‘prudential rights’ (Gewirth 1978, 70–71).
The label ‘prudential’ is, however, slightly misleading because the oughts are not just any prudential oughts, but
only those prudential oughts that it is dialectically necessary for the agent to accept (Gewirth 1978, 71). That is,
Gewirth’s prudential oughts refer to those oughts that any agent necessarily has to accept independent of his or her
specific purposes. I follow Beyleveld in using the terminology of ‘categorical instrumental oughts’ (and categorical
instrumental rights) in order to prevent misunderstanding (see for instance Beyleveld 2013, 209 where Beyleveld
defines prudential in terms of categorical instrumentality). I will come back to this below.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
225
If it could be shown that it is not only dialectically necessary for the agent to claim a
right to his generic features of agency but also necessary to assign a right to the generic
features of agency of all other agents, Gewirth would thus succeed in justifying a categorical
and universal principle of interpersonal morality. In the next section, I will therefore
discuss Gewirth’s argument for the universality of a right to the generic features of agency:
the argument from the sufficiency of agency. I understand this to be the most controversial
step of the argument.
4. Part 2: Universal Rights to the Generic Features of Agency
The fact that Gewirth’s argument proceeds from the first-person perspective of an agent
makes it seem mysterious how the argument could ever lead to normative requirements on
behalf of other agents.14 In this section, I analyse Gewirth’s argument for a universal right to
the generic features of agency. In the next section, I will briefly revisit the logical objection
to transcendental arguments from the first person, according to which there is a gap
between what an agent ought to want for him- or herself (the necessary means to her ends)
and what other agents are required to do for him or her or what an agent is required to do
for other agents
Just like Korsgaard’s argument from the first person that I have discussed in the
previous chapter, Gewirth relies on the argument from the sufficiency of agency to argue
for interpersonal morality. In the previous chapter, I assumed that the argument from the
sufficiency of agency is valid and I subsequently argued that even if the argument from the
sufficiency of agency is valid, Korsgaard fails to justify interpersonal morality. Now that we
have seen that Gewirth succeeds in justifying a categorical reason to claim a right to the
necessary conditions of agency, we should evaluate whether the argument from the
sufficiency of agency can justify a universal right to the generic features of agency. I will
first briefly describe Gewirth’s argument and then I will argue why I think the argument
succeeds, which, I think, becomes clear once it is accepted that it is dialectically necessary to
accept the principle of instrumental rationality.
14
Objections along this line have been put forward by (MacIntyre 1985, 2nd:80; Williams 1986, 61; McMahon
1986; Korsgaard 1996, 133; Habermas 1990a, 101; Kramer and Simmonds 1996; Kramer and Simmonds 1998;
Chitty 2008). For Gewirth’s response to McMahon see (Gewirth 1988). For his response to Kramer and Simmonds
see (Gewirth 1997). For Beyleveld’s response to Chitty see (Beyleveld and Bos 2009). For his response to Williams
see (Beyleveld 2013).
CHAPTER 7
226
How does Gewirth argue from the judgement that I ought to claim a right to the
generic features of agency to the judgement that all agents have a right to the generic
features of agency? Gewirth’s argument proceeds as follows: first, Gewirth puts forward ‘the
criterion of relevant similarities’, which goes as follows:
[I]f some predicate P belongs to some subject S because S has the property Q
(where the ‘because’ is that of sufficient reason of condition), then P must also
belong to all other subjects S1, S2,..,Sn that have Q. If one denies this implication
in the case of some subject, such as S1, that has Q, then one contradicts oneself.
For in saying that P belongs to S because S has Q, one is saying that having Q is a
sufficient condition of having P; but in denying this in the case of S1, one is saying
that having Q is not a sufficient condition of having P (Gewirth 1978, 105).
The criterion of relevant similarities thus states that if Q is a sufficient condition of having
P, one should accept that any Q has P.
Subsequently, Gewirth applies this criterion to the argument discussed in the previous
section. Gewirth claims an agent has to claim a right to the generic conditions of agency
simply by virtue of understanding themselves as an agent (and not, for instance, by virtue
of having certain particular purposes) (Gewirth 1978, 109). He writes that
[agency is] a sufficient condition of the justifying reason he must adduce for his
having the generic rights. If the agent were to maintain that his reason must add
some qualifying restriction to this description, and must hence be less general than
his simply being a prospective purposive agent, then he could be shown to
contradict himself. Let us designate by the letter D such a more restrictive
description. Examples of D would include, “My name is Wordsworth
Donisthorpe,” “I am highly intelligent (or benevolent),” and other descriptions ...
Now let us ask the agent whether, while being an agent, he would still hold that he
has the rights of freedom and well-being even he were not D. If he answers yes,
then he contradicts his assertion that he has these rights only insofar as he is D ...
But if he answers no, that is, if he says that while being an agent he would not hold
that he has these rights if he were not D, then he can be shown to contradict
himself with regard to the generic features of action. For, as we have seen, it is
necessarily true of every agent both that he requires freedom and well-being in
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
227
order to act and that he hence implicitly claims the right to have freedom and wellbeing (Gewirth 1978, 110).
Agency is thus a sufficient condition of claiming a right to the generic features of agency,
because one has to claim this right simply by virtue of being an agent and not by virtue of
any particular characteristic (I briefly come back to this below). This leads to the following
premise:
5) It is dialectically necessary to recognize that purposive agency is a sufficient
condition of claiming a right to the generic features of agency.
Combining the criterion of the relevant similarities with the idea that agency is a sufficient
condition of claiming a right to the generic features of agency leads to the next premise:
6) If it is dialectically necessary to accept that if agency is a sufficient condition of
having a right to the generic features of all agents, it is also dialectically necessary
to accept the claim that all agents have a right to the generic features of their
agency (this is the argument from the sufficiency of agency (Gewirth 1978, 110)).
This, according to Gewirth, is the same as saying that it is dialectically necessary to accept
the principle of generic consistency:
7) Principle of generic consistency: “act in accord with the generic rights of your
recipients as well as of yourself” (Gewirth 1978, 135).
Gewirth’s argument for interpersonal morality can be roughly summarized as follows: if
one necessarily has to claim a right insofar as one understands oneself as an agent, one
necessarily has to accept, assuming the criterion of relevant similarities is true, that not only
oneself but also any other agent has a right to the necessary preconditions of agency. This
follows because your self-understanding as an agent is a sufficient condition of claiming
this right. The consequence of this argument is that you have to accept that no one should
interfere not just with your freedom and well-being but also with the freedom and wellbeing of other agents. In other words, you have to respect the freedom and well-being of
other agents. Because the same argument applies to the first-personal standpoint of any
CHAPTER 7
228
agent, all agents necessarily have to respect the freedom and equality of all other agents.
This implies that other agents also have a reason to respect your freedom and well-being
(insofar as you are part of ‘all’ agents). This last step is crucial because it shows how, from
the first-person perspective of agents, moral reasons can be justified: the first step is to
understand that you have to accept from your own first-person perspective that all agents
have rights. The second step is to see that other agents (from their first person prescriptive)
also have to grant that all agents have rights.
The justification of moral rights is thus accomplished by claiming that any agent, from
his or her own first-person perspective, should ascribe equal rights to all other agents.
Through this universalization step, the ‘instrumental’ rights (premise 4) are thus
transformed into moral rights because all agents have to respect the freedom and well-being
of all other agents. These rights are moral rights because they hold universally (i.e. for all
agents) and categorically (i.e. independent of one’s contingent desires but as a necessary
precondition of having any purpose whatsoever).
How successful is this argument? I think that whether or not the argument from the
sufficiency of agency works to justify this conclusion crucially depends on the reason why
exactly it is necessary to claim a right to the generic features of agency for oneself (premise
4). If one does indeed claim this right simply by virtue of being an agent (premise 5), then
the conclusion seems to follow from the argument. If, however, the justification for premise
4 refers to a particular characteristic of the agent which is not shared with other agents, the
conclusion does not follow.
Most, if not all objections, to this stage of Gewirth’s argument (i.e. the argument from
the sufficiency of agency) state that the universalization of the right to the generic features
of agency fails because the first stage of Gewirth’s argument essentially refers to the selfinterest of the particular agent.15 There are two variants of this objection. On the first
variant, critics construe Gewirth’s argument as claiming that one ought to claim a right to
the generic features of agency because it is in one’s (generic) self-interest to do so. On this
purely prudential reading of Gewirth’s argument, it does not follow that one ought to grant
a right to the generic features of another agent, because it is not in one’s (generic) self-
15
Several variants of this objection can be found in MacIntyre (1985, 2nd:80), Williams (1986, 61), McMahon
(1986), Korsgaard (1996, 133), Habermas (1990a, 101), Kramer and Simmonds (1996 & 1998) and Chitty (2008).
Gewirth already anticipated this kind of objection in Reason and Morality (Gewirth 1978, 112–27). To my mind,
the most convincing response to this kind of objection is provided by Beyleveld and Bos (2009). I briefly outline
this response below.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
229
interest to respect the generic features of another agent. The second variant of this
objection denies that agency is a sufficient condition of claiming a right to the generic
conditions of agency. According to this objection, one only has to claim a right to the
generic features of agency by virtue of being the particular agent one is. I will briefly discuss
both objections in turn.
Let me start by discussing the purely prudential interpretation of Gewirth. In The
Sources, for instance, Korsgaard describes Gewirth’s argument as follows:
[T]he individual is thought to be self-interested, or, at least, self-interest is taken to
be an uncontroversial source of rational norms. Rational justifications of morality
must then show that self-interest gives the individual some reason to participate in
a moral system (Korsgaard 1996, 132).16
Bernard Williams describes the first step of Gewirth’s argument as follows:
Since I necessarily want my basic freedom, I must be opposed to courses of action
that would remove it. Hence I cannot agree to any arrangement of things by which
others would have the right to remove my basic freedom. So, when I reflect on
what arrangement of things I basically need, I see that I must claim a right to my
basic freedom. In effect, I must lay it down as a rule that they respect my freedom
(Williams 1986, 59–60).
Both Korsgaard and Williams thus seem to think that the agent has a prudential reason for
claiming a right to the generic features of agency. See also Kramer and Simmonds, who
claim that “Gewirth relies exclusively on prudential rights and duties” (Kramer and
Simmonds 1996, 313). Chitty stresses the “fundamentally prudential character of the basic
argument [i.e. the first part of Gewirth’s argument]” (Chitty 2008, 20). Illies also
presupposes the purely prudential interpretation when he writes that
the rationale behind the prohibition [i.e. the prohibition on others to interfere
with the generic conditions of agency] appears to be that I consider any negative
16
This is Korsgaard’s description of the project of Hobbes and Gauthier. But she claims that “some neo-Kantian
justifications [among which Gewirth] proceed, or anyway might be thought to proceed, in a similar way
(Korsgaard 1996, 132–33).
CHAPTER 7
230
interference to contradict my necessary value judgement. But why should there be
a clash? The value judgement I must make is simply not explicit about whether it
is a value only for me or also for others; its scope (and hence the scope of the
obligation) remains crucially underdetermined ... Gewirth has not shown a
contradiction with respect to my agency when I claim that my freedom is a
necessary value for me but not for others (Illies 2003, 122–23).
The question is thus whether or not the reason for claiming a right to the generic features of
agency is first and foremost a prudential reason. Beyleveld and Bos recently responded to
this kind of objection by claiming that “the epistemological status of my rights claim is ...
neither merely nor essentially prudential, but (for me) also and essentially dialectically
necessary” (Beyleveld and Bos 2009, 3; See also Beyleveld 2015, 584). 17 The reason for this
is that it is dialectically necessary to accept the instrumental principle insofar as one
understands oneself as an agent. So although the claiming of a right to the generic features
is prudentially or categorically instrumentally required insofar as one pursues any purpose
whatsoever, the judgement that one categorically instrumentally ought to claim a right to
the generic features of agency is itself not prudential or instrumental but dialectically
necessary (cf. Beyleveld 2015, 594). In other words, although it is true that the agent has a
(categorical) instrumental reason to claim a right to the generic features of agency, this
judgement should be understood in the context of the method of dialectical necessity.
Denying that one has a right to the generic features of agency is thus not only
prudentially or instrumentally irrational; it also contradicts the idea that one understands
oneself as an agent, insofar as accepting the principle of instrumental rationality is
dialectically necessary. In other words, it is both instrumentally and dialectically necessary
to claim that one has a right to the generic features of agency. It is therefore incorrect to
think that, according to Gewirth, there are exclusively prudential reasons to claim a right to
the generic features of agency.
Beyleveld and Bos subsequently stress that the dialectical necessity of claiming a right
to the generic features of agency shows that it can be successfully universalized:
17
To at least some extent, Gewirth himself is responsible for this purely prudential reading by talking about
“prudential rights to freedom and well-being” (Gewirth 1978, 95). Gewirth, however, uses ‘prudential’ in a very
specific sense as referring to claims that have to be accepted from the first person as dialectically necessary claims
before it is shown that others have the same rights. Critics of Gewirth, however, typically understand prudential in
terms of self-interest (which indeed comes closer to the everyday usage of the word). I try to clarify the relation
between prudence and dialectical necessity in the text.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
231
[W]hile I have two sets of reasons for considering that I have the generic rights, I
only have one set of reasons for granting the rights to others. I must claim the
generic rights for myself in my own self-interest and to avoid contradicting that I
am an agent. On the other hand, I must grant the generic rights to others only in
order to avoid contradicting that I am an agent. In other words, the generic rights
of other agents are for me not prudential but they are dialectically necessary
(Beyleveld and Bos 2009, 13).
Denying that other agents have a right to the generic features of their agency contradicts
my self-understanding as an agent, because if I am an agent I also necessarily have to accept
that agency is a sufficient condition of claiming a right to the generic features of agency. It
is thus not prudentially or instrumentally necessary to respect the rights of other agents, but
it is dialectically necessary for me to respect their rights insofar as I have to accept the
criterion of relevant similarities and insofar as I have to accept the fact that agency is a
sufficient condition of claiming a right. Beyleveld therefore stresses that “it is irrelevant to
the validity of the Gewirthian argument that, for instance, ‘Albert [or any other agent] does
not necessarily want Brenda to pursue/achieve her chosen purposes or that he does not
categorically need her to have the [generic features of agency] in order to achieve his
chosen purposes’” (Beyleveld 2015, 584).
The argument from the sufficiency of agency therefore succeeds if agency is a sufficient
condition of having a right to the generic conditions of agency and once it is acknowledged
that it is dialectically necessary to accept the principle of instrumental reason. Once it is
acknowledged that one not only has a prudential reason to claim a right to the generic
features of agency but that this judgement is dialectically necessary, it follows that one has
to respect the rights to the generic features of agency of all other agents.
This brings me to the objection that agency is not a sufficient condition of claiming a
right to the generic features of agency (Williams 1986, 66–70; Scheuermann 1987, 304).
Note that Gewirth’s argument can only work if it is indeed agency as such that is a sufficient
condition of having a right to the generic features of agency, and not someone’s particular
agency. If one has to claim a right to the generic features of agency by virtue of being the
particular agent that one is, the universalization would only result in the conclusion that
anyone who is the particular agent that he or she is has a right to the generic features of
CHAPTER 7
232
agency, which would amount to saying that only you have a right to the generic features of
agency.
The question is thus why one has to claim a right to the generic features of agency by
virtue of being an agent and not by virtue of being the particular agent one is. Gewirth
anticipates this objection, which he refers to as the ‘individualizability objection’ in Reason
and Morality. He describes the objection as follows:
[T]he individualizability objection holds that when the agent gives a sufficient
justificatory reason for his having the generic rights, this reason may be so
individualized that it pertains only to the agent himself. The point is that since the
agent claims the generic rights for himself, his reason for this claim must consist in
his own having of his own purposes that he wants to fulfill for himself. Hence, as
formulated by the agent, his reason for his right-claim must include “egocentric
particulars” – expressions whose denotation is relative to the speaker himself, such
as “I” and “my” ... Thus the universalization of the agent’s right-claim would be
not, “All prospective agents who have purposes they want to fulfill have the
generic rights,” but rather, “All prospective agents who have my purposes that I
want to fulfill have the generic rights” (Gewirth 1978, 115–16).
Gewirth’s and Beyleveld’s reply to this objection is basically to claim that one can only be a
particular agent if one is also a purposive agent, i.e. that one can only be a particular agent
insofar as one also conceives of oneself as an agent as such (Gewirth 1978, 116–19).
Beyleveld, for instance, writes: “Albert can only think of himself as the particular agent he is
by also thinking of himself as just one member of the class of beings who stand in the
universal relation to their own contingently chosen purposes of necessarily valuing them”
(Beyleveld 2013, 220; Beyleveld 1991, 288–300). The point of this reply is thus not to deny
that the argument proceeds from the perspective of a particular agent, but to stress that one
can only understand oneself as a particular agent by also understanding oneself as an agent
as such. Beyleveld therefore writes: “The argument proceeds in terms of Albert’s selfreflection on what it is for him (the particular being he is who does X voluntarily in order to
pursue his chosen purposes) to be an agent (a particular being who does X in order to
pursue his chosen purposes)” (Beyleveld 2013, 212). And because the method of dialectical
necessity shows that one has to claim a right to the generic means of agency independent of
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
233
the contingent content of one’s purposes, one has to assign the same right to every other
agent, i.e. to any other member of the class of beings who voluntarily pursue ends.
The relation between understanding oneself as the particular agent one is and as an
agent deserves more discussion than I can provide here. But let me conclude this brief
discussion by saying that I find it hard to see how one could be the particular being who
voluntarily pursues purposes while denying that one is a kind of being who voluntarily
pursues purposes. This means that the argument does not work by abstracting away from
the particular being one is, but by analysing what it means to understand oneself as the
particular being one is. However, at the same time, it is hard to see how one could argue for
this conclusion, instead of simply denying what the objection affirms (which also raises the
question of the burden of proof). For this reason, I think that more discussion is needed
about the relation between particular agency and agency, but in what follows, I will assume
that agency is a sufficient condition of claiming a right to agency.
5. The Logical Objection Revisited
The discussion in the previous section suggests that it is possible to justify a universal
principle of interpersonal morality on the basis of a transcendental argument from the first
person. The logical objection, i.e. the objection that a transcendental argument from the
first person cannot lead to any principle of interpersonal morality, seems to derive much of
its force from mistakenly reducing transcendental argumentation to prudential reasoning.
Once it is acknowledged that it is not just a prudential requirement to claim a right to the
generic features of agency but that this is (also) dialectically necessary, there is nothing
mysterious about the universalization of this rights claim.
So what should we make of Korsgaard’s objection in The Sources that a transcendental
argument from the first person tries to justify ‘public reasons’ on the basis of ‘private
reasons’ and that it is therefore flawed from the very start (Korsgaard 1996, 133)? Recall
that Korsgaard states that an agent has a private reason if the reason only has normative
force for the individual who has this reason. An agent has a public reason if the reason also
has normative force for others. Subsequently, Korsgaard argues that the problem with
transcendental arguments from the first person is that they try to argue from private
reasons to public reasons, stating that:
CHAPTER 7
234
consistency can force me to grant that your humanity is normative for you just as mine
is normative for me. It can force me to acknowledge that your desires have the status of
reasons for you, in exactly the same way that mine do for me. But it does not force me
to share in your reasons, or make your humanity normative for me. It could still be true
that I have my reasons and you have yours, and indeed that they leave us eternally at
odds (Korsgaard 1996, 134).
If Korsgaard is correct in suggesting that a transcendental argument from the first person
presupposes that reasons are essentially private, the argument would necessarily fail
because in principle a private reason cannot be a public reason.
As should have become clear by now, a transcendental argument from the first person
does not, however, try to argue from private reasons to public reasons. In response to
Korsgaard, Beyleveld writes:
The Gewirthian argument does not contend that Albert’s private reasons “logically
commit [him] to taking other people’s reasons into account ... starting from the
assumption that reasons for action are private.” It presumes neither that reasons
for action are necessarily private nor that they are necessarily public, but
demonstrates that dialectically necessary normative commitments, those that are
categorically binding on agents (and only such commitments) provide necessarily
public reasons for action. Consistency with the idea that Albert is an agent as such
requires him to treat Brenda’s dialectically normative commitments as his own
(Beyleveld 2015, 579).
Korsgaard’s dichotomy between private and public reasons should thus be rejected, at least
if private reasons are understood as those reasons which cannot in principle lead to public
reasons. First, this distinction simply postulates, without argument, that only an account
according to which reasons are intrinsically public can justify interpersonal morality,
because by definition private reasons cannot lead to public reasons. Second, Korsgaard does
not explain why Gewirth’s argument, or any other transcendental argument from the first
person, is necessarily committed to private reasons. In other words, if we accept
Korsgaard’s framing of the debate in terms of private reasons versus public reasons, it
might seem mysterious how a transcendental argument from the first person could ever
lead to a principle of intersubjective morality. But I do not think that this framing has to be
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
235
accepted in the first place.
So what would be needed in order to provide a logically valid argument for a
categorical and universal principle of interpersonal morality on the basis of a
transcendental argument from the first person? Summarizing the two steps of the
argument, I think the success of a transcendental argument from the first person ultimately
hinges on the acceptance of the following three claims:
1) It is dialectically necessary to accept the principle of instrumental reasoning.
2) There are not only means that are contingent on pursuing a specific end but also
means that are necessary to engage in any action whatsoever (the generic features
of agency).
3) One has to assign a right to the necessary means of agency to any agent, because
agency is a sufficient reason for ascribing a normative status to the necessary
means of agency (cf. Beyleveld 2016, 20).
Although I have not provided a comprehensive defence of these premises, I hope to have
shown that a transcendental argument from the first person is at least much more plausible
than is often assumed when it is understood in this way. The reason for this is that, taken by
themselves, all three premises seem to be plausible. Although the instrumental principle is
not uncontroversial, most critics of transcendental arguments from the first person accept
it (see also chapter 4). The reason for this is that the cost of denying the instrumental
principle is scepticism about practical reasoning (Skidmore 2002, cf.), which is an excessive
cost few are willing to pay. In addition, it seems plausible enough that there are at least
some generic features, such as being alive. The next question, however, is what exactly the
content of these generic features is. This question will be taken up in the next chapter.
Finally, although the last premise seems to be the most controversial premise, I have tried
to show that it follows from applying the criterion of relevant similarities to the other two
premises. In order to argue against a transcendental argument from the first person one
thus has to deny the instrumental principle, the existence of generic features of agency or
the criterion of relevant similarities (and thereby the argument from the sufficiency of
agency). Although much more needs to be said in defence of these premises, one cannot
simply claim that it is logically impossible to justify a principle of interpersonal morality on
the basis of a transcendental argument from the first person.
Let me end this section by stressing that this line of argumentation is not necessarily
CHAPTER 7
236
exclusive to Gewirth and his specific conception of the generic features of agency. The
argument from the sufficiency of agency works independent of one’s conception of the
necessary means of agency. According to Gewirth, the necessary means of agency are
freedom and well-being. The first person structure of this argument could, however, be
used to argue for different moral theories, which might also include more social or
interpersonal conditions of possibility among the necessary means of agency. These
theories, however, would differ not in the kind of argument they make but in the substance
of the argument: what, exactly, the agent requires in order to understand him- or herself as
an agent. In fact, if my criticism of transcendental arguments from the second person is
valid (chapter 5), these arguments have to presuppose such a transcendental argument
from the first person to lead to a categorical and universal principle of interpersonal
morality.
For now, let me conclude that it is possible to justify a principle of intersubjective
morality according to which every agent has a right to whatever are the necessary
preconditions of agency on the basis of a transcendental argument from the first person.
However, even if one is ultimately convinced that a transcendental argument from the first
person, such as Gewirth’s argument from purposive agency, can logically justify
interpersonal morality, one might still think that there must be something wrong with this
kind of argument. I will turn to this worry now.
6. The Wrong Kind of Reason Objection Revisited
As I have already mentioned in chapter 5, one might think that a transcendental argument
from the first person justifies interpersonal morality for the wrong kind of reason.
Although it is difficult to pin down this objection, it seems to go back to the intuition that
there must be something odd about an argument for interpersonal morality on the basis of
how I understand myself. Of course, thinking that there is something odd about an
argument is not itself an objection to an argument. In this section, I therefore briefly
discuss the question of whether there really is something to this idea that transcendental
arguments from the first person justify interpersonal morality for the wrong kind of reason.
The difference between the logical objection and the ‘wrong kind of reason’ objection is
that whereas the logical objection claims that there is no reason for being committed to a
universal principle of interpersonal morality, the wrong kind of reason objection accepts
that there might be ‘a’ reason for this conclusion, but claims that it is a reason of the wrong
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
237
kind. Now that we have seen how exactly a transcendental argument from the first person
justifies a principle of intersubjective morality, we can assess the plausibility of the wrong
kind of reason objection.
Recall that, according to Darwall, wrong kinds of reasons are reasons that refer, for
instance, to the desires of individuals (Darwall 2006, 17) or to “the agent’s valuing an
outcome” (Darwall 2006, 211). Right kinds of reasons, on the other hand, are described as
follows:
To be a reason of the right kind, a consideration must justify the relevant attitude
in its own terms. It must be a fact about or feature of some object, appropriate
consideration of which could provide someone’s reason for a warranted attitude of
that kind toward it (Darwall 2006, 16).
As I have already mentioned in chapter 5, Darwall does not explicitly address the question
of why exactly a transcendental argument from the first person for morality would provide
the wrong kind of reason for morality. His wrong kind of reason objection is mainly
addressed to Humean accounts of morality which justify morality by reference to the
contingent desires of individuals. About Humean accounts of morality, Darwall writes that
“considerations of the desirable are reasons of the wrong kind to justify claims of justice or
to warrant anything as a norm of justice” (2006, 185).
Although Darwall does not explain why transcendental arguments from the first
person provide the wrong kind of reason, he claims, positively, that only a second-personal
reason can be a right kind of reason for a moral claim. For instance, in response to an
objection raised by Korsgaard about the optionality of the second-person standpoint,
Darwall writes: “The way we get into the second-person standpoint is not by seeing a nonsecond-personal reason for doing so and then taking it up. These would not give us reasons
on which we could genuinely take it up anyway. They would be ‘reasons of the wrong
kind’” (Darwall 2007, 59).
As I have already discussed in more detail in chapter 5, Darwall thus seems to believe
that all reasons are essentially second-personal and that there can be no reasons to engage
in the second-person standpoint. This implies that any moral theory which does not rely on
the second-person standpoint, including not only Humean theories but also transcendental
arguments from the first person, rely on the wrong kind of reasons. How convincing is
Darwall’s suggestion that transcendental arguments from the first person by definition fail
CHAPTER 7
238
to provide the right kind of reason for morality?
The wrong kind of reason objection seems to derive much of its initial plausibility from
the criticism of Humean accounts of morality. Perhaps Darwall is right that moral reasons
grounded in people’s contingent preferences are reasons of the wrong kind.18 But it’s not
clear that the objection simply extends to transcendental arguments from the first person.
As should have become clear by now, there are important differences between a Humean
approach to morality and a transcendental argument from the first person. First, the
argument does not focus on contingent purposes and its contingent means, but on agency
itself and the necessary means of agency. Second, and closely related, the method used is
not a method of dialectical contingency, but a method of dialectical necessity. The
argument does not focus on the claims that an agent contingently has to accept given his or
her acceptance of certain particular evaluations, but on the claims that any agent necessarily
has to accept insofar as he or she understands him- or herself as an agent. Once we
appreciate the distinction between those Humean theories and a transcendental argument
from the first person, in terms of a distinction between dialectical contingency and
dialectical necessity, it is not clear if and how the wrong kind of reason objection also
applies to the first-personal transcendental argument.
I think that whether or not a transcendental argument from the first person puts
forward a reason of the wrong kind ultimately depends on what one takes morality to be. If
we understand morality in terms of categoricity and universality (as Darwall does), we have
no reason to think that a Gewirthian account does not provide the right kind of reason for
morality. After all, it is unclear why this conception of morality would require anything
more than showing that any agent, from his or her first-person perspective, necessarily has
to accept the rights of all other agents. To claim that a transcendental argument from the
first person provides the wrong kind of reasons one would therefore have to rely on
another, more specific, conception of morality. But in that case, it is no longer clear why
that conception of morality, i.e. the conception of morality which simply rules out a
justification of morality on the basis of a transcendental argument from the first person,
should be accepted.
Darwall apparently thinks that my obligation towards other human beings is grounded
18
I say ‘perhaps’ because Darwall’s objection to Humean theories ultimately reduces to the idea that a Humean
theory cannot justify any categorical and universal moral judgements. The problem is not that the Humean
provides a reason of the wrong kind, but that she provides no reason for categorical and universal morality. I come
back to this below.
TRANSCENDENTAL ARGUMENTS FROM THE FIRST PERSON II
239
in my second-personal relationship with them, and he thinks that any claim to the contrary
is counterintuitive (provides reasons of the wrong kind). But one might wonder for whom,
exactly, this would indeed be counterintuitive. As I have tried to show in chapter 5, I think
that it is more problematic to hold that my obligation towards all other persons, such as not
to violate their rights, is grounded in my (potential) dialogue or specific I–you interaction
with a person. For one thing, it makes his or her rights dependent on our (contingent)
second-personal engagements. If my criticism of Darwall, and other transcendental
arguments from the second person in chapter 5, is plausible, it is not possible to justify
interpersonal morality, at least as understood in terms of universality and categoricity, by
reference to second-personal reasons. In fact, it does not seem out of place to think that
second-personal reasons in the way Darwall understands them are reasons of the wrong
kind.
But what, then, would make first-personal reasons for morality reasons of the right
kind? Basically, my thought is this: if moral demands apply to everyone, one would expect
that they spring from everyone. This is exactly what a transcendental argument from the
first person tries to show. So I do not even think we have to agree with the idea that it
strikes us as somehow strange that my reason to accept someone else’s rights in some
important sense must have its source in me, and all those relevantly like me.
Obviously, more can and needs to be said about the wrong kind of reason objection. As
it stands, however, I think that there is little reason to be worried about this objection.
7. Conclusion
In this chapter, I have analysed Gewirth’s argument for the principle of generic consistency
as a transcendental argument from the first person. Although I have not provided a
comprehensive defence of this argument, I have argued that, at least when it is
acknowledged that it is dialectically necessary to accept the instrumental principle, the
argument succeeds in justifying a universal principle of interpersonal morality according to
which each agent has a right to whatever are the preconditions of her agency. In addition, I
have suggested that there is no reason to think that this argument provides the wrong kind
of reason for interpersonal morality.
However, even if a transcendental argument from the first person can justify a principle
of intersubjective morality, there is another important challenge to be answered. Despite
everything I have said so far, the necessary conditions of agential self-understanding could
240
CHAPTER 7
be so trivial or minimal that they could not lead to any interesting substantive moral
requirements. The question is thus whether any substantive moral content can follow from
agential self-understanding and, if so, how exactly substantive norms of action are to be
constructed. This issue will be taken up in the next, and final, two chapters of this thesis.
Part 4: Substance
Chapter 8
The Emptiness Objection and the Appeal to
Anthropology
1. Introduction
In the previous chapters, I have tried to show that a transcendental argument from the first
person could succeed in justifying a categorical and universal principle of interpersonal
morality. According to this principle, all agents have a right to whatever are the necessary
preconditions of agency. However, up to this point it is still unclear whether such an
argument can provide any substantive moral guidance. Can the kind of supreme principles
of morality that are typically justified through a transcendental argument lead to any
concrete and substantive moral permissions, prohibitions, requirements or constraints?
This will be the topic of the final two chapters of this thesis. In what follows, I will thus
assume that it is indeed possible to justify a principle of interpersonal morality on the basis
of a transcendental argument from the first person, and subsequently I will analyse the
substantive moral potential of Kantian constructivism.
There is widespread scepticism about this substantive moral potential. Though I do not
intend to make any exegetical claims about the history of this kind of scepticism, it seems
fair to say that such scepticism goes back as far as Hegel’s criticism that Kant’s categorical
imperative is nothing more than an ‘empty formalism’. The emptiness objection can be
formulated as a variant of the potential general dilemma for Kantian constructivism: the
objection is that either the starting point of a transcendental argument is inescapable, but in
that case nothing substantive follows from the argument, or substantive conclusions do
follow from the argument – but in that case the starting point is escapable or the argument
CHAPTER 8
244
relies on additional, unvindicated premises or on what Hegel calls, “material from outside”
(Hegel 1991, 162;135).
It is impossible to provide a comprehensive discussion of the substantive moral
ambition of Kantian constructivism and a convincing response to the emptiness objection
in the remainder of this thesis. Instead, the goal of the next two chapters is to evaluate the
general strategy that Kantian constructivists have used or could use to try to argue from a
supreme principle of morality to certain substantive constraints on action. The goal of the
following two chapters is thus not to defend a specific set of moral norms of action but to
get a better understanding of the substantive ambition of Kantian constructivism and the
way in which Kantian constructivists try to justify substantive morality given their reliance
on transcendental argumentation.
The aim of this chapter is twofold. In the first part of the chapter, I analyse what it
would mean for a moral theory to be ‘empty’ or ‘substantive’ and when Kantian
constructivism would problematically rely on unvindicated premises in the construction of
substantive morality. The goal of this discussion is to get a better understanding of both the
substantive ambitions of Kantian constructivism and the scepticism about these substantive
ambitions. In the second part of the chapter, I discuss a general Kantian constructivist
strategy of relying on anthropological facts in the construction of substantive norms of
action. Although Kantian constructivists, and Kantian ethicists in general, have different
ideas about how exactly to justify substantive moral claims, there seems to exist a
widespread agreement about this general strategy. The point of this discussion is to get a
first, general idea of how Kantian constructivists aim to justify substantive norms of action,
and, more specifically, I argue that Kantian constructivism can, and indeed ought to, rely
on anthropological facts for the justification of substantive norms of action, i.e. that relying
on anthropological facts is necessary for justifying substantive norms. In the next chapter, I
will analyse whether the reliance on these anthropological facts is also sufficient for the
justification of substantive norms of action.
I should stress from the outset that the goal of these two chapters is not to discuss the
extent to which the emptiness objection applies to Kant’s ethics. Although I briefly discuss
Hegel’s emptiness objection to Kant, the goal of this discussion is merely to illustrate the
emptiness objection and its persistence in ethics. The main aim is to evaluate the extent to
which recent Kantian constructivists succeed in justifying substantive norms of action in
light of the persistent emptiness objection to Kantian ethics. In addition, as I will explain in
section 2, I will say nothing about the ‘question of judgement’, i.e. the question of how
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
245
Kantian constructivism can lead to and justify concrete judgements about specific cases.
Instead, I will focus exclusively on the justification of certain norms of action which apply
not to specific cases but to types of actions (such as a rejection of severe injury, or basic
human rights).
The structure of the chapter is as follows: first, I say something about the structure of
justification of Kantian constructivism in order to make clear how the question of
substantive morality is situated in the broader Kantian constructivist project (section 2). In
section 3, I introduce the recurring emptiness objection to Kantian ethics. I then briefly
explain how I understand the emptiness objection by discussing the question of what it
means for a moral theory to be empty or substantive and what it would mean for Kantian
constructivism to illegitimately rely on unvindicated premises or material from outside
(section 4). In section 5, I discuss the general Kantian (constructivist) strategy of relying on
anthropological facts in the construction of substantive norms of action.
2. The Structure of Justification of Kantian Constructivism
Let me start by explaining how the question of substantive morality should be located in the
broader Kantian constructivist project. In this context, it is useful to make a distinction
between at least three levels of abstraction of a moral theory:1
1) A supreme principle of morality. Examples include, for instance, the categorical
imperative or the principle of utility and various interpretations thereof.
2) Norms of action are normative rules for action that apply to a subclass or certain
types of actions. Norms of action prescribe, forbid or allow certain types of
actions, e.g. human rights concern those rights every human has simply by virtue
of being human, principles of distributive justice concern the allocation of burdens
and benefits in society, and principles of beneficence concern the duties persons
have to help others. The defining feature of moral norms of action is that they do
not concern the evaluation of particular, determinate actions or judgements, but
1
I am not suggesting that all moral theories include all three levels of abstraction. Particularists, for instance, deny
that the first two levels play any role in making moral judgements. Act-utilitarians deny the need for norms of
action because they think that the principle of utility can be applied directly to actions. Rossian intuitionists deny
the need for a supreme principle of morality, because they think that we can directly intuit norms of action and
consequently derive specific judgements from them. Nevertheless, I think the distinction between these three
levels is a useful heuristic tool to distinguish between different structures of moral theories.
CHAPTER 8
246
apply to a certain subclass of actions.2 Norms of action come in different degrees
of abstractness. They range from highly abstract principles, such as a prohibition
on injury or Rawls’ difference principle, to more specific norms of action, such as
those entailed by a human right to periodic holidays with pay.
3) Concrete judgements morally evaluate specific actions. Whereas norms of action
apply to action types, concrete judgements apply to action tokens. For example:
should we increase taxation on property? What are my duties in relation to people
living in extreme poverty? Should ‘bed, bath and bread’ be provided for asylum
seekers whose asylum application has been rejected?
When I talk about substantive morality, I am referring to both norms of action and
concrete judgements about specific actions (although, as I explain below, my focus in the
remainder of the thesis will only be on the former). As should have become clear from the
previous chapters, Kantian constructivism first and foremost tries to justify a supreme
principle of morality, i.e. a principle of morality that should necessarily be accepted by all
agents (categorical) and which expresses moral concern for all agents (universal). Kantian
constructivists, however, typically also think that this supreme principle of morality has
implications for how we ought to act, i.e. that it leads to concrete and substantive moral
permissions, prohibitions, requirements or constraints.
Gewirth, for instance, states that he tries to provide an answer to what he calls the “the
substantive question” of morality (Gewirth 1978, 3). The substantive question of morality,
Gewirth explains, concerns the question “of which interests should favorable account be
taken. Which interests are good ones or constitute the most important goods” (Gewirth
1978, 3). O’Neill claims that the principle of followability can be used to construct some
“abstract principles, [which] may be used as the framework for constructing more
determinate principles, so ultimately for guiding action and policy making and for building
institutions and practices” (O’Neill 1996, 155). Finally, as I have already briefly alluded to in
2
O’Neill claims that “norms pick out types of action that fall under specific act descriptions: it is their
propositional structure and content that make norms apt for reasoning. However, by the same token the fact that a
norm picks out a type of action means that it can offer no more than a reason for doing some action of a specified
type, and not a reason for doing a particular act of that type” (O’Neill 2007b, 394).
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
247
chapter 6, Korsgaard thinks that it follows from the universal value of humanity that
“Enlightenment morality is true” (Korsgaard 1996, 123).3
It is crucial to note that Kantian constructivists do not try to justify these kinds of
substantive norms directly through a transcendental argument. Instead, in Kantian
constructivism, the question about the justification of substantive morality can only be
answered through the application of the supreme principle of morality. According to
Kantian constructivism, justification and application are thus two different but
interdependent steps of an argument leading to substantive moral claims. First, a supreme
principle of morality is justified through a transcendental argument. Second, this supreme
principle of morality is applied to specific contexts of action in order to justify substantive
morality. This means that intermediate steps are needed to apply the supreme principle of
morality both to specific action contexts, in the case of norms of action, and to specific
actions, in the case of concrete judgements. Norms of actions and concrete judgements are
thus not directly justified through a transcendental argument, but indirectly through the
application of the supreme principle of morality which is justified with a transcendental
argument.
The structure of justification of Kantian constructivism can be contrasted with the
3
Darwall’s argument from the second-person standpoint might be one exception to this rule. Recall that Darwall
claims that a commitment to the equal dignity of free and rational persons follows from the second-person
standpoint, but it is not clear if these commitments entail any substantive restrictions on our actions. In a
revealing discussion of the practice of slavery, Darwall denies, for instance, that his theory implies that “practices
like slavery are impossible, necessarily confused, or pragmatically self-contradictory” (Darwall 2006, 268). To be
clear: Darwall claims that both the slaveholder and the slave are necessarily “committed to the presupposition that
addresser and addressee share an equal normative standing as free and rational persons” (Darwall 2006, 268). That
is, he claims that they take up the second-person standpoint and are thereby committed to its normative felicity
conditions. What he seems to deny, however, is that the practice of slavery violates the equal normative standing
of the slave as a free and equal person (the normative felicity conditions of second-personal address). The
normative felicity conditions of the second-person standpoint thus do not seem to normatively discriminate
between different kinds of social interaction nor entail any constraints on morally legitimate actions. Another
potential exception is Habermas’s discourse ethics. According to Habermas, moral theory cannot “make any kind
of substantive contribution” (Habermas 1994, 211; See also his claim that “discourse ethics does not set up
substantive orientations” Habermas 1990b, 122). The reason for this is that “moral philosophy does not have
privileged access to particular moral truths” (Habermas 1994, 211). Habermas’s criticism is specifically directed to
Kant’s ethical theory and to Rawls’ theory of justice. Habermas criticizes Kant for believing that in order to
establish the validity of a norm it is “enough for the individual to reflect on whether he can assent to a norm”
(Habermas 1990a, 67). In a similar way, Habermas criticizes Rawls for understanding ‘justice as fairness’ as the
outcome of a “theory of justice, which he [Rawls] as an expert is qualified to construct” (Habermas 1990a, 66). For
reasons to do with space, I cannot do justice to Habermas’s criticism of philosophical ethicists’ substantive
ambition. Let me stress, however, that there is a tension between Habermas’s explicit rejection of substantive
morality and his defence of certain rules of discourse (Habermas 1990a, 89). I agree with Rawls, who claims that
Habermas cannot “avoid relying on substantive content” (Rawls 2005, 424). But this obviously requires further
analysis.
CHAPTER 8
248
structure of justification of reflective equilibrium and various forms of intuitionism. 4
According to the method of reflective equilibrium, moral principles, moral norms and
moral judgements about specific actions are all justified in roughly the same way. The
method of reflective equilibrium justifies moral claims at all levels of abstraction (i.e. moral
principles, norms or judgements about specific actions) by analysing whether or not a claim
coheres with all considerations relevant to the issue at hand. As opposed to reflective
equilibrium, Kantian constructivism proposes a more complex and hierarchical structure of
justification.
Kantian constructivism can also be contrasted with various forms of intuitionism.
Following Henry Sidgwick (1981), a distinction can be made between perceptual, dogmatic
and philosophical intuitionism. According to a perceptual intuitionism, or ‘particularism’,
as it is often referred to in recent discussions, we can intuitively see whether or not a
particular action is morally right or wrong. The classical example is provided by Gilbert
Harman, who writes, “If you round a corner and see a group of young hoodlums pour
gasoline on a cat and ignite it, you do not need to conclude that what they are doing is
wrong; you do not need to figure anything out; you can see that it is wrong” (Harman 1977,
4).5 Thus, according to perceptual intuitionism, there is no need for either moral principles
or moral norms because, according to perceptual intuitionism, it is possible to directly
intuit the wrongness or rightness of a specific action. According to dogmatic intuitionists,
on the other hand, the proper objects of intuitions are norms of action, or so-called midlevel principles. A prominent example of this view is David Ross’s theory of prima facie
duties (Ross 2002). The dogmatic intuitionist thus claims that we can intuitively know the
value of, for instance, justice, beneficence or fidelity. These mid-level principles or values
are not justified by reference to a supreme principle of morality. Instead, these mid-level
principles are supposed to be self-evidently true. Subsequently, these self-evident mid-level
principles can be used to justify judgements about specific actions.
Finally, a philosophical intuitionist claims that only first principles can be discovered
through intuitions. Historically, this position is often coupled with a defence of the
principle of utility as the first principle of morality (Sidgwick 1981; de Lazari-Radek and
Singer 2014). This first principle is either applied directly to judgements about actions (act4
I take the idea of a structure of justification from Timmons (1987), although I do not necessarily agree with his
categorization of ethical theories.
5
More recent defenders of a similar view include McDowell (1998) and Dancy (2004). For an overview of different
particularist theories see Dancy (2013).
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
249
utilitarianism) or applied to actions through the formulation of certain rules or norms of
action (rule-utilitarianism).
The structure of justification of Kantian constructivism comes closest to philosophical
intuitionism insofar as they both defend a supreme, or first, principle of morality. There
are, however, important differences. First, Kantian constructivism and intuitionism
obviously have different ideas about the justification of the first principle of morality.
Whereas a philosophical intuitionist believes that a first principle is self-evident, Kantian
constructivism puts forward a transcendental argument for its first principle. Second,
Kantian constructivism should also be specifically distinguished from the act-utilitarian
variant of philosophical intuitionism. Act-utilitarians claim that the principle of utility can
be directly applied to evaluate specific actions, practices or institutions, which means that in
act-utilitarianism there is typically no room for norms of action types.
A comprehensive account of the substantive implications of Kantian constructivism
would need to explain both how substantive norms of action can be justified on the basis of
a supreme principle of morality and how, subsequently, judgements about specific actions
can be justified on the basis of these norms of action.6 In the remainder of this thesis, I will
only focus on the first question of application, because answering this question would go a
long way towards answering the emptiness objection. The second question of application,
the question of judgement, needs a treatment of its own and needs to wait for another day.7
One crucial point of clarification is in order. Although Kantian constructivists try to
justify substantive norms of action through the application of a supreme principle of
morality, they typically do not think that norms and judgements can simply be deduced
from a supreme principle of morality. That is, they reject the idea that principles and norms
can function as some kind of algorithm for action. O’Neill, for instances, stresses
throughout her work that principles and norms cannot “provide wholly definite
instructions for each context and can[not] specify exhaustively what must be done in living
up to them” (O’Neill 2007b, 399). I think it might be more helpful instead to understand
the substantive ambition of Kantian constructivism as providing a method through which
to think about substantive moral questions and as providing at least some constraint on
6
For a similar distinction between these two questions of substantive morality see O’Neill (2007b, 399).
I will thus not try to convince the perceptual intuitionist or the particularist. The reason for this is not that this
question is unimportant but because I cannot discuss the complex question of judgement within the scope of this
thesis. O’Neill has provided one of the most comprehensive accounts of moral judgements in the context of
Kantian constructivism that I am aware of (O’Neill 1996, 178–83; O’Neill 2000c; O’Neill 2001; O’Neill 2007b).
7
CHAPTER 8
250
action (O’Neill 2007b, 400). Consequently, the norms of action that are proposed by
Kantian constructivists should be understood as fallible and potentially revisable attempts
at articulating the substantive implications of a supreme principle of morality, allowing for
the fact that the content of norms of action will differ according to the specific time and
context to which the supreme principle is applied (I will come back to this below and in the
next chapter).
Having stressed the anti-algorithmic, fallible, revisable, contextual understanding of the
substance of morality, the goal of Kantian constructivism is to justify at least some norms of
action. Before discussing the way in which Kantian constructivist try to justify such norms,
I will first say more about the existence of widespread scepticism about this substantive
ambition and the challenge that this scepticism poses for Kantian constructivism.
3. The Recurring Emptiness Objection to Kantian Ethics
The emptiness objection to Kantian constructivism states that even if Kantian
constructivism succeeds in justifying a universal and categorical principle of interpersonal
morality, this is not sufficient as a justification of objective moral claims. After all, despite
everything I have said, this universal and categorical principle of interpersonal morality
might be nothing more than an empty formalism, i.e. a principle which does not lead to any
substantive constraints on action.
The emptiness objection can be formulated as a variant of the general dilemma for
Kantian constructivism. The objection is that either the starting points of a transcendental
argument are inescapable, but in that case nothing substantive follows from the argument
(I will come back to what I mean by ‘substantive’ below), or substantive conclusions do
follow from the argument – but in that case the starting point is escapable or the argument
relies on additional, unvindicated premises (I will come back to what I mean by
unvindicated premises below).
Sharon Street, for instance, criticizes Kantian constructivism by “den[ying] that the
rabbit of substantive reasons can be pulled out of a formalist hat” (2010, 370). Evan Tiffany
writes that “there may be ways of understanding agency that are inescapable, but … any
such understanding would have to be too minimal to ever generate substantive content”
(Tiffany 2012, 224). Karl Schafer claims that given the commitment of Kantian
constructivism to a “conception of agency … [that is] inescapable from the practical point
of view” (Schafer 2015a, 692), it is “difficult to see how the resulting view could possibly
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
251
deliver a vindication of anything like substantive, universal moral norms” (Schafer 2015a,
692). Onora O’Neill aptly summarizes the emptiness objection in the form of the following
rhetorical question: “what conjuring trick could build much worth having out of so little?”
(O’Neill 1996, 155).
Although these claims are highly metaphorical and typically not developed beyond this
metaphorical level, I think the general idea is sufficiently clear: critics doubt that Kantian
constructivism can justify (sufficiently) substantive constraints on action (‘substantive
reasons’, ‘substantive content’), because it operates with a very minimal or basic idea of
agency and because it justifies only formal principles of action. The idea is thus that even if
a principle of interpersonal morality can be justified through a transcendental argument,
this principle cannot provide (sufficient) substantive guidance. In addition, some suggest
that in the justification of substantive norms of action, Kantian constructivism (implicitly)
relies on certain unvindicated premises, i.e. premises which are in tension with the basic
methodological commitments of Kantian constructivism (see e.g. Tiffany 2012).
This kind of emptiness objection to Kantian constructivism is a recurring objection to
Kant’s ethics and to Kant-inspired ethics. The emptiness objection goes back as far as
Hegel’s emptiness objection to Kant. In the remainder of this section, I will briefly discuss
Hegel’s objection. The goal of this discussion is not to understand Hegel’s emptiness
objection in the context of his moral philosophy, nor to assess to what extent the emptiness
objection applies to Kant ethics. Instead, I merely use this discussion as an illustration of
the emptiness objection, and subsequently I will try to understand what exactly the
emptiness challenge amounts to in the context of Kantian constructivism. My interest here
is thus purely systematic and not exegetical.
Hegel focused specifically on Kant’s Formula of Universal Law (FUL) of the categorical
imperative, which states as follows: “act only in accordance with that maxim through which
you can at the same time will that it become a universal law” (Kant 1998, 4:421; 31). Hegel
famously suggested that the FUL is nothing more than an “empty formalism” (Hegel 1991,
162;135).
Hegel’s emptiness objection can be interpreted in different ways, but is typically
interpreted as consisting of (at least) two separate but closely related objections (See e.g.
Wood 1990, 154; Korsgaard 1985; Freyenhagen 2012, 45–46; Stern 2012b, 75). The first,
and most fundamental, objection is that, according to Hegel, the FUL cannot justify any
moral content, i.e. it cannot lead to any constraints on action because:
CHAPTER 8
252
no immanent theory of duties is possible. One may indeed bring in material from
outside and thereby arrive at particular duties, but it is impossible to make the
transition to the determination of particular duties from the above determination
of duty as absence of contradiction, as formal correspondence with itself, which is no
different from the specification of abstract indeterminacy; and even if such a
particular content for action is taken into consideration, there is no criterion
within that principle for deciding whether or not this content is a duty. On the
contrary, it is possible to justify any wrong or immoral mode of action by this
means (Hegel 1991, 162;135).
The idea here seems to be that the FUL is purely formal and that it therefore cannot lead to
any substantive norms of action, except by relying on “material from the outside”, which
Hegel seems to assume would be illegitimate in the framework of Kant’s ethics.8 I will say
more about what it could possibly mean to rely on material from outside below, but for
now I understand the idea of material from outside in terms of premises which cannot be
justified in the context of the Kantian (constructivist) framework, given its own
methodological commitments.
Hegel illustrates this objection by reference to the example in The Groundwork of
someone who finds himself needing to borrow some money but who knows that he will not
be able to repay the debt. The question is whether this person should make the promise of
returning the money, even though he knows that he will not be able to keep his promise.
According to Kant, making a false promise on the basis of the maxim that one can make a
false promise when one is in need would violate the FUL. The reason for this is that the
practice of promises could not exist if this maxim were to be universalized:
8
The empty formalism is also sometimes read as applying to the justification (and not just the application) of the
categorical imperative. Sally Sedgwick writes that “there is another version of the empty formalism critique ... This
other (and, I believe, Hegelian) version aims its attack at the supreme practical law itself. It challenges the
purported status of the law as a priori, as derived from pure practical reason rather than experience. It also calls
into question the supreme law’s universal and necessary validity” (Sedgwick 2011, 37). And Robert Stern writes:
“On the view of Hegel I want to put forward ... , Hegel’s objection to Kant may be compared with a form of
intuitionism, where this is to be understood not primarily as an epistemological doctrine (‘we know moral truths
or propositions by intuition’), but as a doctrine that rejects the idea that morality has any single highest principle,
and thus the view that there might be any ‘supreme principle of morality’ at all, whether that is the FUL, the FH, or
any other principle of a Kantian or non-Kantian kind (such as the utilitarian principle of maximizing happiness or
well-being)” (Stern 2012b, 78). I will not discuss this empty formalism objection as applied to the justification of
the categorical imperative because I take it that I have dealt with such an objection in the previous chapters.
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
253
[T]he universality of a law that everyone, when he believes himself to be in need,
could promise whatever he pleases with the intention of not keeping it would
make the promise and the end one might have in it itself impossible, since no one
would believe what was promised him but would laugh at all such expressions as
vain pretenses (Kant 1998, 32;4:322).
Kant’s idea is thus that by pointing out a contradiction between willing a maxim and being
able to universalize the maxim, it can be shown which maxims of actions are morally
prohibited and which are permitted or required.
However, according to Hegel, this argument does not work. The problem, Hegel
argues, is that although the application of the FUL might show that I cannot at the same
time want a maxim and the universalization of this maxim, this by itself does not yet show
which of the two premises has to give way: Kant assumes that if I cannot make a false
promise, because this would undermine the very possibly of promising, I should not make a
false promise. However, the other option is that the practice of promise-making has to give
way to the satisfaction of my present needs. This would be a strange outcome, of course, but
that’s precisely the point. According to Hegel, it only follows that I should not make false
promises if there is an independent reason to maintain the practice of promise-making or
to not steal money from other people. 9 In the absence of these kinds of substantive
considerations, contradiction by itself cannot determine what is wrong or right. Hegel
therefore claims that “a contradiction must be a contradiction with something, that is, with
a content which is already fundamentally present as an established principle. Only to a
principle of this kind does an action stand in a relation of agreement or contradiction”
(Hegel 1991, 162–163;135).
Hegel’s second objection pulls in the opposite direction. Whereas the first objection
claims that FUL cannot rule out any line of action, including actions which we would
normally consider to be highly immoral, the second objection claims that the FUL could
also lead to false negatives. A false negative would be a case in which something which we
would normally consider to be morally permissible or even morally praiseworthy can be
deemed to be immoral according to the FUL. One example that is mentioned by Hegel is
9
For a discussion see Freyenhagen (2012, 56–57).
CHAPTER 8
254
the maxim of giving aid to the poor. According to Hegel, this maxim cannot be
universalized. Hegel writes that
the determinacy of helping the poor expresses abolition of the determinacy that is
poverty; the maxim whose content is that determinacy, when tested by being
raised to universal legislation, would prove itself false, for it would annihilate itself.
If it is thought that the poor should be helped universally, then either there would
no longer be any poor, or there would be only poor and then there would be no
one who could help them; and so in both cases the help ceases (Hegel 1999, 127).
I understand Hegel as making the point that applying the FUL could lead to ruling out
certain actions which seem to be at least intuitively praiseworthy, such as helping the poor.
Taken together, these objections claim that the FUL cannot justify any substantive
norm of action and/or that if it can lead to a substantive norm of action it would rule out
actions that are at least intuitively praiseworthy. Again, I should stress that my goal is not to
assess the plausibility of Hegel’s empty formalism objection as applied to Kant’s ethics.10
Instead, I merely want to illustrate that the emptiness objection is a persistent objection to
Kant’s ethics that is not peculiar to recent ethics but has been constantly recurring ever
since the publication of Kant’s work. In what follows, I will abstract away from Hegel’s
specific emptiness objection to Kant’s ethics and will discuss how I understand the
emptiness objection and what it would mean to give a convincing answer to it.
10
For recent overviews of different Kantian answers to the emptiness objection see (Stern 2012b; Freyenhagen
2012; Sedgwick 2011). There are roughly three lines of responses put forward by proponents of Kantian ethics
(which are not mutually exclusive). One response is to try to show that, once properly understood, FUL does
generate constraints on actions after all (Korsgaard 1985; O’Neill 1986). The second response is to emphasize that
Hegel and other critics of Kant focus exclusively on FUL and neglect other formulations of the categorical law such
as the Formula of Humanity. Allen Wood, for instance, stresses that “Hegel and other critics will not have shown
Kantian ethics to be empty of content until they have demonstrated the emptiness of these other formulas [i.e. the
Formula of Humanity and the Formula of Autonomy] along with that of the FUL” (Wood 1990, 156). A third
response is to emphasize the ‘empirical’ aspects of Kantian ethics, i.e. to show that the justification of norms of
action can only succeed by relying on general empirical anthropological assumptions and that this does not
undermine the Kantian project (Louden 2000, 167–70; Herman 1985; Herman 2007; Formosa 2013; Smit and
Timmons 2013). I will come back to the final response below, not so much as a response to Hegel but to illustrate
the general way in which Kantian constructivism tries to justify substantive norms of action. In addition to these
three responses, a final response by Kantians is to stress that there is a tendency among critics of Kantian ethics to
focus exclusively on the Groundwork and the examples that Kant puts forward in The Groundwork and that critics
misconstrue the role of the examples in The Groundwork and neglect the derivation of duties from the categorical
imperative in the Metaphysics of Morals (see e.g. Smit and Timmons 2013; Flikschuh 2000, 3).
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
255
4. The Emptiness Dilemma
As I have tried to illustrate above, the emptiness objection comes in the form of a dilemma:
the first horn of the dilemma thus states that Kantian constructivism is empty. The second
horn of the dilemma states that it can only justify a substantive moral theory by relying on
material from outside. It is not entirely clear, however, what it would mean to a moral
theory to be (sufficiently) substantive or what Hegel and other critics mean by unvindicated
premises and ‘material from outside’. In fact, in light of how widespread this objection is, it
is rather surprising to see how little systematic reflection there is on the question of what
exactly the objection amounts to. In this section, I therefore briefly discuss both horns of
the dilemma in order make clear how I understand the emptiness objection.
4.1. First Horn: No Moral Substance
The first horn of the dilemma states that Kantian constructivism is empty. But what does it
mean to say that a moral theory is empty? Is the objection that Kantian constructivism
cannot justify any substantive norm of action (a strong emptiness objection), or that it is
not sufficiently substantive (a weak(er) emptiness objection)? Allan Wood describes the
difference between these two variants of the emptiness objection as follows in the context of
the FUL: “even if the FUL shows some actions or maxims to be wrong, it may still fall short
of providing a fully adequate account of duties if there are some cases that it fails to cover or
in which it yields the wrong results” (Wood 1990, 154–55).
I think that there is something unsatisfying about both the strong and the weak
emptiness objection. Answering the strong emptiness objection might be a bit too easy for
Kantian constructivism, because all that is needed to answer this objection is be show that
Kantian constructivism can justify some norm of action, however abstract or limited
(although it would still need to be proven that this is possible). Answering the strong
emptiness objection, in other words, is still a far cry from justifying ‘enlightenment
morality’ or the rather comprehensive moral theories that are proposed by, for instance,
O’Neill and Gewirth (see next chapter).
The problem with the weak emptiness objection, however, is that it is surely
problematic to presuppose that there is an independent standard against which to measure
CHAPTER 8
256
the (substantive) adequateness of a moral theory.11 To briefly come back to Hegel’s example
of helping the poor: of course, I agree that it would extremely counterintuitive if helping the
poor was immoral according to Kant or Kantian constructivists, and I share the intuition
that there must be something wrong with the moral methodology and the resulting moral
theory that would lead to this conclusion (although I am not convinced that FUL leads to
this conclusion (cf. Korsgaard 1985)). However, the fact that I (or we) find this result
counterintuitive is, at least on the transcendental methodology, not itself a reason to revise
the moral claim, the norm of action, just like the fact that there were (recent!) times in
history when many people had the intuition that slavery was morally acceptable, that
homosexuality was a moral sin and that women should not have certain social and political
rights is not a reason to think that slavery is morally acceptable, that homosexuality is a
moral sin or that women should not have those rights. We should, in other words, be very
careful when assessing moral theories by appealing to what we find morally
(counter)intuitive. A moral theory might be right while going against (some of) our moral
intuitions.
The potentially counterintuitive implication of a transcendental argument might
provide an incentive to carefully reconsider all the steps of the transcendental argument
that lead to this conclusion, but it is not a reason to either accept or reject a moral claim.12
In other words, referring to pre-established moral theories or to moral intuitions is
question-begging from the perspective of transcendental argumentation. In the absence of a
plausible moral methodology that can account for the epistemic status of our intuitions,
these intuitions cannot play such a fundamental role in moral theorizing.
But although the question of when a moral theory is sufficiently substantive, and how
to answer this question in the first place, deserves more attention, I think the idea
underlying the first horn of the dilemma is sufficiently clear: Kantian constructivism
promises to provide an answer to the substantive question of morality and to provide a
11
In a recent criticism of the substantive contribution of Kantian ethics, Noriaki Iwasa criticizes the categorical
imperative, among other reasons, for failing to lead to the judgement that creating graffiti and littering are
immoral actions (Iwasa 2013, 73) This criticism assumes that there is an independent moral standard that can be
applied to determine whether or not an action is immoral, whereas the standard for moral justification is precisely
the topic under consideration. The standard of evaluation cannot be whether the judgements that are generated by
the categorical imperative fit with someone’s intuitions. In addition, it assumes that a moral theory would have to
make claims about things such as graffiti in the first place.
12
In other words, although intuitions cannot play a justificatory role in Kantian constructivism, they might have a
heuristic function; intuitionism might be part of the context of discovery but not part of the context of
justification. I will come back to this in the next subsection.
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
257
framework in which to construct a wide range of rights and principles, such as human
rights or principles of justice. Although there is no independent standard against which to
measure the substantive success of Kantian constructivism, I understand the emptiness
objection, first and foremost, as raising doubts about the ability of Kantian constructivism
to live up to its own substantive ambition and promises.13 Answering the first horn of the
emptiness dilemma therefore requires showing how, exactly, substantive norms of action
are justified by reference to a supreme principle of morality. I think this is a legitimate
challenge even in the absence of a very clearly articulated idea of what it would mean to
justify a sufficiently substantive moral theory.
4.2. Second Horn: Unvindicated Premises or Material from Outside
This brings us to the second horn of the dilemma: the idea that Kantian constructivism can
only justify substantive norms of action by (implicitly) relying on unvindicated premises or
material from outside. Just like the question of what it means to justify a sufficiently
substantive theory, the question of when Kantian constructivism would rely on
unvindicated premises or material from outside, or even what material from outside refers
to in the first place, is not entirely clear.
I think it is most helpful to understand unvindicated premises or material from outside
in terms of premises which cannot be justified within the Kantian constructivist
framework. But this, of course, still leaves open the question of what kind of premises a
Kantian constructivist can legitimately rely on. One clear case of relying on material from
outside would be if a Kantian constructivist simply relied on his or her moral intuitions in
the justification of norms of action. The reason why a Kantian constructivist would rely on
material from outside if he or she would rely on moral intuitions is that in the Kantian
constructivist methodology there is no room for moral intuitions as things that play a
justificatory role. Moral intuitions are material from the outside given Kantian
constructivists’ own fundamental methodological commitments. The challenge for Kantian
constructivism, insofar as it aims to provide an alternative to reflective equilibrium and to
13
I say ‘first and foremost’ because there are also variants of the emptiness objection which focus on ‘false
negatives’ and ‘false positives’ (see e.g. Iwasa 2013). For reasons mentioned in the text, I do not think these variants
of the emptiness objection are the most interesting objections because they seem to presuppose an independent,
substantive moral standard.
CHAPTER 8
258
intuitionism, would be to show how a transcendental argument can lead to the justification
of norms of action without relying on moral intuitions.
Another way in which Kantian constructivism would rely on material from the outside
would be if its conception of agency changed between the justification of a supreme
principle of morality and its application. 14 This needs explaining. Recall that a
transcendental argument starts from an inescapable description of our agential selfunderstanding. However, because ‘agency’ can be understood in many different ways, one
might be tempted to change one’s understanding of agency between the justification and
application stage, because, obviously, it is easier to generate substantive norms of action on
the basis of a richer, but potentially optional, agency conception. Recently, Evan Tiffany
(2012), for instance, put forward an emptiness objection to Christine Korsgaard (and other
Kantian constructivists) by arguing that Korsgaard equivocates between different
conceptions of agency in the justification and application stage of her theory.15 The point
here is not to assess Tiffany’s objection to Korsgaard, but merely to note that if a Kantian
constructivist equivocates between different conceptions of agency, he or she would rely on
material from outside (this will be analysed in more depth in the next chapter).
The reliance on moral intuitions and the reliance on a so-called ‘optional’ conception
of agency are two ways in which Kantian constructivism could rely on material from
outside. These two cases would be cases of relying on material from outside because they
are incompatible with the basic methodological commitments of Kantian constructivism.
The first is incompatible with the aim of providing an alternative to both reflective
equilibrium and intuitionism. The second is incompatible with the aim of justifying
categorical moral claims.
Hegel and other critics, however, seem to go beyond these two ways in which Kantian
constructivism could rely on unvindicated premises or material from outside. Hegel
suggests that any substantive norm would necessarily have to rely on material from the
14
This material from outside could also be interpreted as a variant of relying on intuitions, because arguably the
challenge is that certain rich conceptions of agency are not inescapable (as the Kantian constructivist would claim)
but merely articulate the moral intuitions of Kantian constructivists. Nevertheless, I think it might be helpful to
distinguish these two kinds of material from outside.
15
According to Tiffany, a conception of agency cannot generate moral content because any plausible conception
of agency – e.g. the idea that an agent is someone who deliberates among alternatives and acts on the basis of
reasons – “places no restrictions on how one can exercise one’s agency” (2012, 228). His argument for this
conclusion is the familiar Humean observation that, at least on this minimal conception of agency, “what counts as
a reason is thoroughly subjective; thus we need to know something about the contingent psychology of particular
agents in order to know what reasons she has” (2012, 228–29). Tiffany therefore concludes that the idea that a
minimal notion of agency could generate substantive moral content is simply “wildly implausible” (2012, 231).
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
259
outside because FUL, as a formal principle, is not just contingently but necessarily empty.
Thus, even if Kantian constructivism would not rely on moral intuitions or an optional
conception of agency, Kantian constructivism would still have to rely on material from
outside to avoid being empty. Hegel writes about FUL in terms of “a formal identity which
necessarily excludes every content and determination” (Hegel 1991, 163;135). This gives us
an idea of what, according to Hegel, constitutes material from outside, namely any
substantive premise. It is, however, not obvious that Kantian constructivism cannot rely on
any substantive premise, even if is true that Kantian constructivism cannot rely on moral
intuitions or optional conceptions of agency. Or at least, this is something that deserves
further attention before we conclude that Kantian constructivism cannot in principle justify
any substantive constraints on action.
So even though it is not entirely clear what is meant by material from outside, it is
sufficiently clear to understand what the challenge is for Kantian constructivism. Kantian
constructivism would have to show that in its justification of substantive norms of action it
does not rely on certain premises which are clearly brought in from outside the
transcendental methodology, such as intuitions.
In this section, I have briefly discussed both horns of the emptiness dilemma by
discussing the question of what it means for a theory to be sufficiently substantive and what
it would mean for Kantian constructivism to rely on unvindicated premises or material
from outside. I have not tried to provide a comprehensive answer to these questions, but I
have tried to point out that even though it is not entirely clear what critics mean by
(sufficiently) substantive or material from outside, it is nevertheless sufficiently clear what
the challenge amounts to. The challenge for Kantian constructivism is to show how certain
substantive norms of action could be justified on the basis of a supreme principle of
morality and to show how this is possible without relying on unvindicated premises, i.e.
premises cannot be justified on the basis of the basic methodological commitments of
Kantian constructivism. My hope is that this very brief discussion of the two horns of the
dilemma regarding the emptiness objection does justice to both critics of Kantian
constructivism, who, I think, legitimately raise questions about the way in which Kantian
constructivism hopes to justify certain substantive norms of actions, and to Kantian
CHAPTER 8
260
constructivists (and Kantian ethicists more generally16), who, I think, also legitimately
complain that as it stands the emptiness objection is not a convincing objection.17
I will now analyse the way in which Kantian constructivists try to justify substantive
norms of action.
5. Kantian Constructivism and Anthropology
The question is thus how Kantian constructivists aim to justify substantive norms of action
on the basis of a supreme principle of morality. Although interpreters of Kant and Kantian
constructivists try to justify substantive norms in many different ways (I will come back to
this in the next chapter), they share the idea that in the justification of substantive norms of
action, Kantian constructivism has to rely on general, empirical premises about human
nature, which I will refer to as an ‘anthropology’. Very roughly, the idea is that relying on
anthropological facts makes it possible to move from a supreme principle of morality to
substantive norms of action and that this reliance is not the same as relying on
unvindicated premises or material from outside, but is legitimate within the Kantian
constructivist project. One crucial question is how exactly to understand the status of these
anthropological considerations. I will briefly come back to this question at the end of this
section, although I will not try to provide a comprehensive answer to this question. For
now, I will understand anthropological premises to refer to certain facts about human
nature, where facts are simply understood as descriptions of our capacities and
vulnerabilities.18 Examples of such anthropological facts are the fact that human beings are
vulnerable and dependent, i.e. that they can he harmed by others, and that they need, for
instance, oxygen to function.
In section 5.1, I will briefly illustrate how widely shared this appeal to anthropology is
among both interpreters of Kant and Kantian constructivists. In section 5.2, I will discuss
what exactly this anthropology involves and why I think an appeal to anthropology is both
legitimate and necessary within the Kantian constructivist framework. In the next chapter, I
will discuss in detail the accounts of recent Kantian constructivists against the background
16
There is no strict distinction between Kant scholars and Kantian constructivists. Both Korsgaard and O’Neill, for
instance, present their theories at the same time as interpretations of Kant and as a systematic position in recent
debates. But I think that it is roughly clear what I mean by this distinction.
17
For references see footnote 10 of this chapter.
18
I have no specific (e.g. positivist) conception of what (empirical) facts are. For current purposes, we can take
such facts to be similar to, for example, the ‘fact’ that birds can fly or the ‘fact’ that they lay eggs, etc.
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
261
of this general strategy of generating moral content on the basis of a supreme principle of
morality, and I analyse whether this appeal to anthropology is not only necessary but also
sufficient for the justification of substantive norms of action.
5.1. The Appeal to Anthropology
If it often thought that Kantian ethics is a purely a priori ethics, completely divorced from
all empirical considerations. Kant, after all, clearly states in the opening pages of The
Groundwork that it is “thought to be of the utmost necessity to work out for once a pure
moral philosophy, completely cleansed of everything that may be only empirical and that
belongs to anthropology” (Kant 1998, 2; 4:389).
Both Kant scholars and Kantian constructivists, however, claim that this is only one
side of the coin and that in the application of the categorical imperative, Kant relies on
general anthropological considerations, and rightly so. In addition, defenders of Kantian
ethics claim that relying on empirical anthropology is not just an ad hoc modification of
Kantian ethics in response to the emptiness objection, but that empirical anthropology
already plays an important role in Kant’s own ethical system.
Kant scholars refer, for instance, to passages in The Metaphysics of Morals where Kant
claims that
we shall often have to take as our object the particular nature of man, which is
known only by experience, in order to show in it what can be inferred from
universal moral principles. But this will in no way detract from the purity of these
principles or cast doubt on their a priori source. This is to say, in effect, that a
metaphysics of morals cannot be based upon anthropology but can still be applied
to it (Kant 1996, 6:217).
In addition, many theorists point to The Groundwork, in which Kant already relied on
anthropological considerations in deriving duties from the categorical imperative. Take, for
instance, Kant’s argument for a duty of beneficence (Kant 1998, 33;4:423). In arguing for
the duty of beneficence, Kant argues that even though we could, without contradiction,
imagine a world in which everyone acted according to purely self-interested maxims,
CHAPTER 8
262
a will which decided in this way would be in conflict with itself, since many a
situation might arise in which the man needed love and sympathy from others,
and in which, by such a law of nature sprung from his own will, he would rob
himself of all hope of the help he wants for himself (Kant 1998, 33;4:423).
Kant’s argument for a duty of beneficence thus relies on an empirical anthropological
consideration about human beings, more specifically the consideration that human beings
will encounter situations in which they need and want help from others. Kant’s argument is
that insofar as this claim about human neediness and interdependence is a general
anthropological fact about human beings, we necessarily have to accept a duty of
beneficence, insofar as we are not just finite rational beings, but also human beings.
Other examples can be found in The Metaphysics of Morals. Otfried Höffe, for instance,
refers to Kant’s argument for private property in ‘The Doctrine of Right’. According to
Höffe, “the institution of private property makes sense only because ‘corporeal’ human
beings need living space and goods in order to satisfy their needs and interests” (Höffe
2013, 228). Private property can thus only be justified by relying on empirical
considerations regarding the neediness and embodiment of human beings (of course,
together with the claim that private property is the best, or only way, to satisfy human
needs in light of their embodiment).
In light of these kinds of examples, Robert Louden concludes that Kant “strongly
endorses the claim that moral theory is inapplicable to the human situation (indeed, to any
situation) without a massive infusion of relevant empirical knowledge” (Louden 2000,
168).19 Louden claims that “the aim of impure ethics is ... to make these principles [i.e.
universal and categorical principles] efficacious in human life” (Louden 2000, 170). Barbara
Herman likewise concludes that “Kant does not need to argue (nor does he) that the
content of morality is to be determined without regard to the empirical nature of things”
(Herman 1984, 590). In addition, Paul Guyer characterizes Kant’s strategy of deriving
substantive norms of actions in ‘The Doctrine of Right’ in the following way:
19
Louden specifically singles out four ways in which an emphasis on the impure part of Kant’s ethics can rebut the
empty formalism charge: the role of moral education, the importance of institutional support for the functioning
of human beings, the importance of judgement in moving from abstract principles to specific case and, finally, the
importance of empirical information for the realization of moral ideals (Louden 2000, 168–70).
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
263
[F]or Kant the principles or duties of justice are derived by determining what
principles of property rights and social and political organization are required for
agents who live in our circumstances—that is, who because of their own
embodiment need some control of other physical objects, beginning with land,
and the labor of others in circumstances in which, ultimately because of the
finitude of the surface of the globe, they cannot avoid interaction with each
other—if they are to act in conformity with the moral law. With the moral law in
hand, Kant considers only the most general facts about human life in order to
derive the principles of justice (Guyer 2013, 177–78).
Finally, Smit and Timmons claim that the derivation of substantive norms of action in The
Metaphysics of Morals relies on “fundamental and thus universal empirical facts about
human beings” (Smit and Timmons 2013, 232).
The reliance on empirical anthropology is not only an important topic within Kant
scholarship; it also plays an important role in the work of recent Kantian constructivists
who likewise stress the importance of empirical anthropological facts for the justification of
substantive morality. Consider, for instance, the moral theory of Gewirth. In the previous
chapter, I have already stressed that Gewirth’s theory is meant to apply to this specific
world and that his starting point is not rational agency simpliciter, but human rational
agency. Gewirth, for instance, stresses that his “concern is to differentiate, from the many
and varied real features of human behaviors, those that constitute human action in the
relevant sense” (Gewirth 1978, 29 my emphasis). The idea of freedom and purpositivity is
put forward as a characterization of the inescapable self-understanding of human beings,
and Gewirth’s concrete normative claims, for instance with regard to rights to various
forms of freedom and well-being, clearly draw on general anthropological considerations
concerning the needs, vulnerabilities and abilities of human beings.
Similar considerations can be found in the work of O’Neill. She emphasizes throughout
her work that she starts from a realistic, empirically informed account of the capacities and
vulnerabilities of actual human beings: “the construction of ethical principles could start
from abstract but adequately accurate accounts of the ‘agents of construction’ or ‘builders’,
whose capacities and capabilities determine and whose vulnerabilities correspondingly
limit what they can build out of the ‘materials’ which they dispose” (O’Neill 1996, 62). In
other words, O’Neill relies on an “empirically realistic view of the capacities and capabilities
agents have” (O’Neill 2000d, 7). In addition, O’Neill explicitly stresses that the construction
CHAPTER 8
264
of norms of action can only succeed by including empirical anthropological premises
among the materials of construction (O’Neill 2007a, 162).
Of course, the fact that Kant, interpreters of Kant and Kantian constructivists rely on
anthropological premises for the justification of substantive moral norms does not yet show
that this reliance on anthropology is indeed legitimate (neither does it show that it is
sufficient, but this is the topic of the next chapter). So is empirical anthropology
problematic material from outside or unproblematic material from inside which can be
used to justify substantive norms of action?
5.2. Anthropology as Material from Inside
Hegel assumes that a formal principle is necessarily empty. Barbara Herman, however,
claims that we should distinguish between a principle being formal and it being empty.
According to Herman, the claim that a principle is formal does not mean that it is empty; it
only means that it does not rest on contingent materials, i.e. that it does not rest on the
desires and interests of particular agents (Herman 1996, 217). Herman therefore claims that
“purely formal principles do not have no content; they have noncontingent content”
(Herman 1996, 217). More specifically, she argues that the content of formal principles
should be provided by what is “constitutive of the kind of beings we are” (Herman 1996,
217f19) and by the “details of what we are like or the circumstances in which we act”
(Herman 2007, vii).
Herman labels the transition from the formal principle of the categorical to the
substantive norms of action the ‘middle theory’ of Kantian ethics. The middle theory is put
forward as “the translation of a formal conception of value into terms suitable to the
particular contexts of human action and deliberation” (Herman 1996, 240). Middle theory,
in other words, is the theory that applies the categorical imperative to human agents by
introducing anthropological premises.
Herman makes a distinction between different kinds of anthropological premises. First,
general anthropological premises are needed to determine the “limits and weaknesses of
human agency” (Herman 1996, 233). These include insights into the fact that there is a
plurality of interdependent and vulnerability human beings (Herman 1996, 233–36). These
general anthropological premises are non-contingent because they are among the
circumstances of morality, meaning that in the absence of these anthropological facts,
morality would have no application in the first place (cf. O’Neill 1996, 98–101). These
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
265
general facts are therefore, according to Herman, “constitutive of effective rationality”
(Herman 1996, 234). Together with the categorical imperative, they lead to abstract norms
of action, such as a rejection of maxims of coercion and deceit (Herman 1996, 233) and a
duty of beneficence (for the latter see previous subsection).
Second, contingent anthropological premises are also needed to justify more specific
norms of action. Contingent anthropological facts are those facts which are only valid in a
specific socio-historical context and which can possibly change throughout time. Herman
illustrates the relevance of these contingent anthropological premises in the context of a
rejection of the maxim of deceit and coercion:
[W]hat modes of treatment in what circumstances amount to deceit or coercion is
in part a contingent matter, conceptually dependent on our morally informed
knowledge of institutions, practices, and their effects. What is not contingent is
that if one acts on a maxim of deceit one is acting impermissibly; what is
contingent are the ways of acting that are deceitful (Herman 1996, 234).
The same point applies to any abstract norm of action: although general anthropological
premises might be sufficient to justify highly abstract norms of action, contingent
anthropological facts are needed to determine the particular conditions under which an
abstract norm of action would be violated. Although these anthropological facts are
themselves contingent, reliance on these contingent facts is necessary in order to determine
more specific norms of action. I will say more about this and will give more examples of
contingent anthropological facts in the next chapter.
Again, I will leave aside the question of to what extent this constitutes a plausible
interpretation of Kant’s moral theory. The crucial question for the current discussion is
whether Kantian constructivists can legitimately help themselves to these anthropological
premises in the application of a supreme principle of morality.
In answering this question, it is important to stress that in Kantian constructivism
general anthropological considerations already play a role in the justification of a supreme
principle of morality. In the previous chapter, I have argued that Gewirth’s transcendental
argument from the first person, or any transcendental argument from the first person, only
applies to the self-understanding of human purposive agents and that, for instance, his
claim that agents dialectically must want others not to interfere with the generic features of
their agency presupposes that there is a plurality of vulnerable and interdependent agents.
CHAPTER 8
266
If not, it would not be necessary to claim a right to the generic features of agency, because
in that case agents could not be harmed or helped by others in the first place. Imperatives
do not apply to angels; there’s no room for them to judge or act immorally.
Given that these general anthropological facts already play an important role in the
justification of the principle of generic consistency, there should be no problem in relying
on the same facts in the application of the principle of generic consistency (I come back to
the role of contingent anthropological facts below). On a Kantian constructivist account,
anthropological facts would thus only count as material from outside if one assumes that a
transcendental argument in ethics is supposed to apply to all possible worlds. But this is not
the ambition of Kantian constructivism, nor do I see a reason why this should be the
ambition of Kantian constructivism.
The reliance on general anthropological facts in the application of the supreme
principle of morality is not only compatible with its justification; it is even necessary to rely
on these facts in its application. Recall that the supreme principle of morality grants a right
to whatever are the necessary means of human agency. The very idea of necessary means of
agency already presupposes general anthropological facts: there would be no necessary
means of agency if agents were not vulnerable and interdependent, because in that case they
would be no way in which agency could either be undermined or enhanced (compare
angels). In addition, in order to know what exactly the necessary means of agency are, one
cannot but rely on anthropological facts. One can only determine what the generic features
of agency are by looking to the specific ways in human agency could either be undermined
or enhanced. This obviously cannot be done without relying on anthropological facts.
This is also where contingent anthropological facts come in. Arguably there are some
very abstract generic means of agency which can be determined on the basis of very general
anthropological facts. But in order to determine more specific norms of action one would
have to rely on contingent, and possibly social and historically relative, facts about human
capacities and vulnerabilities. The generic means of agency could, for instance, be realized
in a multitude of ways (e.g. in order to be hydrated one could drink water or fruit juice). In
addition, these means require different kinds of concrete actions depending on the sociohistorical context (e.g. freedom of movement might require different things in a modern
Western society compared to 300 years ago).
The point here is not to defend any specific necessary means of agency, nor to assess
how exactly norms of action are justified on the basis of a supreme principle of morality
through relying on anthropological facts, but only to show that if a supreme principle
THE EMPTINESS OBJECTION AND THE APPEAL TO ANTHROPOLOGY
267
which grants rights to whatever are the necessary means of agency can be justified through
a transcendental argument, it is not only legitimate but even necessary to rely on
anthropological facts in the justification of substantive norms of action (on different levels
of abstraction).
Let’s take stock. The critics of Kantian constructivism claim that Kantian
constructivism can only justify substantive norms of action by relying on material from
outside, such as moral intuitions or optional conceptions of agency. In this section, I have
tried to show that, pace Hegel, not all substantive premises qualify as ‘material from
outside’. Anthropological facts are not arbitrary and unvindicated premises but are among
the circumstances of morality, and they are not only crucial in the application of a supreme
principle of morality but also for its justification. This means that if it is indeed possible to
justify substantive norms of action by relying on anthropological facts, Kantian
constructivism can answer the emptiness objection.
Of course, this discussion raises more questions than it answers. For instance, the
reliance on anthropological facts raises the question of how exactly to understand the status
of these kinds of facts and their place in a radically first-personal justificatory strategy such
as transcendental argumentation. In addition, and closely related to this, it raises the
question of how to determine which alleged anthropological facts are truly general facts
about human beings and how these are distinguished from more contingent facts.20 These
fundamental questions deserve more attention than I can give them within the constraints
of this thesis. What I will do in what follows is assume that the reliance on anthropology is
indeed legitimate and that it is possible to determine and distinguish general facts about
human beings from more contingent facts. In the next chapter, I will discuss the question of
whether this reliance on anthropology is sufficient. The relevance of this question is that
emphasizing the importance of anthropology is, as I have tried to illustrate above, often
taken as a sufficient answer to the emptiness objection. In the next chapter, however, I will
argue that the success of this strategy hinges on the specification of the conception of the
agency in Kantian constructivism and that this question has not received sufficient
attention.
20
I discuss these kinds of questions in the context of Rawls’ theory of justice in De Maagt (2014). Rawls, for
instance, assumes, without argument, that it is a fact about human beings that they need incentives in order to be
motivated to work (this lies at the basis of Rawls’ justification of the difference principle).
CHAPTER 8
268
6. Conclusion
In this chapter, I have introduced the emptiness objection to Kantian ethics in general and
Kantian constructivism in particular. I have tried to clarify what it means for a theory to be
empty versus substantive and when Kantian constructivism would illegitimately rely on
‘material from outside’ in the justification of substantive norms of action. With regard to
the former question, I have argued that one should not presuppose a specific substantive
standard against which to measure the emptiness of a moral theory. With regard to the
latter question, I have argued that although Kantian constructivism should not rely on
moral intuitions and optional conceptions of agency, Kantian constructivism can and even
should rely on both general and contingent anthropological facts in the construction of
substantive norms of action.
In the next chapter, I will critically evaluate recent attempts at justifying substantive
norms of action on the basis of a supreme principle of morality against the background of
the discussion in this chapter. The goal of the next chapter is thus to evaluate whether the
reliance on anthropology is not only necessary but also sufficient for the justification of
substantive morality.
Chapter 9
The Emptiness Objection and Kantian
Constructivism
1. Introduction
In the previous chapter, I have discussed the emptiness objection to Kantian constructivism
and the general Kantian (constructivist) strategy of relying on both general and more
contingent anthropological facts in the justification of substantive norms of action. In this
chapter, I analyse to what extent Kantian constructivists succeed in justifying substantive
norms of action in light of the discussion in the previous chapter. I do not focus on the
details of their substantive moral theory, but rather on the way in which they try to argue
for a rather ambitious range of substantive norms of action on the basis of a supreme
principle of morality.
More specifically, I critically evaluate two different Kantian constructivist approaches
to the justification of substantive moral norms through the application of a supreme
principle of morality: Onora O’Neill’s application of the principle of followability and Alan
Gewirth’s application of the principle of generic consistency. The reason for discussing
O’Neill’s and Gewirth’s justification of substantive morality is that they have written
extensively on both the justification of a supreme principle of morality and its application. I
therefore consider O’Neill and Gewirth to have provided the most extensive defence of
substantive morality on the basis of a transcendental argument.1
1
One might wonder why I discuss O’Neill’s application of the principle of followability given my criticism of
O’Neill’s justification of the principle of followability in chapter 5. I will come back to this question at the
beginning of section 3.
CHAPTER 9
270
The conclusion of this chapter is moderately optimistic (or moderately pessimistic
depending on the way one looks at it). I will argue that Kantian constructivism can succeed
in justifying at least some substantive norms of action and that therefore the strong
emptiness objection to Kantian constructivism, i.e. the objection that Kantian ethics cannot
justify any substantive norm of action, fails. However, I will argue that, as it stands, Kantian
constructivists fail to give a convincing argument for the wide range of substantive norms
they themselves defend. Here, I partly agree with the critics of Kantian constructivism that,
at least as it stands, Kantian constructivism is susceptible to the objection that it (implicitly)
relies on a potentially escapable conception of agency in the application of a supreme
principle of morality. I disagree with the critics, however, that this would show that Kantian
constructivism fails to justify any interesting substantive norms of action beyond the most
minimal or trivial moral requirements. Rather, it shows that the conception of agency in
Kantian constructivism is underspecified and that specifying and justifying their
conception of agency should be the first priority of Kantian constructivists.
The structure of this chapter is as follows: in section 2, I provide some prima facie
reasons to question the sufficiency of relying on anthropological considerations in
answering the emptiness objection and discuss the need to provide a specification of agency
in the application of a supreme principle of morality. Subsequently, I assess two Kantian
constructivist arguments by focusing on their (explicit or implicit) specification of the
conception of agency in the application of the supreme principle of morality. In section 3, I
analyse O’Neill’s application of the principle of followability. In section 4, I analyse
Gewirth’s application of the principle of generic consistency. Finally, in section 5, I abstract
away from O’Neill’ and Gewirth’s theories to discuss the general implications of my
discussion for the substantive ambitions of Kantian constructivism.
2. Which Conception of Agency?
The picture that has emerged from the previous chapters is this: a transcendental argument
from the first person leads to the conclusion that all agents have a right to whatever are the
necessary conditions or means of agency. In order to specify these necessary means or
conditions of agency, Kantian constructivism needs to rely on both general and contingent
anthropological facts. Taken together, the justification and the application of the supreme
principle of morality thus lead to a set of norms of action which protect the necessary
preconditions of action, allowing for the fact that what exactly these necessary
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
271
preconditions of action are will depend on the specific socio-historical context to which the
supreme principle of morality is applied.
Before discussing the concrete attempts to justify substantive norms by O’Neill and
Gewirth, I aim to provide in this section some prima facie reasons to question the
sufficiency of relying on anthropological considerations in answering the emptiness
objection. The main reason why reliance on anthropology is not sufficient is that, as I have
explained in the previous chapter, anthropological facts, both general and contingent, are
needed in order to identify the necessary means of agency. That is, anthropological facts are
needed in order to determine the ways in which, and the conditions under which, agency
can be either undermined or enhanced. Anthropological facts are not morally relevant in
themselves, but only insofar as they are necessary to identify the necessary means or
preconditions of agency. However, in order to identify the necessary preconditions of
agency through the reliance on anthropological facts, we first need to know what it means
to understand oneself as an agent. After all, if moral substance is determined by reference to
the necessary preconditions of agency, we need to know what exactly is meant by ‘agency’.
In other words, although anthropology is necessary in order to determine the ways in
which our agency might be undermined or promoted (whatever conception of agency one
proposes), for such an appeal to also be sufficient, we need to know the precise content of
this conception of agency in order to determine its necessary preconditions. In the
remainder of this section, I will try to illustrate this point.
Until now I have discussed agency at a rather abstract level, in terms of acting for a
reason or a purpose. This abstract notion of agency was sufficient for the justification of a
supreme principle of morality, insofar as the supreme principle of morality requires an
inescapable, but not necessarily determinate or specific, conception of agency to proceed.2
Now that we are turning to discuss the substantive implications of Kantian constructivism,
however, we have to take a closer look at the conception of agency that it operates with,
because this abstract idea of acting for a reason or a purpose can be specified in different
ways.
In what follows, I briefly discuss different possible specifications of agency by briefly
2
Cf. Beyleveld: “we can run the argument for the PGC [the principle of generic consistency] in terms of relatively
unspecified generic-dispositional freedom and well-being, leaving the application of the PGC as, partly, an
empirical matter ... if we do this, then the argument will be that PPAs [purposive agents] must assent to the
proposition that PPAs have rights to the generic features of agency (without it being specified precisely what this
amounts to)” (Beyleveld 1991, 89).
CHAPTER 9
272
discussing several authors who can be taken to have offered different specifications of the
abstract notion of agency outside the specific context of Kantian constructivism.3 Although
these authors specify agency in different ways, they all draw a distinction between what we
might call more minimal and more maximal conceptions of agency. These more minimal
and more maximal conceptions of agency could be interpreted as different ways in which
an abstract conception of agency could be specified.4
In a recent article, Fabian Schuppert, for instance, distinguishes between “minimal
agency” and “some thicker account of free and autonomous rational agency” (2013, 7).
Schuppert defines minimal agency as follows:
Minimal agency refers to the idea that an agent is able to form intentions, identify
and process reasons, and act on the basis of his/her intentions and reasons.
However, being able to do so, does not mean that one always acts on reasons
which are either one’s own, or most conducive to achieving one’s goals. Minimal
agency, then, is about fulfilling some basic rationality criteria and being able to
execute one’s will in a limited, imperfect sense (Schuppert 2013, 7).
More maximal or, as Schuppert calls it, “free and autonomous agency” on the other hand, is
the ability to form and to act on one’s own plan of life or one’s own conception of the good
(Schuppert 2013, 12). Schuppert goes on to argue that the needs or preconditions of
minimal agency are not the same as the needs for maximal agency:
The requirements for achieving minimal agency, however, are relatively limited.
They include the mental capacity to identify oneself, to form intentions, to
consider reasons, the physical security to act purposefully, and the fulfillment of all
relevant needs for securing one’s (prolonged) survival. The basic needs necessary
for survival, meanwhile, are enough nutritional food and clean water, a minimally
healthy natural environment, basic health care, protective shelter and clothing,
and basic physical security. However, need-claims on the basis of thicker accounts
3
In the previous chapter, I have already briefly discussed Evan Tiffany’s criticism of Korsgaard along these lines
(Tiffany 2012). The following could be read as an elaboration and generalization of Tiffany’s criticism of
Korsgaard.
4
In what follows, I will often talk about minimal and maximal conceptions of agency, but it should be kept in
mind that there are not two different conceptions or specifications of agency, but a whole range of possible
specifications of agency (as will also become clear in the discussion below).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
273
of personal autonomy, or full rational agency, are simply beyond the scope of basic
needs, as the lack of fuller (i.e. more substantial) personal autonomy and free
rational agency – while certainly being harmful and detrimental to the agent’s
ability to live a good life – cannot be considered to be existentially necessary for
the being of the need-claimant. Phrased differently, even if one lacks the
conditions to reason independently and act freely, one is still an agent who enjoys
minimal agency (Schuppert 2013, 7–8).
Without going into the details of Schuppert’s description of minimal agency and maximal
agency and the precise needs of minimal agency versus maximal agency, the crucial point is
that there are different possible specifications of the abstract idea of agency. The central
message is that this, in turn, has implications for the needs or preconditions of agency.
Others have made similar points regarding minimal/maximal agency. Carol Gould, for
instance, distinguishes between ‘basic agency’ and ‘flourishing agency’. She writes:
I distinguish a basic sense of agency characteristic of human beings, and consisting
in intentionality or choice as a feature of human action, from the exercise of this
agency in the development of capacities or the realization of long-term projects or
goals (Gould 2015, 181; see also Gould 2004, 152).
Just like Schuppert, Gould stresses that these two conceptions of agency have different
necessary preconditions:
The first set of human rights – to the conditions of human agency in general –
include the basic rights, especially those of life and liberty, where “life” signifies
not only a right not to be killed but also access to means of subsistence as well as
rights to health care, basic education, and certain others. The second set of what
we might call nonbasic human rights include those to the conditions for the higher
development of human agency in differentiated forms (Gould 2004, 152).
Finally, Claassen (forthcoming) makes a distinction between ‘participational agency’, which
he defines as “a form of participation in a social practice, i.e. a cooperative structure
characterized by a common set of institutions” (Claassen forthcoming, 3) and ‘navigational
agency’, which he defines as “a second, higher-level form of agency: the ability, not to
CHAPTER 9
274
participate in practices, but to navigate between them” (Claassen forthcoming, 3). Claassen
stresses that the question of whether we have to understand ourselves not just as
participational but also as navigational agents has important implications for the necessary
preconditions of agency, because the capabilities for participational agency and
navigational agency are not the same (Claassen forthcoming, 16–18).
Although these authors thus clearly differ in how they define the different, specific
conceptions of agency, they all stress that there are more ‘minimal’ and more ‘maximal’
conceptions of agency (whatever these exactly entail), that the abstract idea of rational
agency is itself indeterminate about these different ways (minimal vs. maximal) in which
the abstract idea of agency could be specified, and that the ‘choice’ of a minimal or maximal
conception has implications for the necessary preconditions of agency.5 The point is thus
that even if the general strategy of determining the content of morality in terms of the
necessary preconditions of agency can be successful, the content of the moral theory, and
the range of moral norms that are justified through this method, depends on how exactly
one specifies agency.
This raises the following question: what exactly do Kantian constructivists mean by
agency, and is this conception closer to a minimal conception of agency or closer to a
maximal conception of agency? Importantly, the choice of a more specific conception of
agency in the application of a supreme principle of morality cannot be made independently
of the justification of the supreme principle of morality. After all, the argument starts from
a notion of agency that is supposed to be phenomenologically inescapable (chapter 4), and
it is not evident that a more specific conception of agency can either be derived from the
abstract conception of agency or that it is itself inescapable.
In the following two sections, I will turn to two recent Kantian constructivist attempts
to justify substantive norms of action on the basis of a supreme principle of morality and I
will evaluate to what extent they succeed in justifying the substantive norms of action which
they propose by focusing on their (explicit or implicit) specification of an abstract notion of
agency and their justification of this more specific conception of agency. I will try to show
that although these theorists put forward substantive norms of action relating to the
necessary preconditions for a rather maximal account of agency, they say very little in
defence of this maximal account of agency. Although this does not show that we have to
5
I use scare quotes when I say ‘choice’ because strictly speaking Kantian constructivism does not choose a
conception of agency but tries to give a description of an inescapable aspect of our phenomenology. I will come
back to this below.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
275
reject their substantive theories, it shows that more work is needed in order to apply a
supreme principle of morality to specific contexts of actions. More specifically, I will argue
that the discussion should focus on the appropriate specification of agency.
3. Onora O’Neill’s Justification of Substantive Morality
3.1 The Principle of Followability
In chapter 5, I have criticized Onora O’Neill’s justification of the principle of followability.
In this section, I critically analyse O’Neill’s application of the principle of followability in
the context of the justification of justifying certain substantive norms of action. In what
follows, I thus assume, for the sake of argument, that the principle of followability can be
justified (for instance by providing a first person justification of this principle), and
subsequently I analyse to what extent the principle of followability can lead to the
justification of substantive norms of action. The reason for this is that although I think that
O’Neill fails to justify the principle of followability, she has written extensively on the
application of the principle of followability; I therefore discuss her attempt to apply the
principle of followability as one possible way in which Kantian constructivists might try to
apply a supreme principle of morality.
Recall that O’Neill puts forward the principle of followability as an interpretation of the
Formula of Universal Law of the Categorical Imperative, which, according to O’Neill,
essentially boils down to the question of whether “what one proposes for oneself could be
done by others” (O’Neill 1989a, 94). More specifically, O’Neill argues that the principle of
followability entails that “a principle [i.e. a maxim or a norm of action] cannot be taken to
be universalizable if it cannot be viewed as a principle for all, because its universal
adaptation (per impossibile) would render some unable to act, a fortiori unable to adopt that
principle” (O’Neill 1996, 163).
Crucially, the principle of followability does not require that all agents should actually
be able to act on a principle or norm of action – because there might always be reasons or
situations that make it imprudent, immoral or impossible to actually (all) act on a certain
principle. In addition, there are many principles which at face value seem morally
permissible but which it is simply not possible for all to act on at the same time. The
question, according to O’Neill, is whether a principle could be coherently adopted by all as a
CHAPTER 9
276
principle for action, i.e. whether it would in principle be possible for all affected to act on
the principle. O’Neill writes:
The requirement that reasoned principles be such that all others (in a restricted or
an inclusive domain) could adopt them demands not that all of them could here or
now, or on some anticipated occasions, or simultaneously, or with some defined
level of success, act on those principles, but that they can be judged principles
which could coherently be adopted by all (O’Neill 1996, 59).
This implies that principles should not just be followable ‘in thought’, i.e. that principles
should not be merely intelligible for all affected, but should also be (coherently) capable of
being adopted as principles of action for which intelligibility is a necessary but not a
sufficient condition (O’Neill 1996, 57; O’Neill 2000b, 25). Obviously, a lot hinges on what
exactly it means to say that a principle could be coherently adopted as a principle of action
by all. I come back to this question below when discussing O’Neill’s justification of
substantive norms of action on the basis of the principle of followability.
If the principle of followability can be used to justify substantive norms of action, it
would, first and foremost, provide constraints on action. O’Neill writes that “what has been
vindicated as the end of the story is not a set of principles whose adoption will determine
thought and action, but only principles that constrain adoptable sets of principles” (O’Neill
1989c, 25). Nevertheless, O’Neill believes that the principle of followability has clear
implications for substantive morality.
In arguing for substantive norms of action on the basis of the principle of followability,
O’Neill follows the general Kantian constructivist strategy of relying on anthropological
facts. O’Neill writes that
once we make reasonably plausible background assumptions, a range of principles
of duty can be identified. In my view these background assumptions include the
thoughts that we act as one of a plurality of agents; that each agent’s action bears
on other agents; and that each agent’s capacities for action are indeterminate,
limited and vulnerable. One can think of these as assumptions about the plurality,
connectedness and finitude of human (or other rational) agents. If any of these
assumptions did not hold, we might find that we could not derive many—or any—
more specific principles of duty. If we were to assume only a single agent, we could
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
277
at most offer an account of duties to self (even this might raise problems). If we
were to assume a plurality of agents who could not affect one another (for
example, because they live on different planets or in different epochs) similar
limitations would follow. If we were to assume a plurality of agents with unlimited
powers to affect one another, other incoherencies arise (O’Neill 2007a, 162).6
The application of the principle of followability proceeds in two steps; first, general
anthropological facts about the plurality, connectedness and finitude or vulnerability of
human beings are used to justify abstract norms of action. Second, specific, contextdependent anthropological facts about the specific ways in which human being are
connected and vulnerable are used to justify more specific norms of action, including
implications for institutional design (I will give examples of this two-stage procedure
below). O’Neill describes this two-stage procedure as follows: “the initial stage of
construction will identify only very abstract principles, but these may be used as the
framework for constructing more determinate principles, so ultimately for guiding action
and policy-making and for building institutions and practices” (O’Neill 1996, 155).
In what follows, I focus on O’Neill’s justification of one specific norm of action, the
rejection of the principle of severe injury, and the more specific norms to which this
abstract norm is supposed to give rise. The reasons for focusing on this particular norm are
that this norm is at the centre of O’Neill substantive moral theory7 and that she has written
extensively on the derivation of both the abstract norm of the rejection of severe injury and
the more specific norms following from this abstract norm (O’Neill 1996, 161–78). She
writes, for instance, that “the principle of rejecting injury is ... fertile enough to ground a
complex web of requirements of justice” (O’Neill 1996, 166f14). I therefore take her
discussion of the rejection of severe injury to be typical of her general strategy of justifying
substantive norms on the basis of the principle of followability in combination with
anthropological facts.8
I will not directly discuss O’Neill’s (1989a) influential article ‘Consistency in Action’
about the application of the formula of humanity of the categorical imperative. In this
6
For a further discussion of these three anthropological assumptions see O’Neill (1996, 100–106).
O’Neill writes: “the core of injustice is action on principles which if universally acted on would foreseeably injure
at least some; the core of justice is rejection of principles whose attempted universal adoption would foreseeably
injure at least some” (O’Neill 1996, 164).
8
In Justice and Virtue, O’Neill uses the principle of followability not only to construct an account of justice but
also to construct an account of virtue. I will limit my discussion to her account of justice.
7
CHAPTER 9
278
article, O’Neill argues that maxims of becoming a slave or slaveholder and maxims of
deception of coercion should be rejected because when universalized these maxims are
conceptually incoherent (contradiction in conception) (O’Neill 1989a, 96).9 In addition,
O’Neill claims that there are certain maxims which, although they are not conceptually
incoherent when universalized, can nevertheless not be willed as a universal law
(contradiction in the will). As an example, she mentions Kant’s defence of a duty of
beneficence. If I universalized a maxim of not helping others when they are in need, I will
not be helped when I am in need. However, given that one wants to be helped when one is
in need (given the hypothetical imperative), one cannot will universalizing a principle of
non-beneficence: “it is only because intending a maxim of nonbeneficence as a universal
law requires commitment to that very absence of help when needed, to which all rational
intending requires assent, that nonbeneficence cannot coherently be universalized” (O’Neill
1989a, 99).
The reason for not further discussing these kinds of arguments is that even if O’Neill’s
arguments in ‘Consistency in Action’ are successful, the resulting duties are highly abstract
and indeterminate (O’Neill 1989a, 100). 10 O’Neill, for instance, acknowledges that the
principle of beneficence “does not specify whom it is morally worthy to help, to what
extent, in what ways or at what costs, but only that it would be morally unworthy to adopt
an underlying intention [i.e. maxim] of nonbeneficence” (O’Neill 1989a, 100).11 Because
O’Neill in her later work, and more specifically in Towards Justice and Virtue, tries to justify
more concrete norms of action on the basis of principle of followability, I will focus on this
work, which is substantively more ambitious than the argument in ‘Consistency of Reason’.
3.2 The Rejection of a Principle of Severe Injury
In the remainder of this section, I will critically evaluate O’Neill’s two-step argument for the
rejection of maxims of severe injury. I will first briefly describe these two stages of O’Neill’s
argument, and subsequently I will evaluate to what extent she succeeds in justifying the
9
About maxims of becoming a slave or slaveholder, she writes, for instance: “action on either of the
nonuniversalizable maxims of becoming a slave or becoming a slaveholder would reveal moral unworthiness: it
could be undertaken only by one who makes of himself or herself a special case.”(O’Neill 1989a, 96).
10
In addition, ‘Consistency of Reason’ is put forward first and foremost as an interpretation of Kant and not as an
application of her own principle of followability
11
This distinction is somewhat artificial because O’Neill does not clearly distinguish between the formula of
humanity and the formula of universal law.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
279
specific, substantive norms she proposes.
According to O’Neill, the rejection of an abstract principle of severe injury follows
directly from the principle of followability because
nobody can coherently think of a principle of injuring severely as adoptable by all
(in any domain): if it were so adopted, some would succeed in acting on it, hence
others would become their victims, so unable to act on it. There is no possible
world in which adoption of a principle of severe injury by all could be achieved;
there is nothing agents could do that would count as or contribute to bringing
about such a world (O’Neill 1996, 59).
The reason for rejecting a principle of severe injury is thus, according to O’Neill, that if one
adopted this principle as a principle of action, others could in principle not act on it
(because they would be severely injured). For instance, if I killed you, as a consequence of
adopting the maxim of severe injury, you could not in principle adopt the same principle as
a principle of action because one obviously has to be alive in order to adopt a principle in
the first place. O’Neill describes the reason for rejecting the principle of severe injury as
follows:
The ordinary and predictable results even of widespread, let alone of ostensibly
universal, commitment to an inclusive principle of injuring within some domain
would injure some, so disable them from acting, so show the principle itself nonuniversalizable. Those who adopt principles which they cannot view as principles
for all must see what they do as non-universalizable, that is as available to some
only on condition that at least some others not do likewise (O’Neill 1996, 163).
This argument clearly relies on general anthropological facts. First, it presupposes that
there is a plurality of human beings. Second, it presupposes that this plurality of human
beings is connected, so human beings have to interact with each other. Finally, it
presupposes that human beings are vulnerable, so they can be injured by others in their
interactions. O’Neill summarizes this stage of the argument as follows, saying,
putting matters simply, principles of action will not be universalizable if attempted
universal adoption would forseeably injure capacities and capabilities for action of
CHAPTER 9
280
some of those within the relevant domain of ethical concern, thereby ensuring that
they cannot adopt those principles (O’Neill 1996, 164).
O’Neill suggests that similar arguments can be made to rule out principles of deception and
coercion (which she considers to be a specific instantiation of the general principle of the
rejection of injury). Deception and coercion make it in principle impossible for some
persons in a certain domain to adopt these principles as principles of action, because these
types of actions undermine the very possibility of accepting these principles as principles of
action for all involved (O’Neill 1996, 174; see also O’Neill 1989a, 96).
The rejection of the principle of severe injury shows, according to O’Neill, that both
our lives and our institutions should be based on a rejection of the principle of severe injury
(O’Neill 1996, 166). O’Neill contrasts the rejection of a principle of injury with adopting a
principle of non-injury. Whereas a principle of non-injury would require, for instance,
becoming a pacifist, rejecting injury “is a matter of not making a principle of injury
fundamental to lives, institutions or practices” (O’Neill 1996, 166). O’Neill’s point is thus
that we should not so much positively adopt a principle of non-injury but that we should,
negatively, reject a principle of injury. Although in practice the positive principle of noninjury and the rejection of a principle of injury overlap, this need not be the case, as the
example of pacifism shows: it might be needed to sometimes injure people as a result of
rejecting the principle of injury (for instance by preventing people from injuring others).
Although O’Neill claims that more specific norms of action cannot be simply
“deduced” (1996, 168) from the principle of rejecting injury, she nevertheless believes that it
is possible to construct more specific norms of action by “tak[ing] account of actual
capacities and capabilities for action of those involved and of their counterpart
vulnerabilities to injury” (O’Neill 1996, 167–68). This brings us to the second step of her
argument, in which O’Neill invokes more contingent anthropological facts to justify more
specific norms of action.
O’Neill distinguishes two types of injury (O’Neill 1996, 168). Direct injury directly
affects agents and their capacity for action. O’Neill puts forward the following examples of
direct injury: “direct injury ... may be inflicted by killing or destroying, by wounding or
maiming, by threatening or coercing, by forced labour and starvation, by intimidation and
terror, by detention and deportation. Evidently no principle of acting in these ways can be
universalized” (O’Neill 1996, 168).
Indirect injury, on the other hand, damages “the conventions trust, traditions and
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
281
relationships by which pluralities of agents maintain a social fabric and complex
capabilities” (O’Neill 1996, 168) or damages “the natural and man-made environments,
which provide the material basis both for life and lives and for the social fabric” (O’Neill
1996, 168). Indirect injury thus entails damaging the social and natural preconditions of
action.
The rejection of these two types of injury requires different responses. Limiting direct
injury requires overcoming avoidable conditions under which injury typically happens, e.g.
by limiting both the power of the stronger and the vulnerabilities of the weaker. According
to O’Neill, limiting direct injury has important institutional implications:
The rejection even of direct injury will require institutions that limit and control
the power that some have over others, and hence also the vulnerability of those
others. Limiting power and vulnerability may require the construction of legal
orders which guarantee certain universal liberty rights, of political orders which
limit powers of government, of economic orders which secure tolerable levels of
subsistence, and of social orders which limit subordination and dependence, and
of
information
orders
which
regulate
communication
and
above
all
telecommunication (O’Neill 1996, 169–70).
The criterion of the adequacy of these rules of conduct, institutions and practices is the
question of which system is most effective in limiting injury (O’Neill 1996, 171). Writing on
the question of what sort of economic order should be preferred, O’Neill, for instance,
claims that “the most just system for given conditions will have to do better than others in
reliably sustaining capacities and capabilities for action, so limiting vulnerabilities, in those
conditions” (O’Neill 1996, 170f17). The same principle applies to the prevention and
remediation of indirect injury. Limiting indirect injury mainly concerns the limitations of
various forms of deceit, which undermines the free availability of information and relations
of trust, which are both necessary for a functioning society: “deception is unjust because it
injures indirectly: systematic or gratuitous deception destroys trust and creates
vulnerabilities, which facilitate direct injuries. By destroying the social fabric deception
injures those whose lives are linked by that fabric” (O’Neill 1996, 174). Just like the
limitation of direct injury, the limitation of indirect injury also has institutional
implications, although, according to O’Neill, institutions can never completely rule out the
possibility of indirect injury. She explains that:
CHAPTER 9
282
a just political and legal order can at best go some way – perhaps a long way –
towards avoiding and preventing systematic and gratuitous deception, for example
by establishing criminal and civil law, by securing forms of free speech and
freedom of information, and by demanding and enforcing some degree of probity
in public, commercial and professional life (O’Neill 1996, 175–76).
In addition to the creation of institutions which aim to limit various forms of deceit,
limiting indirect injury also requires the limitation of damage done to the natural world, or
at least insofar as the natural world is a necessary precondition of our ability to act and to
adopt rules. The limitation of this second form of indirect injury also has important
institutional implications. According to O’Neill,
environmental justice is ... a matter of transforming natural and man-made
systems only in ways that do not systematically or gratuitously destroy the
reproductive and regenerative powers of the natural world, so do not inflict
indirect injury (O’Neill 1996, 177).
The reason I have quoted O’Neill at length is to illustrate the way in which she tries to
derive specific, substantive norms of action from the principle of followability and to grasp
the scope of the norms which she proposes. O’Neill first derives the rejection of an abstract
norm of severe injury from the principle of followability by relying on certain general
anthropological facts. And subsequently she formulates more specific norms of action,
including institutional implications, by applying the rejection of the abstract norm of severe
injury to specific contexts of ethics through a reliance on contingent anthropological facts.
This leads to a rather comprehensive, although still partly indeterminate, moral theory
which includes a range of environmental, political, economic and social demands (O’Neill
1996, 177).
3.3 Injury, Agency and the Emptiness Objection
How did we get from the principle of followability to this comprehensive moral theory?
Recall that, according to the principle of followability, “reasons for action must be held
capable of being followed or adopted by others” (O’Neill 1996, 57), meaning that a reason
or a maxim of action is only followable if it could coherently be adopted by all others. It
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
283
remains ambiguous, however, on which grounds exactly (specific) norms of action should
be rejected or accepted on the basis of the principle of followability. On the one hand,
O’Neill claims that a principle should be rejected if (some) others cannot coherently adopt
this as their principle of action. On the other hand, however, O’Neill suggests that injury is
prohibited if it undermines a person’s agency directly. Consider, for instance, O’Neill’s
remark concerning the rejection of direct injury on the basis that it affects “agents and their
capacity for action” (O’Neill 1996, 168). On this interpretation, the focus is thus not on the
agent’s ability to adopt the specific principle under consideration, but directly on his or her
capacities for agency. Of course, these two possible interpretations of O’Neill’s derivation of
substantive norms from the principle of followability do not necessarily have to come apart
insofar as one needs to be an agent in order to able to act on any principle whatsoever, but,
as I argue below, they can come apart (because agency might entail more than just acting
on a principle). Let me therefore discuss both interpretations in turn.
First, regarding the criterion that a norm should be rejected if it cannot coherently be
adopted by all, I think it is clear that certain norms of action can indeed not be coherently
adopted by others, such as norms concerning killing others. If one killed another person,
they could not in principle adopt this norm as a norm of action. If there is such a thing as a
necessary condition of being able to follow a principle, it is being alive. I therefore think
that O’Neill succeeds in answering the strong emptiness objection, which claims that
Kantian constructivism cannot justify any substantive norm of action. But as we have seen,
O’Neill’s substantive moral theory is much more ambitious than trying to justify a
prohibition on killing.
Whether or not the principle of followability can lead to the justification of the
substantive norms of action that O’Neill proposes depends on what this standard of
coherency in the principle of followability amounts to. What exactly does it mean to say
that a principle can(not) be coherently adopted as a principle of action by all? At least at
face value, this standard does not seem to rule out most actions which we would normally
conceive as injurious and, more importantly, it does not justify the more specific norms of
action proposed by O’Neill. True, if I were to adopt the principle of killing another person,
then obviously some others could not adopt this principle as a principle for action. But say
that I adopt the norm of cutting off the finger of a colleague if he or she disagrees with me.
Would that make it impossible for my colleague to adopt this principle of action? It seems
not. Or say that I act in such a way that other people would as a result be undernourished,
their natural environment would be damaged or they would lack adequate healthcare.
CHAPTER 9
284
Although we would normally consider these kinds of actions to constitute extreme forms of
injury, it is not clear how it would be incoherent for me to hold a principle that would lead
to this result. It seems that this would only be the case if, say, damaging someone’s natural
environment would make it in principle impossible for that person to also damage
someone else’s natural environment. Incoherency seems to be too minimal a standard to
rule out most, if not all, of the specific norms of action that O’Neill proposes. Or at least, it
is unclear how incoherency is to lead to these kinds of specific norms of action – especially
when taking into consideration that, according to O’Neill, the principle of injury does not
require that all affected can actually act on the principle, which would clearly be impossible
in the case of damaging the environment.
O’Neill anticipates this kind of objection. She calls this the objection of ‘trivial injury’,12
which holds that “the world in which everybody inflicts trivial injury may be beastly, but it
is quite coherent to suppose that everybody gets in on the act” (O’Neill 1996, 165). O’Neill
responds to this objection by saying that
if certain inclusive principles can be shown non-universalizable (in this case an
inclusive principle of injuring), laws, institutions, practices and policies must be
designed to meet the requirements of rejecting that principle. It may well be that
an overall commitment to rejecting injury will turn out to allow (or even to
require) certain sorts of trivially injurious action; but what it allows or requires will
not be generalized trivial injury. A serious commitment to rejecting injury will
only be properly expressed in social, political and other structures that put rather
strong prohibitions on injury and make exceptions only for equally serious
reasons. Whether trivial injury is permissible in some contexts will emerge in the
course of considering just what it takes to make the rejection of injury sans phrase
a basic principle of lives and institutions ... the central point is to see what must be
done if those inclusive principles which can’t be principles for all are rejected,
rather than to discover whether related principles could be principles for all
(O’Neill 1996, 165).
12
The term ‘trivial’ is a bit misleading in this context because my examples (cutting off fingers or damaging the
natural environment in which people live) do not seem to be trivial at all. I think O’Neill’s ‘trivial injury’ should be
read as a technical term which refers to any kind of injury which does qualify as severe injury on O’Neill’s
definition of severe injury, i.e. any injury which can be universalized (versus severe injury, which, according to
O’Neill, cannot in principle be universalized).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
285
Insofar as I understand O’Neill’s response to this objection, it seems to hold the middle
ground between biting the bullet and simply dismissing the idea that trivial injury can be
universalized. On the one hand, she admits that certain ‘trivially injurious action’ might still
be allowed under the general rejection of an abstract norm of severe injury, thereby
potentially allowing that the specific norms of the rejection of injury which follow from the
abstract norm have a narrower scope than she herself suggests. On the other hand, she
denies that ‘generalized trivial injury’ can be universalized. She seems to give an
instrumental justification for this conclusion: rejecting (severe) injury can only be realized
if institutions and practices prohibit a wide range of injury (including trivial injury).
I am not sure how convincing I find this response. The problem is that what is at stake
is what counts as injury on O’Neill’s moral theory in the first place. Her point that one
should not try to discover whether a principle of trivial injury can be a principle for all
therefore misses the mark, because the point of the objection is to press O’Neill to make
clear whether or not instances of what we would normally consider to be injurious count as
injurious on her definition of injury. Given that she defines injury in terms of the
impossibility of acting on a certain norm of action, it is not evident that what we would
normally consider as injurious counts as injurious on O’Neill’s account, because not all
things that we would normally consider to be injurious make it impossible for others to act
on these principles of trivial injury. I do not see how O’Neill’s response answers this worry.
The problem with the first interpretation of O’Neill’s argument, the interpretation
which claims that a norm should be rejected if it cannot coherently be adopted by all, is
thus that it fails to rule out most forms of injury, except for the most extreme forms of
injury. Or at least, there is a gap between the rejection of principles of extreme injury that
are obviously incoherent to adopt as universal principles and the rather comprehensive
substantive moral theory that O’Neill proposes.
Let me now turn to the second interpretation: the interpretation which focuses directly
on the implications of adopting a principle for the capacities for agency as such (and not
indirectly via the coherency of universalizing a specific principle). I aim to show that
whether or not this interpretation succeeds in justifying the specific norms of action that
O’Neill proposes depends on how exactly one understands agency, and that O’Neill fails to
both specify and justify the conception of agency she implicitly relies on to justify the
norms of action she proposes.
In consideration of what is discussed in section 2, we can make a distinction between
more minimal and more maximal conceptions of agency. On a minimal conception of
CHAPTER 9
286
agency, a principle should be rejected if it makes it impossible for an agent to act at all, i.e. if
it “destroy[s] ... capacities and capabilities to (inter)act” (O’Neill 1996, 168). Although this
conception of agency is clearly presupposed by the principle of followability (because in
that case it is in principle impossible to act on that, or any other, principle), it seems to rule
out only the most extreme forms of injury, because even in the most deprived situations
persons might still be able to act in some sense.13 Relying on a minimal conception of
agency thus does not go beyond the interpretation of the principle of followability in terms
of the incoherency of adopting a principle. Or at least, it is unclear how this minimal
conception of agency would be undermined except for the most extreme forms of injury.
In order to go beyond these minimal norms of action, O’Neill would thus have to rely
on a more maximal interpretation of agency. According to a more maximal conception of
agency (and the one that seems to be used by O’Neill throughout her discussion of concrete
implications of the rejection of injury), a principle would be ruled out if it diminishes or
“damages” (O’Neill 1996, 168), rather than completely undermines or destroys, other
agents’ agential capabilities.14 The point is thus that it seems that one could still be a
‘minimal’ agent, i.e. being able to act at all, while failing to have (all) the capabilities of a
more maximal agent.
Focusing on agential capabilities in combination with the introduction of the criterion
of the prohibition on diminishing agential capabilities might indeed lead to the more
specific norms of action that O’Neill proposes. It is questionable, however, whether this
more maximal conception of agency follows from the principle of followability. There is a
crucial difference between claiming that people’s agency as such should not be undermined
and the claim that the adoption of a principle should not diminish, injure or affect an
agent’s capabilities for action (whatever they might be).
For instance, in some sense my capabilities for action might be diminished if my
income were to be lowered (I might no longer be able to buy myself coffee, or go to Bob
Dylan concerts), but this does not mean that my agency as such is undermined: I still have
13
I elaborate on this point in section 4, which discusses Gewirth’s theory.
O’Neill does not distinguish between these two interpretations of agency. She writes, for instance, that “injury
can destroy or damage bodies (including minds), bodily (including mental) functioning and so capacities and
capabilities to (inter)act and to respond” (O’Neill 1996, 168 my emphasis). Instead of (just) defining injury in
terms of destroying agency altogether, she also presupposes a certain standard of agency which should be
guaranteed, although she does not specify what this standard is supposed to look like. O’Neill claims that the
criterion to judge institutions is that they should “limit both systematic and gratuitous injury and secure some
measure of the most important capabilities for action for all” (O’Neill 1996, 171 my emphasis).
14
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
287
the capacities to act for reasons, i.e. the capacities to adopt a principle as a principle of
action. My not being able to buy tickets to see Bob Dylan, it seems, means my agency is
diminished in some sense, but it is not obvious that it is diminished in any morally relevant
sense. Note that I do not want to deny that it might be diminished in a morally relevant
sense; my point is only that this needs to be argued for and that it is not obvious that
O’Neill has the resources to show that this more maximal conception of agency is justified
by her argument. In other words, the focus on one’s agency being diminished rather than
altogether removed is helpful insofar as it opens up a way of accounting for substantive
morality, but it is also problematic because it raises the question of what counts as the
relevant agency in the first place.
My point here is easily misunderstood. Let me therefore stress that I do not want to
claim that a more minimal specification of agency should be the default specification of
agency and that the burden of proof is on O’Neill to justify a more maximal conception of
agency. The point rather is that O’Neill (implicitly) seems to rely on a (maximal)
specification of agency which she fails to justify and that any specification of agency,
whether minimal or maximal, is in need of justification (I will come back to this in section
5).
In addition, let me stress that these minimal and maximal accounts of agency are not
the only possible conceptions of agency. There might be many conceptions of agency in
between these two extremes. The point of distinguishing these two conceptions of agency is
to illustrate that O’Neill’s conception of agency is ambiguous: she says very little about how
she understands agency and why a certain specific conception of agency would be justified.
What she does say about conceptions of agency is mainly negative. For instance, she
criticizes John Rawls for relying on an ‘ideal’ account of agency, as opposed to relying on
just an abstract account of agency. Abstraction, according to O’Neill, involves “pruning
away assumptions whose truth cannot be ascertained, and relying on a meager and
parsimonious set of plausible assumptions” (O’Neill 1996, 39–40) Abstraction “is a matter
of bracketing, but not of denying, predicates that are true of the matter under discussion”
(O’Neill 1996, 40). An idealization, on the other hand, involves the ascription of “predicates
– often seen as enhanced, ‘ideal’ predicates – that are false of the case in hand, and so denies
predicates that are true of that case” (O’Neill 1996, 41).
O’Neill claims that she presupposes “only an abstract, hence non-idealizing and banal,
account of agents and of conditions of action” (O’Neill 1996, 48). But she does not specify
what this ‘banal account of agency’ amounts to, besides stressing that it relies on an
CHAPTER 9
288
“empirically realistic view of the capacities and capabilities agents have” (O’Neill 2000d, 7).
This emphasis on the actual capacities and capabilities of agents is necessary but not
sufficient to justify the substantive norms of action that O’Neill proposes, because we do
not just need an account of the conditions under which agency should be undermined; we
also need to know which sort of agency deserves moral protection. In other words, saying
that agency deserves moral protection is in itself insufficient, because this claim is still
indeterminate about whether one needs to ensure that agents need to be able to engage in
any action whatsoever (minimal agency), or that any limitation of one’s agential capabilities
would be morally problematic (maximal account of agency), or anything in between these
two options.
O’Neill would thus have to say more about how, exactly, she understands agency and
how this account of agency could lead to the specific norms of action that she defends
without relying on problematic idealizations or unvindicated, substantive ideals of agency.
In the absence of such an argument, she is susceptible to the objection that she relies on an
optional and unvindicated conception of agency in the application of the principle of
followability.
4. Alan Gewirth’s Generic Features of Agency
In this section, I discuss Alan Gewirth’s approach to the justification of substantive norms
of action in order to analyse whether Gewirth succeeds where O’Neill fails. I try to show
that despite the differences between these two authors, for instance in terms of the principle
of morality that they defend (the principle of followability versus the principle of generic
consistency), and despite the fact that Gewirth has written extensively on the identification
of the necessary means of agency, they both fail to convincingly specify and justify the
conception of agency that is presupposed in their defence of substantive norms of action.
In chapter 7, I have discussed Gewirth’s argument for the principle of generic
consistency. In the discussion of the argument for the principle of generic consistency, I
have assumed that there are at least some means that are necessary to engage in any action
whatsoever (such as being alive), but I have not specified the content of the generic features
of agency. In this section, I discuss the substantive implications of the principle of generic
consistency by focusing on the way in which Gewirth specifies the generic features of
agency and Gewirth’s specification of purposive agency. I try to show that the specification
of the generic features of agency and Gewirth’s specification of agency are closely related
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
289
because the specification of the generic features of agency ultimately hinges on how exactly
Gewirth’s conception of agency should be understood.
Let me stress that, just like in chapter 7, I draw on the recent elaboration and
clarification of Gewirth’s moral theory by Deryck Beyleveld. Again, I do not pursue the
question of the extent to which Beyleveld’s Gewirth and the Gewirth of Reason and
Morality are necessarily the same, except when there are clear differences between their
positions about the details of Gewirth’s theory.
4.1 The Application of the Principle of Generic Consistency
Recall that, according to Gewirth, any agent, simply by virtue of understanding him- or
herself as an agent, must admit, on pain of self-contradiction, that s/he ought to act so as to
respect the right of all agents. More specifically, Gewirth argues that any agent, simply by
virtue of understanding him- herself as an agent, has to accept the principle of generic
consistency, which claims that all agents have an equal claim right to the generic features of
agency, i.e. a right to those goods that are necessary for action in general.
Gewirth claims that the principle of generic consistency is an improvement on the
categorical imperative when it comes to the justification of substantive morality. The reason
for this is that, according to Gewirth, the principle of generic consistency combines the
formal aspects of the categorical imperative (consistency) with an emphasis on the
necessary contents of morality (the generic features of agency). Gewirth writes that “the
basic problem [with the categorical imperative] lies in the fact that Kant’s appeal to
consistency and universalizability provides only a formal and quantitative specification,
which is not sufficient as a criterion of morally justified contents. To attain such a criterion,
the formal criterion of consistency must be supplemented by a substantive or material
necessity” (Gewirth 1993, 500).15
The extent to which Gewirth succeeds in justifying substantive morality thus crucially
depends on his specification of the generic features of agency, since the generic features of
agency are introduced as the required substantive or material necessity of a supreme
principle of morality. In Gewirth’s theory, moral content is to be determined by providing
an account of the features/goods that are necessary for action in general, independent of the
15
The point here is not to evaluate to what extent Gewirth’s criticism of Kant is valid, but to illustrate how exactly
Gewirth aims to justify substantive morality, i.e. through an analysis of the generic features of agency.
CHAPTER 9
290
content of one’s contingent ends, purposes or desires. The content of these generic features
of agency, in turn, depends on the conception of agency which forms the starting point of
Gewirth’s argument.
Recall that Gewirth defines an agent as someone who understands him- or herself as a)
acting for a purpose, i.e. acting for some end or purpose that constitutes his or her reason
for acting, and b) acting voluntary, i.e. the performance of an action is under the agent’s
control in that s/he unforcedly chooses to act as s/he does. Gewirth’s idea is that this
conception of ‘purposive agency’ does not express a contingent ideal of agency, but
constitutes an inescapable aspect of our self-understanding (see also chapter 7).
According to Gewirth, purposive agency has two generic conditions, i.e. features that
are necessary in order to pursue any purpose whatsoever, which parallel the two features of
action (I only mention these categories here and I come back to their justification below).
The first generic feature of agency is freedom, which concerns the ability to control one’s
behaviour by one’s voluntary choice (what Gewirth calls occurrent freedom) and the
longer-range ability to exercise such control (dispositional freedom) (Gewirth 1978, 52).16
Beyleveld further specifies this abstract notion of freedom by claiming that “basic freedom
includes the freedom to act according to one’s chosen purposes and freedom of thought”
(Beyleveld 2011, 2f1).
The second generic feature is well-being, which concerns the abilities and (material and
social) conditions of achieving one’s purposes. Well-being is further subdivided into three
levels. Basic goods concern the necessary preconditions of action. Gewirth defines basic
goods as follows:
[B]asic goods, which are the general necessary preconditions of action, comprise
certain physical and psychological dispositions ranging from life and physical integrity
(including such of their means as food, clothing, and shelter) to mental equilibrium
and a feeling of confidence as to the general possibility of attaining one’s goals
(Gewirth 1978, 54).
Non-substractive goods concern the goods necessary for the maintenance of agential
16
Gewirth seems to use the term ‘dispositional’ to refer to a diachronic element of freedom, i.e. a longer-range
ability. I will follow Gewirth in this usage of the term dispositional.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
291
capabilities.17 To lose a non-substractive good “is to have one’s level of purpose-fullfillment
lowered” (Gewirth 1978, 55). According to Gewirth, non-substractive goods include both
basic goods and “whatever else, before his acting, the agent has and regards as good”
(Gewirth 1978, 55). Beyleveld specifies these non-subtractive goods by giving the following
examples: “non-subtractive needs include provision of accurate information and being able
to rely on promise” (Beyleveld 2011, 2f1).
Additive goods concern the goods necessary for the improvement of agential
capabilities. Beyleveld, for instance, writes that “additive needs include access to new
information and opportunities to learn new skills” (Beyleveld 2011, 2f1). Gewirth gives the
following examples of additive goods: “education, self-esteem, and opportunities for
acquiring wealth and income” (Gewirth 1985, 237).
It is important to distinguish between freedom and well-being as abstract conditions of
agency and the specification of freedom and well-being in terms of food, clothing,
education and mental equilibrium, for instance. Beyleveld, for example, stresses that “the
identification of generic conditions requires careful analysis, and these examples are
illustrative only” (Beyleveld 2011, 2f1). Whereas Gewirth thus claims, on the most abstract
level, that freedom and well-being are necessary preconditions of agency, the specification
of freedom and well-being in terms of specific goods and institutional requirements do not
(and do not have to) carry the same necessity. The reason for this is that freedom and wellbeing can be a) realized in a multitude of ways (e.g. in order to be hydrated one could drink
water or juice) and b) require different kinds of concrete actions depending on the sociohistorical context. Beyleveld writes that
what will, in practice, constitute the necessary means to the GCAs [i.e. the generic
features of agency] abstractly specified can and will vary according to contingent
factors, such as the biological constitution of an agent, and resources available to
an agent. For example, air of a certain composition is a necessary means to the life
of human agents, but it need not be for agents with a different biological makeup
(if such exist). Also, some things, like peanuts, which can threaten the life of an
17
Gewirth uses capabilities to refer to ‘real opportunities’ (which corresponds to the 'minimal' definition of
capabilities put forward by Robeyns 2011). Gewirth (1998, 63f5) briefly discusses the relation between his usage of
capabilities and Amartya Sen’s more comprehensive capability theory. For discussion of the differences between
Gewirth and Nussbaum see Claasen and Düwell (2013).
CHAPTER 9
292
agent allergic to them can be, and are, healthy foods for other agents (Beyleveld
forthcoming, 9–10f47).
Gewirth and Beyleveld thus use the by now familiar two-step procedure in which Kantian
constructivism, first, relies on general anthropological facts for the identification of some
abstract generic means and, secondly, relies on more contingent anthropological facts for
the specification of these generic means.
Although contingency thus increases with the further specification of the generic
features of agency, this does not undermine the necessity of the generic features as such.
Freedom and well-being are necessary preconditions of agency, but what freedom and wellbeing require in a specific context might change depending on contingent anthropological
facts. Gewirth’s argument thus combines the claim to the necessity of the generic features of
agency with flexibility when it comes to their realization in concrete contexts and for
particular persons.
In addition, Gewirth proposes specific guidelines for this further specification, for
instance by introducing the idea of ‘degrees of necessity for action’ in order to settle
potential conflicts between generic features of agency (Gewirth 1978, 62–63). The idea of
the degree of necessity entails that some goods are more necessary for action than others,
e.g. basic goods are more necessary than additive goods, and that, in case of conflict, goods
that are more necessary for action have priority (I come back to this below).
Besides specifying the generic features of agency along the lines indicated above,
Gewirth applies the principle of generic consistency to different contexts of action. In
Reason and Morality, Gewirth, for instance, discusses both direct and indirect applications
of the principle of generic consistency by discussing the requirements imposed by the
principle of generic consistency on the actions of individual agents (direct application) and
the implications for social rules and institutions (indirect) (Gewirth 1978, 200). In The
Community of Rights (1996), he further explores the implications of the principle of generic
consistency for social justice, arguing for the importance of socio-economic rights and
defending a specific idea of the welfare state. Moreover, Gewirth argues that the principle of
generic consistency can justify human rights (Gewirth 1992; See also Beyleveld 2011).
Gewirth’s application of the principle of generic consistency, through a specification of
the generic conditions of agency, thus has the following structure:
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
293
1) The definition of a conception of agency in terms of the constitutive features of
agential self-understanding: purpositivity and voluntariness.
2) A definition of a set of generic features of agency in terms of those abstract means
and conditions which are necessary to engage in agency as defined under (1):
freedom (occurrent and dispositional freedom) and well-being (basic goods, nonsubtractive goods and additive goods).
3) A further specification of the generic features of agency that allows for variability
in relation to concrete contexts of application (the direct and indirect applications
of the principle of generic consistency) and is guided by the principle of degrees of
necessity for action.
Just like O’Neill, Gewirth believes that the application of a supreme principle of morality
can be used to justify a rather comprehensive moral theory, which includes far-reaching
political, social and economic requirements. According to Gewirth, the generic features of
agency not only include rights to the most basic preconditions of agency (basic goods) but
also rights to the maintenance (non-subtractive goods) and improvement of one’s
purposive fulfilling capabilities (additive goods), for instance the right to life (basic), the
right to the provision of accurate information (non-subtractive) and the right to education
(additive). In addition, generic freedom not only includes occurrent freedom, the freedom
to control one’s behaviour at a particular moment, but also dispositional freedom, the
longer-range ability to exercise such control. If the Gewirthian argument succeeds in this, it
would thus succeed in justifying a wide range of norms of action on the basis of a
transcendental argument from agency.
In what follows, I will analyse to what extent Gewirth succeeds in justifying substantive
norms of action, and more particularly the wide range of norms of actions which Gewirth
believes can be justified on the basis of the principle of generic consistency. I will not go
into the details of the specific applications of the principle of generic consistency that
Gewirth and Gewirthians such as Beyleveld have proposed. Instead, I want to focus on the
specification of the generic features of agency and the conception of agency that is
presupposed in this specification. I will focus specifically on the role of additive goods and
dispositional freedom in the application of the principle of generic consistency, because
they play a crucial role in going beyond the most minimal moral requirements (as I will
argue below).
CHAPTER 9
294
4.2 Identifying the Generic Features of Agency
Recall that the generic features of agency do not express any contingent features of action
but should be understood as the “invariant features that pertain generically to all actions”
(Gewirth 1978, 25). So, before proceeding, it will be helpful to briefly consider how the
generic features of agency should be identified in the first place. Only then is it possible to
evaluate the success of Gewirth’s argument.
Gewirth distinguishes between a ‘particular occurrent’ and a ‘generic-dispositional’
interpretation of the generic features of agency, and he claims that the generic features
should be understood in generic-dispositional terms (Gewirth 1978, 58). On a particular
occurrent interpretation of the generic features, they “consist in the particular purposes any
person may actually try to fulfil by his actions, including maintaining particular basic
goods, retaining the particular goods he already has, and obtaining further particular
goods” (Gewirth 1978, 58). On a generic-dispositional interpretation, the generic features
“consist in the general conditions and abilities required for fulfilling any such particular
purposes. In the generic-dispositional view, the emphasis falls not on particular goods for
which the agent occurrently acts, but rather on the general necessary conditions that enable
him to act for any purpose he might have and that he would hence regard as generically
good” (Gewirth 1978, 58).
The generic features of agency, understood in generic-dispositional terms, thus single
out the necessary means for acting for a purpose in general and not the (contingent) means
for acting for a specific, contingent purpose. For example, it might be a means for my
contingent purpose of getting a coffee that there actually is a place to buy coffee where I can
order a coffee, but this is not a necessary means of action as such. There being a place where
I can buy coffee is certainly not a necessary condition of your making curry, say. What is
necessary to engage in this kind of coffee-pursuing action (or any other purposive action
whatsoever) is, for instance, to be alive and to have a certain freedom of movement, etc.
These are the conditions that my getting a coffee and your making curry share.
Beyleveld therefore stresses that the generic goods should not be identified by
pinpointing “some particular occurrent freedom that every [agent] must want” (Beyleveld
1991, 86). Instead, Beyleveld stresses that generic goods are “‘abstract’ conditions in that
their content does not depend on the content of my particular occurrent purposes, only my
having such purposes” (Beyleveld 1991, 86). Gewirth and Beyleveld also refer to these
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
295
conditions as “second-order goods” (Gewirth 1978, 59) or “second-order capacities”
(Beyleveld 1991, 363, 394).
How does one have to go about trying to identify generic features of agency? Gewirth
gives the following example:
Consider, for example, some of the standard extreme cases of divergent purposes,
such as those of the religionist who wants to dedicate himself to God even if this
means losing his earthly life, or of the intellectual who welcomes blindness because
it enables him to concentrate on his philosophical inquiry. It may be thought that
the capabilities of action required for fulfilling such purposes have little or nothing
in common with those required for fulfilling the purposes of the person dedicated
to this-worldly success or to militaristic prowess. Nevertheless, as long as the
religionist and the intellectual are agents, acting for purposes they want to achieve,
they must satisfy the necessary general conditions of agency. They must control
their behaviors by their unforced choices, they must be able to make certain
minimal plans for what they will do in order to achieve what they want, and they
must have or maintain the general abilities and conditions required for exerting
such control and making such plans (Gewirth 1978, 59).
Thus, in order to establish what counts as a generic feature of agency, we should analyse
what is necessary not just for a specific action but for engaging in certain types of actions,
e.g. one should ask the question ‘Which necessary means of getting a coffee are not just
necessary for this specific activity but for action in general?’ These generic means are the
necessary means of agency as such, as opposed to the necessary means of a particular
action.
4.3 Additive Goods and Dispositional Freedom
Given that the generic features of agency are identified in terms of the necessary means of
purposive action, we need to know how exactly to understand purposive agency: what does
it mean to be a purposive agent, and when does one no longer qualify as a purposive agent?
If the generic features are the “invariant features that pertain generically to all actions”
(Gewirth 1978, 25), we need to know when one ceases to be a purposive agent in order to
determine which features are truly needed for all actions and which conditions are only
CHAPTER 9
296
needed for particular actions. What are the abstract conditions of agency, or the capacities
of agency?
Gewirth defines an agent as someone who understands him- or herself as a) acting for a
purpose, i.e. acting for some end or purpose that constitutes his or her reason for acting,
and b) acting voluntarily, i.e. the performance of an action is under the agent’s control in
that s/he unforcedly chooses to act as s/he does. Understood at this level of abstraction, it
might seem the generic features are very limited in range. The reason for this is that it
might seem that a person can still be a purposive agent while not living in a liberal
democracy or while being subject to the kinds of treatments which we normally deem to be
immoral, for instance while being injured, coerced or threatened, or even when being
imprisoned or enslaved (I made a similar point in the context of O’Neill’s justification of
substantive norms of action). These kinds of interference might undermine some of his or
her purposes but not his or her voluntariness, or his or her purposive agency in general.
There are many actions the prisoner or the slave cannot pursue because of his
imprisonment or enslavement, but there are also many actions that are still open to the
prisoner or the slave – this might include the choice of when and how to hum a song or
when and how to stroll across his cell or hut, etc. This seems to suggest that the range of
generic features is much more restricted than Gewirth suggests – hence I shall refer to this
as the ‘range’ objection.18 I will refer to these cases as potential limiting cases of purposive
agency.19
As I show below, Gewirth’s and Beyleveld’s response to the range objection is to claim
that it overlooks the fact that the generic features of agency not only include occurrent
freedom and basic goods but also dispositional freedom and additive goods. This allows
Gewirth and Beyleveld to say that although some generic features might still be in place in
some cases, such as imprisonment or slavery (basic goods and occurrent freedom),
mentioned above, not all generic features are (additive goods and dispositional freedom).20
This response, however, highlights that it is crucial that dispositional freedom and additive
goods should be considered to be generic features of agency. The question therefore is
18
Variants of this objection are put forward by Narveson (1980, 658), Danto (1984), Regan (1999), Richardson
(1999), Illies (2003, 118–21). I will not go into the details of these specific objections. Instead, I will focus on the
general line of response taken by Gewirth and Beyleveld in light of the range objection and the potential limiting
cases of purposive agency.
19
I say ‘potential’ limiting cases, because the question to consider is whether these cases are limiting cases of
purposive agency or cases in which one’s purposive agency is undermined.
20
Note, again, the similarity between this response and my discussion of minimal and maximal conceptions of
agency in the context of O’Neill’s justification of substantive norms of action.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
297
whether dispositional freedom and additive goods should be considered as generic features
of agency.21 This question will be the topic of the remainder of this section. Let me first
discuss the response to the range objection in terms of stressing the importance of both
dispositional freedom and additive goods.
Recall that Gewirth distinguishes dispositional freedom, i.e. the longer-range ability to
exercise control over one’s behaviour, from occurrent freedom, i.e. the freedom to control
one’s present behaviour by exercising one’s unforced choice (Gewirth 1978, 52). In line
with this, Gewirth admits that “a certain degree of voluntariness or freedom is exhibited
even in forced choice” (Gewirth 1978, 52). Even the slave has some degree of freedom.
What Gewirth, denies, however, is that dispositional freedom is respected in forced choices
and that therefore forced choice violates rights to the generic features of agency. Consider,
for instance, Gewirth’s discussion of imprisonment and slavery in Reason and Morality,
when he says that
a person’s freedom of movement is dispositionally interfered with if he is
imprisoned; here few or no alternatives as to his physical movements are left
available for his effective choice, and a whole host of other behaviors dependent on
his having freedom of movement are debarred to him. This is even more
drastically the case if a person is enslaved. For even if the slave has considerable
occurrent freedom of movement, he has lost control over his behaviour in a
deeper sense; for as a self or person to whom his choices would otherwise belong,
he is now the property of another and must hence do as the latter chooses, not as
he himself chooses. Thus both in imprisonment and in slavery dispositional
freedom is lost (Gewirth 1978, 254; see also Gewirth 1978, 52).
21
In what follows, I will thus focus primarily on dispositional freedom and additive goods and I will bracket the
idea of non-subtractive goods. The reason for this is that dispositional freedom and additive goods seem to have
the most potential to lead to a wide range of norms of action and because dispositional freedom and additive
goods are at the same time the most controversial generic features of agency. In addition, both Gewirth and
Beyleveld first and foremost refer to additive goods and dispositional freedom in the context of illustrating the
substantive potential of the principle of generic consistency. This, however, does not mean that I think that nonsubtractive goods are unimportant for the justification of substantive morality. Non-subtractive goods are
necessary to maintain the same level of purpose-fulfilment, and they might therefore include quite a wide range of
rights (such as rights to the provision of accurate information). A comprehensive analysis of Gewirth’s
justification of substantive morality would therefore also have to include an analysis of non-substractive goods.
This analysis would not only have to include an exploration of the precise content of non-substractive goods
(which I think is an underexplored element of Gewirth’s generic features of agency) but also the question of
whether non-substractive goods are indeed to be considered as generic features of agency in the first place (in line
with my discussion of additive goods and dispositional freedom below).
CHAPTER 9
298
So even though there are many (trivial) actions the prisoner or the slave can still pursue,
such as the choice of when and how to hum a song or when and how to stroll across his cell
or hut, etc., he does not have the longer-range ability to exercise control over his
behaviours. In other words, he might have occurrent freedom, but at the same time lack
dispositional freedom. 22 The generic features of agency are thus violated in cases of
imprisonment or slavery because individuals are denied their dispositional freedom.23
Gewirth and Beyleveld put forward a similar response by referring to additive goods.
Recall that Gewirth distinguishes between basic goods, goods that are necessary for
purposive agency, non-substractive goods, goods that are necessary to maintain the same
level of purpose-fulfilment, and additive goods, goods that increase one’s level of purposefulfilment (Gewirth 1978, 55). Although it might seem that one could be extremely
deprived or even starving and still be engaged in purposive action, this is only partly true,
because even though in such a case one might have basic goods, one obviously lacks
additive goods, and possibly non-subtractive goods. Beyleveld, for instance, writes that
“torture, etc., does interfere generically with my capacity to act, even if it removes it
completely only under extreme circumstances” (Beyleveld 1991, 83). And in a recent paper,
Beyleveld writes:
Generic conditions of agency do not necessarily have an all or nothing effect on
action or successful action. If they did there could be very few generic needs. In
fact, something is a generic condition if its absence to some degree diminishes or,
if its deprivation continues, will tend to diminish an agent’s ability to act or to act
successfully regardless of the agent’s purposes (Beyleveld 2011, 15).
In other words, even though one might still be an agent in a situation of extreme
deprivation, one would obviously lack additive goods such as education, self-esteem, access
to new information and opportunities for acquiring wealth and income. Thus, insofar as
additive goods are among the generic features of agency, the reference to additive goods
22
Beyleveld suggests that the other way around is also possible: “we can conceive of contingencies that present me
with only forced choices [i.e. occurrent unfreedom], [but] I could still have the dispositional capacity to choose
purposes freely” (Beyleveld 1991, 84). I struggle, however, to think what such a case would look like.
23
Gewirth also makes this point in reply to his critics. See e.g. Gewirth (1984, 33), Gewirth (1999, 195–96). See also
Beyleveld (1991, 86–88).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
299
allows the Gewirthian to rebut the range objection (analogous to the reference to
dispositional freedom).
Although in the potential limiting cases of purposive agency, agents might still have
the most basic generic features of action, they lack others (dispositional freedom and
additive goods). Insofar as dispositional freedom and additive goods are among the generic
features of agency, Gewirth can successfully argue for a wide range of rights and duties
(independent of their precise specification). The crucial question is thus as follows: why
should we include dispositional freedom and additive goods among the generic features of
agency?
If the generic features of agency are supposed to be the necessary means for engaging in
purposive action, it is not evident that either dispositional freedom or additive goods are
among the necessary means of agency, whereas basic goods clearly are among the necessary
means for engaging in purposive action. The reason for this is that, at least at first instance,
it might seem to be possible to pursue a purpose without necessarily wanting the freedom
to pursue purposes in the future (dispositional freedom) or to have the capabilities to
pursue more purposes in the future (additive goods).
Although Gewirth (and Beyleveld) has said little about the justification of both
dispositional freedom and additive goods, I think it is possible to identify and reconstruct
two distinct but closely related arguments for the inclusion of both dispositional freedom
and additive goods among the generic features of action. The first argument, which focuses
specifically on additive goods and not so much on dispositional freedom, is that additive
goods are needed for successful agency, i.e. having success in the pursuit of purposes. The
second argument is that both dispositional freedom and additive goods are needed for
prospective agency, i.e. the capacity for pursuing purposes not just in the present but also in
the future (I will explain and distinguish between different elements of prospectivity
below). These arguments could be interpreted as two ways of specifying the abstract concept
of purposive agency in such a way as to make clear why dispositional freedom and additive
goods are among the generic features of agency. I will discuss both arguments in turn. The
question I will ask is whether these respective specifications do indeed lead to the inclusion
of dispositional freedom and additive goods and, if so, whether these specifications of
agency are phenomenologically inescapable or whether they can be derived from the
abstract idea of purposive agency.
CHAPTER 9
300
4.4 Specifying Agency 1: Successful Agency
The first argument for including additive goods among the generic features of agency is
that additive goods are a necessary means to successful agency, i.e. that they are a necessary
means of achieving a purpose.24 Gewirth writes that
the generic features of successful action also compromise ... well-being in all three
of its dimensions [i.e. basic goods, non-subtractive goods and additive goods] ...
For in successful actions the agent achieves his purposes, so that he has not only
the minimal abilities required for any purposive pursuits but also the further
abilities and conditions needed for avoiding losses and making gain (Gewirth
1978, 62; see also Gewirth 1985, 237).
Beyleveld reconstructs the relation between purposive agency, successful purposive agency
and basic, non-substractive and additive goods in the following way:
I need my freedom and basic well-being in order to be able to act at all. No
behavior with the defining conditions of action can occur without these features,
and this is true by definition, as my freedom and basic well-being are the concrete
specifications of the defining conditions of voluntariness and purposiveness. The
other levels of well-being are not necessary for your purposivity as such, but they
are necessary for successful agency; so they are similarly definitionally linked to
successful agency, and this is definitionally linked to purposivity as such, because
PPAs [i.e. purposive agents] act in order to achieve their purposes (Beyleveld 1991,
82)25
The argument is thus, first, that purposive agency should be understood in terms of
successful purposive agency, because, as Beyleveld stresses, purposive agents act in order to
achieve their purposes. In other words, according to Beyleveld, successful agency is not an
24
As I have already noted above, this argument is specifically put forward to justify the inclusion of additive goods
and not so much for the inclusion of dispositional freedom. For reasons of space, I will not try to reconstruct
whether a similar argument can be made for the inclusion of dispositional freedom but will instead focus on
Beyleveld’s and Gewirth’s discussion of the relation between successful agency and additive goods.
25
Beyleveld makes the same point in more recent work: “Non-subtractive and additive needs are needed for
successful action rather than action per se” (Beyleveld 2011, 2f1).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
301
additional or contingent conception of agency; it simply specifies what it means to be an
agent, i.e. what it means to be someone who acts in order to achieve purposes. Second, the
claim is that additive goods should be understood as the goods that are necessary for
successful agency. So if agency should be understood in terms of successful agency, and if
additive goods are among the necessary conditions of successful agency, Gewirth can
answer the range objection.
Let me stress that I agree with Gewirth and Beyleveld that the generic features of
agency should indeed be understood in terms of the necessary means not just of agency
simpliciter but (also) of successful agency, insofar as to understand oneself as an agent just
means to understand oneself as a successful agent. For instance, it is not my purpose to try
to drink a coffee, but to actually drink a coffee, i.e. to actually achieve my purpose of
drinking coffee. If the coffee place were closed, I would not think I had achieved my goal
simply because I had achieved the goal of having tried to get a coffee. So, the specification
of purposive agency in terms of successful purposive agency is legitimate or even necessary,
because successful agency is simply entailed by the very idea of purposive agency (I take
this to be what Beyleveld means by saying it is “definitionally linked”). Successful agency is
thus not an optional conception of agency but a necessary part of our self-understanding as
purposive agents.
The question is, what are the necessary conditions of successful agency? To rebut the
range objection, the answer must be additive goods, i.e. those goods that increase one’s level
of purpose-fulfilment (and dispositional freedom, for a longer-range ability to control one’s
behaviour). Recall that Beyleveld gives the following examples of additive goods: “additive
needs include access to new information and opportunities to learn new skills” (Beyleveld
2011, 2f1). It is not obvious, however, that these kinds of additive goods are indeed
necessary for successful agency.
In arguing for the inclusion of additive goods among the necessary preconditions of
(successful) agency, Gewirth writes:
In acting for a purpose, the agent intends to achieve the purpose for which he acts
and hence to gain something that seems to him to be good. Such gain is an
increase in his level of purpose-fulfilment, and he necessarily regards it as good,
since it directly reflects the conative nature of his action, that for the sake of which
he puts forth whatever effort enters into his action (Gewirth 1978, 56).
CHAPTER 9
302
Gewirth thus claims that increasing one’s level of purpose-fulfilment is a necessary part of
action. However, although I agree with Gewirth that in successful action I intend to gain
something that seems good to me, I do not see why this something should be “an increase in
[my] level of purpose fulfilment”.26 It seems to me that what I necessarily intend to gain in
acting is the purpose of that kind of action, not a general increase in my level of purposefulfilment.
Recall that in order to establish what counts as a generic feature of agency we should
analyse what is necessary not just for a specific action but for engaging in certain types of
actions, i.e. one should ask which necessary means of, for instance, getting a coffee are not
just necessary for this specific activity but for action in general (this is the genericdispositional interpretation of the generic features). These generic means are the necessary
means of agency as such, as opposed to the necessary means of a particular action.
Now, say that I am trying to get a coffee. Why would it be necessary for me to not only
want the necessary means to engage in this kind of action (whatever ‘kind’ of action getting
a coffee is, exactly) but also want a general increase in my agential capacities so that I have
the capacity to achieve more ends in the future – the capacity to be able to get more coffee
(now or in the future) or to be able to engage in more kinds of activities – that is, to have
the means to not only get a coffee but also a piece of carrot cake to go with it? Why would it
be necessary to claim, as Beyleveld claims, a right to, for instance, new information and
opportunities to learn new skills insofar as I engage in activities such as getting a coffee? It
seems to me that one could reject the need for additive goods, i.e. an increase of one’s
purpose-fulfilling capabilities, without failing to understand oneself as an agent or as a
successful agent. In short, I think that successful purposive agency is a legitimate
specification of purposive agency, but that it fails to justify the inclusion of additive goods
among the generic features of agency.
4.5 Specifying Agency 2: Prospective Agency
Let me now turn to the second argument that Gewirth puts forward for the inclusion of
both additive goods and dispositional goods among the generic features of agency. The
second, and closely related, argument for the inclusion of dispositional freedom and
26
At least if this increase in the level of purpose-fulfilment is supposed to refer to something beyond the purposefulfilment as part of this specific action.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
303
additive goods among the necessary conditions of agency is to stress that the argument does
not simply focus on the self-understanding of an actual (successful) agent, but on the selfunderstanding of a prospective agent.27 In Reason and Morality, Gewirth defines prospective
agency as follows:28
The rational agent’s purview is not only occurrent but also dispositional; he
considers himself not only as a present agent but also as a prospective one who has
desires and purposes even when he is not currently acting. Thus [quoting Hobbes]
“the object of man’s desire is not to enjoy once only, and for one instant of time,
but to assure forever the way of his future desire”. Well-being requires that actions
be successful [quoting Aristotle] “in a complete life. For one swallow does not
make summer, nor does one day; and so too one day, or a short time, does not
make a man blessed and happy” (Gewirth 1978, 62).
[T]he agent claims [the rights to the generic features of agency] not only in his
present action with its particular purpose but in all his actions. To restrict to his
present purposes his reason for claiming the rights of freedom and well-being
would be to overlook the fact that he regards these as goods in respect of all his
actions and purposes, not only his present one. To be a prospective agent, then, is
not necessarily to be an actual agent; it is rather to have desires or goals one wants
or would want to fulfil through action. A prospective agent is so called not only
from this prospects, which at some points may be meagre, but also from his
prospecting – his occurrently or dispositionally looking ahead in some way to
acting for purposes he regards as good (Gewirth 1978, 111–12; see also Gewirth
1985, 239).
Gewirth thus claims that an agent not only understands him- or herself occurrently, i.e. as
someone who tries to achieve a certain purpose at this moment in time, but also
27
I say ‘closely related’ because it seems to me that the validity of the argument from successful agency
presupposes the argument from prospective agency, i.e. that the argument from successful agency presupposes
that one will have (potentially different) purposes in the future.
28
The following quotes are the only places in Reason and Morality in which Gewirth explicitly discusses the idea of
prospective agency (I will also briefly discuss Gewirth’s brief elaboration of the idea of prospective agency in a
more recent paper (Gewirth 1999) ). I think the lack of systematic discussion of the idea of prospective agency is
rather surprising given the importance of this concept for the substantive potential of his theory.
CHAPTER 9
304
diachronically or prospectively as someone who also tries to achieve purposes in the future,
“in a complete life” (I come back to the question of how exactly to understand this
diachronic element of agency below).
Prospective agents, Gewirth stresses in a more recent reply to critics, are “agents who
look beyond their present purposes toward further purposes they may have in the future”
(Gewirth 1999, 197). In addition, he claims that “any minimally prudent person will be
aware that his desires may change at some point or in some degree” (Gewirth 1999, 195).
Finally, he states that “agents care about and value their future purposes as well as their
present ones” (Gewirth 1999, 197). In the same paper, Gewirth stresses that “this point
about prospective agents is obviously of first importance in my whole doctrine of rational
agency” (Gewirth 1999, 197; see also Gewirth 1978, 62, 68)
If we have to understand ourselves as prospective agents, it seems that Gewirth might
be right in including both dispositional freedom and additive goods among the generic
features of (prospective) agency. For instance, Gewirth argues that generic-dispositional
freedom “is essential to what I have called being a prospective agent as well as an actual
agent” (Gewirth 1999, 196). The reason for this, Gewirth stresses, is that “as a prospective
agent, one values many freedoms besides the ones required for one’s present purposes”
(Gewirth 1999, 197).
A similar argument on the basis of prospective agency can be made for the inclusion of
additive goods. At one point in Reason and Morality, Gewirth acknowledges that “an agent
may perform some successful actions without having well-being in all three of its
dimensions [e.g. without additive goods]” (Gewirth 1978, 62). But he continues by saying
that “nevertheless, well-being in its three dimensions is a generic feature of successful
action when such action is regularly and predictably accomplished” (Gewirth 1978, 62).
The reason for this is that the agent “considers himself not only as a present agent but also
as a prospective one” (Gewirth 1978, 62).29 In other words, Gewirth claims that the generic
features of prospective agency include both dispositional freedom and additive goods (and
he more or less explicitly acknowledges that in the absence of the idea of prospectivity,
there might be no reason to include dispositional freedom and additive goods among the
necessary means of agency). So if we have to understand ourselves as prospective agents, it
seems to be initially plausible that dispositional freedom and additive goods should be
included among the generic features of agency, which implies that the range of generic
29
This also suggests a close link between successful agency and prospective agency.
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
305
features would be greatly enlarged.30 The slave or the prisoner would have to claim a right
to the dispositional freedom and additive goods, which suggests that at present his rights
are violated.
The crucial question is thus whether it is indeed inescapable to understand oneself not
just as an agent but also as a prospective agent, and what prospective agency precisely
entails. In other words, is prospective agency a legitimate specification of the idea of
purposive agency or a different and potentially escapable description of our selfunderstanding?
Let me start by saying that in a certain sense the idea of prospectivity is simply entailed
by the very idea of purposive agency; agency has a necessary diachronic dimension insofar
as any concrete action extends in time. Getting a coffee is not something that happens at
one particular time-slice but is something that takes time. In this sense, one needs a longerrange ability to exercise control over one’s behaviour, i.e. the longer-range ability that is
necessary to successfully achieve this purpose (e.g. the purpose of getting a coffee).
Gewirth’s idea of prospectivity, however, goes far beyond this truism about the
diachronic nature of action. Gewirth does not focus, first and foremost, on the diachronic
nature of one’s present actions, but on the possibility of being an agent “in a complete life”
(Gewirth 1978, 62). Although Gewirth does not give a clear definition of prospective
agency, his idea of prospective agency seems to have the following three elements (in
addition to the idea that any actual action has a diachronic component):
1) The idea that one not only has purposes in the present but also in the future, i.e.
the idea that agency (and not just action) is diachronic. Recall that “the rational
agent’s purview is not only occurrent but also dispositional; he considers himself
not only as a present agent but also as a prospective one who has desires and
purposes even when he is not currently acting” (Gewirth 1978, 62).
2) The idea that these purposes one has in the future might be different from the
purposes one has in the present. Recall that “any minimally prudent person will be
30
I say ‘initially plausible’ because one might object to the suggestion that, for instance, additive goods are indeed
necessary for prospective agency. I think that whether or not additive goods should be included among the
necessary conditions of prospective agency depends on how exactly one understands prospective agency. First, I
will discuss the different elements of prospectivity and then I will come back to the relation between different
elements of prospective agency and the inclusion of dispositional freedom and additive goods among the generic
features of agency in the next subsection.
CHAPTER 9
306
aware that his desires may change at some point or in some degree” (Gewirth
1999, 195).
3) The idea that the purposes of the agent might not only be different in the future,
but that the (present) agent also values his or her future purposes (whatever they
might be (see 2)). Recall that “agents care about and value their future purposes as
well as their present ones” (Gewirth 1999, 197).
In short, a prospective agent, on Gewirth’s definition of prospective agency, is someone
who understands him- or herself as also having purposes in the future (diachronic agency)
(1), as possibly having different purposes in the future (2), and as someone who also
(evaluatively) identifies with these purposes (whatever they might be) (3). I will refer to the
combination of these three elements as Gewirth’s ‘comprehensive’ conception of
prospectivity. Now the question is as follows: is this (comprehensive) conception of
prospective agency phenomenologically inescapable? Alternatively, is it entailed by
Gewirth’s abstract notion of purposive agency (just like the idea of successful agency was
entailed by the idea purposive agency)?
Recall that Gewirth gives two arguments for the phenomenological inescapability of
agency (see also chapter 4 and chapter 7). The first is that purposive action is the object of
all practical precepts (Gewirth 1978, 27). The second is that purposive action is
presupposed by all human practices (Gewirth 1978, 29). However, although I agree that
purposive action is the object of all practical precepts and that it is presupposed by all
human practices, I doubt whether this also necessarily applies to comprehensive
prospective agency, or least in the way that Gewirth defines it. For instance, although there
is always a forward-looking element in practical precepts insofar as they contain claims
about how one should or should not act, this kind of diachronicity is not the same as
Gewirth’s claim that one could have different purposes in the future and/or that one should
value one’s purposes whatever they might be.31 Gewirth, however, seems to disagree. He
links prospective agency to “the features that are constitutive of agency, features to which
are addressed the moral and other practical precepts that get my argument going in the first
place” (Gewirth 1999, 195). But, unfortunately, Gewirth never tries to show why these
31
Cf. “although it is correct that purposiveness analytically includes some transcendence beyond the present
moment, this move is not unlimited. It ends exactly at the point of having achieved the end or particular purpose”
(Illies 2003, 120).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
307
features (of comprehensive prospective agency) are indeed constitutive of (purposive)
agency.
Comprehensive prospective agency seems to be at least a potentially escapable
description of our self-understanding.32 There are individuals who do not identify with
their later selves (and there are competing philosophical theories about the question of
whether this is a coherent form of self-understanding (see e.g. Parfit 1984; G. Strawson
2004)). 33 Gewirth himself gives examples of these kinds of individuals: “nominalist
philosophers like William of Ockham [and] the ancient Megarians who held, according to
Aristotle’s Metaphysics, that only the actual present exists, that there is no continuity of
motion but only a series of discrete atomistic states, and that future tendencies and
potentialities, as such, have no reality” (Gewirth 1999, 198). In addition, there are
individuals who do identify with their later selves, but who deny that they should care about
their later selves in the same way as they should care about their present selves34 – one
might want to reply that these carpe diem persons are irrational, but this judgement
presupposes a thicker notion of rationality than the one explicitly adopted by Gewirth,
which includes only “the canons of deductive and inductive logic” (Gewirth 1978, 22).
Finally, there are individuals who do identify with and even care about their later selves but
for whom it is, at least from their own perspective, unthinkable that their purposes will
change in the way that Gewirth suggests. Whether or not it is, as Gewirth suggests, prudent
to think that one’s purposes may change at some point depends on how important a
particular purpose is for one’s identity. For instance, for someone who completely identities
with his or her role in society, his or her job or his or her role in the family, the thought that
one’s purposes might be different in the future might not even be conceivable without loss
of identity (cf. Williams 1973).
I therefore do not think that one could “reject this stage of the argument [i.e. the
inescapability of comprehensive prospective agency] only by disavowing and refraining
from all intentional action and from the whole sphere of practice” (Gewirth 1978, 29). I also
do not think that, as Gewirth suggests, denying comprehensive prospective agency is to
reduce oneself to “an automaton” (Gewirth 1999, 196). There simply are people who do not
32
I say ‘potentially’ because one might try to argue that those who reject one or more elements of Gewirth’s
definition of prospective agency misunderstand themselves. I will come back to this in the next subsection.
33
For a recent discussion of different conceptions of diachronicity and, closely related to this, narrativity see
Jongepier (2016).
34
Consider, for instance, Parfit’s Russian Nobleman (Parfit 1984, 326).
CHAPTER 9
308
look beyond their present purposes, or who do not care about the future (even if they look
beyond their present purposes) or for whom it is inconceivable that their purposes change.
Although (some of) these people might have an incoherent self-understanding, Gewirth
needs to say more to convince us of why comprehensive prospective agency is an
inescapable aspect of our self-understanding, even though there are many people who seem
to be capable of acting but who not understand themselves as a prospective agent, or at least
in the way in which Gewirth defines prospectivity.35
4.6 Phenomenological Inescapability and Specifications of Agency
As should have become clear from the discussion above, Gewirth’s comprehensive
conception of prospective agency contains different elements. Prospectivity refers to the
diachronicity of agency, to the idea that one’s future purposes might be different from one’s
present purposes and/or to the idea that one values one’s future purposes, just like one
values one’s present purposes. I claimed that, as it stands, Gewirth fails to argue in favour of
why we inescapably have to understand ourselves as comprehensive prospective beings,
because Gewirth’s general arguments for the inescapability of agency do not seem to apply
to prospective agency. In addition, I noted that there are individuals who deny all aspects of
prospective agency (except perhaps for the idea that a present action has a certain
diachronic nature), which puts further pressure on the idea that prospective agency, in the
way that Gewirth defines it, is inescapable.
However, the fact that there might be individuals who deny certain elements of
prospectivity does not have to mean that these self-descriptions provide a correct
description of their practical self-understanding and that therefore (some or all elements of)
prospectivity has to be rejected. There might, after all, also be people who explicitly deny
that they are agents at all (independent of the specification of agency), but this does not
mean that they do not actually understand themselves as agents. Or to give another
example, the fact that someone says that he or she does not have any experiences does not
necessarily mean that he or she does not actually have any experiences. People who deny
that they understand themselves as agents or experiencers simply seem to misunderstand
themselves. Whether or not a person understands him- or herself as an agent is not (just)
35
When confronted with a similar observation put forward to Donald Regan, Gewirth attempts to shift the burden
of proof by merely stating that “Regan has given no reason for this restriction to present purposes” (Gewirth 1999,
198).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
309
determined by what that person says but is shown in how they act (see also chapter 4).
The possibility that explicit self-reports or self-understandings do not necessarily
determine whether or not a certain conception of agency is phenomenologically
inescapable obviously raises the question of how to determine whether or not a certain selfunderstanding is phenomenologically inescapable. If the appropriate description of one’s
self-understanding is not necessarily determined by what one actually thinks or says, then
how can we decide whether or not a specific description of our self-understanding is
phenomenologically inescapable? This methodological question is all the more pressing if
the focus is not on an abstract conception of agency (e.g. purposive agency) but on
specifications of agency (e.g. prospective purposive agency).
The worry is that even though it might well be true that someone who denies that he or
she understands him- or herself an agent while ordering a cappuccino or being on the way
to work is simply mistaken about his or her self-understanding, it’s not clear how a similar
appeal to a ‘failure of self-understanding’ can be applied to someone who denies that he or
she will have purposes in the future or who denies that these purposes might be different,
and/or that he or she values these purposes, whatever they are (such as a person who
identifies strongly with his or her role in society, his or her job or his or her family). In
other words, how can one provide a convincing account of the phenomenological
inescapability of a specific conception of agency?
Because Kantian constructivists have paid insufficient attention to the specification of
agency, this question has, by implication, not received sufficient attention either. It is
impossible to provide a comprehensive answer to this fundamental question in the
remainder of this thesis. Instead, my point is mainly to illustrate that the question of the
phenomenological inescapability of specific conceptions of agency is underdeveloped (as
the discussion of Gewirth showed) and needs further research. Let me make a start,
however, in answering this question by briefly discussing, by way of example, one specific
element of Gewirth’s definition of prospectivity: the idea that an agent extends into the
future, or that agents (and not just actions) are necessarily diachronic.
As I mentioned in the previous section, there are people who deny that they will have
purposes in the future, because they deny that they will be the same person then (Parfit
1984; G. Strawson 2004). I think, however, that this idea of prospectivity can and should be
defended against those who are sceptical about personal identity over time. Although I
cannot fully argue for it here, those who are sceptical about personal identity over time
typically study personal identity from a third-person perspective. From this perspective,
CHAPTER 9
310
scepticism about diachronic personal identity might be warranted. From the first-person
perspective, however, it is clear that we understand ourselves as diachronic beings, i.e. as
beings who also have purposes in the future. Consider, for instance, Korsgaard’s discussion
of the diachronic nature of agency. She says that
most of the things we do that matter to us take up time. Some of the things we do
are intelligible only in the context of projects that extend over long periods. This is
especially true of the pursuit of our ultimate ends. In choosing our careers, and
pursuing our friendships and family lives, we both presuppose and construct a
continuity of identity and of agency. On a more mundane level, the habitual
actions we perform for the sake of our health presuppose ongoing identity. It is
also true that we think of our activities and pursuits as interconnected in various
ways; we think that we are carrying out plans of life (Korsgaard 1989, 113).
Korsgaard’s idea seems to be this: our actions have a diachronic nature insofar as they take
up time. Although trivial actions take up only a limited amount of time, the ‘pursuit of our
ultimate ends’ extends much further into the future, and most of our trivial actions only
make sense against the background of more ultimate ends. In addition, we pursue a
plurality of purposes which are somehow interconnected and together constitute a partly
overlapping chain of purposes which extends us into the future.
Against those like Parfit, who claims that he is not a diachronic agent in this sense, one
might thus point out that, even though he explicitly denies that he is a diachronic agent, he
is writing books and pursuing a career, i.e. he is pursuing purposes which project himself
into the future. In other words, one might point out that in his practical life, Parfit does
understand himself as a diachronic agent. Diachronic agency is proven to be inescapable
through a rational reconstruction of the actions he performs, not in the metaphysical
descriptions he gives of himself.
The point here is not so much to argue for this kind of diachronicity of agency, but
merely to note that it could be argued for and that even if we reject the more
comprehensive conception of prospectivity that Gewirth defends, there might still be some
truth in the idea that we have to understand ourselves as diachronic agents (one possible
element of prospectivity). On this restricted understanding of prospectivity (understood as
diachronic agency), the only non-prospective agent would be someone who has only one
purpose or only purposes which do not extend beyond the next few minutes. But I do not
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
311
know any human agents who are like that.36
Generalizing from this example, the way in which to argue for the phenomenological
inescapability of a specific conception of agency is thus to provide a rational reconstruction
of the first-person perspective of an individual, and not simply to rely on the self-reports of
individuals. The basic strategy is thus to argue for the inescapability of diachronic agency
by reference to our practical self-understanding, a self-understanding which, I have argued,
is not expressed in people’s self-reports but rather in how (and why) they act.
Importantly, this strategy is in principle also open to Gewirth. Gewirth might claim
that comprehensive prospective agency – the kind of agency which entails all possible
elements of prospectivity, including valuing your future purposes – is likewise expressed in
how we act, even if it isn’t expressed in what we say or how we think about ourselves.
Gewirth might thus claim that my earlier examples of people who have a carpe diem
mentality or who fully identify with their present purposes thus fail to be counter-examples
to comprehensive prospective agency.
However, although I think that one could in principle argue for Gewirth’s idea of
comprehensive prospective agency in this way, it seems much more radical, and much less
plausible, when the claim is that individuals actually value their future purposes, even
though they do not explicitly do so. It’s one thing to say, against someone like Parfit, that
insofar as he has the goal of writing a book or pursuing a career, he must necessarily
understand himself as a diachronic agent (a so-called restricted prospective agent). It’s
36
Illies objects to this kind of move. He writes that “even if it is likely that agents have among their purposes many
that extend quite far into the future, and even if there will be overlapping purposes (new purposes start before all
the old ones have ended), so that a patchwork of purposes will stretch to the end of the agent’s life, this is merely
an empirical matter of fact. It need not be the case. We need a philosophical argument as to why one should not
say rather modestly: ‘Take therefore no thought for the morrow: for the morrow shall take thought for the things
of itself. Sufficient unto the day is the evil thereof’ (Matt. 6:34). Gewirth would only succeed if he could show that
this thought is impossible; namely, that there is a contradiction in the notion of having merely very short-term
purposes. But he cannot” (Illies 2003, 120). In addition, Illies claims that “we need more than phenomenological
support for conceiving ourselves as agents who persist in time” (Illies 2003, 120). I am not sure how exactly to
interpret Illies’ claim (partly because he conflates the different aspects of prospectivity which I distinguished
above), but I think he is right to press Gewirth on his justification of the idea of prospective agency. I thus agree
with Illies that Gewirth has to justify his specification of agency, but I think that I disagree about what this
justification would have to amount to. Illies seems to suggest that the phenomenological inescapability of agency
would not be sufficient. Instead, Illies argues, one has to show that “there is a contradiction in the notion of having
merely very short-term purposes.” I fail to see, however, what this idea could possibly contradict with, if not with
one’s practical self-understanding (cf. the example of Parfit). Unfortunately, Illies says very little about how to
understand this idea of contradiction and why contradiction with one’s practical self-understanding would not be
sufficient. One possible explanation for this disagreement between Illies and me might be traced back to the fact
that Illies aims to use a transcendental argument to argue for moral realism (see also chapter 3). If one tries to
argue for moral realism, it is indeed problematic to rely on the ‘contingent’ phenomenology of human beings. But,
as I argued in chapter 3, a transcendental argument cannot be used to justify moral realism in the first place.
CHAPTER 9
312
quite another to say that, insofar as he is writing a book on personal identity or pursuing a
career, he must necessarily see himself as a comprehensive prospective agent and value his
future goals, whatever they might be (e.g. writing a book on gardening or baking cakes).
It is not evident that this comprehensive prospectivity would qualify as a plausible
reconstruction of our practical self-understanding. For instance, I fail to see why it would
be irrational not to value one’s future purposes (at least on an uncontroversial conception
of rationality). One might even wonder how one could possibly value future goals whose
content is still unknown in the present. Although these observations are by themselves
insufficient to conclude that we have to reject comprehensive prospectivity as a correct
description of practical self-understanding and/or what it means to act at all, I think that in
the absence of an argument to the contrary, this aspect of prospectivity cannot be used in a
transcendental argument for Kantian constructivism. So although what I have called
restricted prospective agency (diachronic agency) might be defended as an inescapable
aspect of our self-understanding, I think it is unlikely (although maybe not impossible) that
Gewirth’s comprehensive conception of prospective agency can be defended in a similar
way.
Of course, this brief discussion of prospectivity leaves many questions unanswered. It is
one thing to say that the truth conditions of a description of our phenomenology are not
determined by the self-reports of individuals, but it is quite another thing to give a
convincing account of the criteria which determine the rationality of a reconstruction of the
first-person perspective of an agent. What makes it correct to say that Parfit (or any other
agent) understands himself as a diachronic agent, even though he (or any other agent)
explicitly denies this? And how can I claim at the same time that Gewirth’s comprehensive
conception of prospectivity is unlikely to qualify as a rational reconstruction of an
inescapable aspect of our phenomenology? Discussing the specification of agency thus
raises fundamental methodological questions, which I cannot answer within the scope of
this thesis, but which should be answered in order to provide a comprehensive account of
Kantian constructivism.37
I will briefly discuss the implications of this discussion for Kantian constructivism in
the next and final section of this chapter. Let me begin this section, however, by briefly
37
One important question is the extent to which the conceptions of agency in Kantian constructivism express
more than just specific, contingent cultural ideas. In this context, the confrontation with different cultural and
intellectual traditions is crucial (see e.g. Flikschuh 2014; Flikschuh forthcoming).
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
313
discussing the relation between the different elements of prospective agency and the
contents of the generic features of agency.
4.7 The Generic Features of Prospective Agency
In the discussion of prospective agency, I have assumed that comprehensive prospective
agency would justify the inclusion of dispositional freedom and additive goods among the
generic features of agency. However, if one rejected one or more elements of prospective
agency (which I think we should), one might wonder whether there is still a reason to
include additive goods and dispositional freedom among the generic features of agency.
To provide an illustration of the relation between the appropriate specification of
agency (i.e. the different elements of prospectivity) and the generic features of agency,
consider Beyleveld’s definition of prospective agency. Interestingly, Beyleveld also seems to
defend a more restricted conception of prospective agency than Gewirth. Beyleveld
disagrees with Gewirth about the question of whether individuals (must) value their future
purposes. He writes:
“My freedom and well-being” ... refers to general abilities or capacities that I
require to pursue/achieve any of my purposes whatsoever. ... to say that these are
general (generic) capacities is to say that they are the same capacities whatever my
purposes might be. So, whether or not I must now value purposes I will (or might)
have, but do not as yet have, I must value these same general capacities, whatever
purposes I value at t or at t+n. It follows that, in order to show that I must value
my freedom and well-being, whatever my purposes, Gewirth does not have to
suppose that I must value my future purposes (which I don’t now have) (Beyleveld
1991, 85).
So, the suggestion is that on Beyleveld’s conception of prospectivity, when I have the goal of
getting myself a coffee, I do not need to value my own future goals. Beyleveld’s idea of
prospectivity seems to be slightly different, and less controversial, than Gewirth’s idea of
prospectivity. This is also evident from the following description of a prospective agent:
We use the term “agent” to cover also those, strictly “prospective agents”, who,
while not actually pursuing purposes voluntarily, have the developed capacity to
CHAPTER 9
314
do so. This characterisation of agency is not based on the empirical features of
human or any other beings. It is based on the characteristics that any beings that
are intelligible subjects of practical precepts (those who issue practical precepts)
and objects of practical precepts (those to whom practical precepts are directed)
must be supposed to have (Beyleveld and Brownsword 2007, 39f1).
A prospective agent, on Beyleveld’s definition, is thus someone who has the general,
developed capacity to pursue purposes (now and in the future). Beyleveld’s definition of
prospective agency is closer to diachronic agency than to Gewirth’s conception of
comprehensive agency, which includes the valuing of future purposes.
If the generic features of agency should be understood in terms of the capacities that
are necessary to be an agent not just at this point but also in the future (diachronic agency),
this still seems to justify the inclusion of dispositional freedom among the necessary
conditions of agency. After all, if one understands oneself necessarily as a restricted
prospective agent, one not only needs the freedom to control one’s behaviour at this
moment (occurrent freedom) but also the longer-range ability to do so (dispositional
freedom). This conclusion is hopeful, for it means that at least dispositional freedom can be
justified, even if one does not accept Gewirth’s comprehensive conception of prospective
agency.
The question, however, is whether this more restricted conception of prospectivity
leads to the inclusion of additive goods among the generic conditions of agency. It is at least
not obvious that in order to have the general capacity for agency that Beyleveld talks about,
one needs to improve or enhance one’s agential capabilities (additive goods). Additive
goods entail the goods that are necessary to increase one’s level of purpose-fulfilment. The
problem is that there only seems to be a reason to want this if one (contingently) thinks that
one’s purposes might be different in the future, combined with the idea that one not only
values one’s present purposes but also one’s future purposes. Why is it necessary to claim a
right to additive goods if one does not think that one might have different or more
purposes in the future which would potentially require more or enhanced capacities and/or
if one fails to evaluatively identify with one’s future self? In the absence of these reasons, it
is unclear why additive goods should be included among the generic features of agency. In
short, the specific conception of prospective agency that one adopts will have implications
for the generic features of agency and, more specifically, for the question of whether
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
315
dispositional freedom and additive goods should be included among the generic features of
agency.
Let me try to briefly summarize this very long section. I have critically evaluated
Gewirth’s attempt to justify a wide range of substantive norms of action on the basis of the
principle of generic consistency. I have tried to show that Gewirth’s success in justifying the
wide range of norms which he himself defends hinges on the question of whether
dispositional freedom and additive goods should be included among the generic features of
agency and that this question in turn depends on the question of how exactly we have to
specify the idea of purposive agency. Subsequently, I have discussed two specifications of
agency: successful agency and prospective agency. I have argued that although successful
agency is simply entailed by the abstract idea of purposive agency, it does not have
dispositional freedom or additive goods among its necessary conditions of possibility.
Finally, I have argued that although Gewirth’s comprehensive conception of prospective
agency might have dispositional freedom and additive goods among its conditions of
possibility, Gewirth says very little in defence of prospective agency, and I have tried to
explain why I am pessimistic about the possibility of justifying Gewirth’s comprehensive
conception of prospective agency. Finally, I have argued that there might be more restricted
forms of prospective agency of which it might seem more plausible that they are
phenomenologically inescapable, and I have discussed the implications of a more restricted
conception of prospective agency for the generic features of agency.
5. The Emptiness Objection Revisited
In the previous sections, I have discussed O’Neill’s and Gewirth’s application of a supreme
principle of morality in order to evaluate to what extent Kantian constructivists succeed in
justifying a wide range of norms of action. As should have become clear by now, I think
that the reliance on both general and contingent anthropological facts is necessary but not
sufficient to justify substantive norms of action and that the substantive potential of
Kantian constructivism hinges on the success of specifying a conception of agency.
Although even on the basis of an abstract conception of agency certain norms of action can
be justified (for instance the prohibition of murder), a specification of agency is needed to
move beyond the most minimal moral norms. As I have tried to show, this task – the task
of specifying agency – has not received sufficient attention from Kantian constructivists.
This means that, as it stands, they fail to justify the substantive moral theory that they
CHAPTER 9
316
themselves defend. In this sense, the critics of Kantian constructivism are right: there is a
certain tendency to rely on an unvindicated and potentially escapable conception of agency
in the application of a supreme principle of morality.
This, however, neither means that Kantian constructivism is (too) empty nor that it is
impossible to justify the wide range of norms that Kantian constructivists typically defend
through a transcendental argument. With respect to the former, it might simply turn out
that the range of norms that a transcendental argument can justify is more limited than
Kantian constructivists themselves typically think or hope for. As I have argued in the
previous chapter, there is, however, no independent standard against which to judge
whether or not a moral theory is sufficiently substantive. So even if the range of norms of
Kantian constructivism is more restricted than Kantian constructivists themselves seem to
believe, this is not an objection to Kantian constructivism.
Pace the critics of Kantian constructivism, however, it is way too early to consider
biting this bullet and conceding that Kantian constructivism cannot justify norms of
actions beyond the most trivial or minimal norms. What I have mainly shown is that the
justification of substantive norms of action requires much more work and that Kantian
constructivists have not sufficiently discussed the specification of an abstract idea of agency.
But it does not follow that it is impossible to justify a more specific conception of agency
than the abstract conception of rational or purposive agency. Although the question of
what constitutes a correct reconstruction of our practical self-understanding deserves more
attention, there are, as I have suggested, good reasons to think that restricted prospective
agency is phenomenologically inescapable, even if the comprehensive idea of prospective
agency that is defended by Gewirth seems to be merely an optional form of selfunderstanding.
There is another important point to make, namely that there is no reason to think that
a more minimal specification is necessarily less controversial than more maximal
conceptions of agency. It’s just not obvious that the more minimal our conception agency
is, the more plausible it is. For all specifications of agency are in need of justification,
including minimal specifications of agency. For instance, I think that it is at least just as
controversial to think that an agent does not have any purposes in the future (cf. Parfit
1984; G. Strawson 2004) as to suppose that an agent necessarily has to think that he might
have different purposes in the future (cf. Gewirth). This is not an attempt to shift the
burden of proof, but a way to take seriously the task of specifying a conception of agency.
The point that an abstract conception of agency is too abstract to lead to determinate
THE EMPTINESS OBJECTION AND KANTIAN CONSTRUCTIVISM
317
norms of action does not imply that we should therefore settle for the most minimal
interpretation of agency. Rather, it means we have to explore which description of agency
provides the most plausible reconstruction of our self-understanding. I doubt whether
minimal conceptions of agency are the most plausible reconstructions of our selfunderstanding, even if we agree with the critics of Kantian constructivism that the maximal
conceptions of agency that are implicitly presupposed by Kantian constructivists are
likewise controversial.
There is a huge gap between inescapable, abstract conceptions of agency and
potentially escapable specific conceptions of agency which, at present, is underexplored.
Both Kantian constructivists and their critics should try to show why a certain specific
conception of agency should be preferred over other possible specifications of agency.
In the current debate, there seems to be little room between radical pessimism on the
one hand and radical optimism about the substantive implications of Kantian ethics on the
other. Critics of Kantian ethics generally think that it is a mystery how the categorical
imperative could possibly lead to any substantive norms of action (recall, for instance,
Sharon Street: “the rabbit of substantive reasons can[not] be pulled out of a formalist hat”
(2010, 370)). For proponents of Kantian ethics, on the other hand, it often seems as if it is
evident that Kantian ethics can lead to substantive norms of action, at least when it is
acknowledged that Kantian ethics relies on anthropological facts and allows the idea that
specific norms of actions might differ according to the specific socio-historical contexts to
which the categorical imperative (or its Kantian constructivist variants) is applied. So,
inevitably, Kantian constructivists will think that I am too pessimistic about the substantive
implications of Kantian constructivism. Critics of Kantian constructivism, on the other
hand, will think that I am too optimistic about its substantive implications. However, what
follows from the foregoing is that the space in between these two radical extremes is worth
exploring. This is the reason why I said in the introduction to this chapter that the
conclusion of this chapter is moderately optimistic or moderately pessimistic, depending
on one’s perspective. I urge that more attention ought to be given to the position in
between these radical ends of the spectrum.
6. Conclusion
In this chapter, I have tried to show that the reliance of Kantian constructivists on
anthropological facts in the justification of substantive morality is not sufficient to respond
CHAPTER 9
318
to the emptiness objection. The reason for this is that the anthropological facts are not
morally relevant in and by themselves, but only insofar as they tell us something about the
necessary preconditions of agency. As I have tried to show in this chapter, this implies that
the central question for Kantian constructivists in the context of the justification of
substantive morality is the question of how exactly to understand agency. I have tried to
argue for this point by providing a detailed analysis of the way in which two prominent
Kantian constructivists, Alan Gewirth and Onora O’Neill, have tried to justify substantive
norms of action on the basis of a supreme principle of morality. I have argued that they
implicitly or explicitly rely on a rather maximal conception of agency in the justification of
substantive morality which stands in need of further justification.
As long as the specification of the abstract conception of agency is not made explicit or
defended, critics are right to complain that it is unclear how substantive norms of action
can be justified on the basis of a supreme principle of morality. However, at the same time,
critics are only justified in their scepticism about the substantive potential of Kantian
constructivism once it is concluded that there is no (interesting) specific conception of
agency that is phenomenologically inescapable.
The conclusion of this chapter is thus that debates about the emptiness objection can
only move forward by discussing conceptions of agency and their phenomenological
(in)escapability. This neither counts as a defence of the substantive ambitions of Kantian
constructivism nor as a criticism of this approach. Instead it shows where the action should
be.
Chapter 10
General Conclusions
1. Summary
I started this thesis with the following question: is it possible to justify objective (i.e.
universal and categorical) moral claims without relying on moral facts? As I have already
mentioned in the introduction to this thesis, an affirmative answer to this question may be
considered to be “the holy grail” of ethics (Street 2016, 165). Although my answer to this
question is not unequivocally affirmative, I go against the grain of mainstream ethics in that
I am moderately optimistic about the possibility of providing a non-realist, constructivist
justification of moral objectivity. That is, a justification of morality which does not rely on
moral facts, but instead justifies morality by reference to the first-person perspective of an
agent. The reason for my moderate optimism is that I think that there is an alternative
method of moral justification to the method of reflective equilibrium (and to moral
intuitionism), namely transcendental argumentation. I have referred to this combination of
moral objectivity, constructivism and transcendental argumentation as ‘Kantian
constructivism’.
I have argued that on a Gewirthian interpretation, a so-called transcendental argument
from the first person can justify a universal and categorical principle of interpersonal
morality. Such a transcendental argument analyses what an agent (i.e. someone who acts to
achieve a certain purpose) is necessarily committed to from his or her first-person
perspective as someone who understands him- or herself as an agent. Moral normativity,
according to this approach to moral justification, follows from what it is to understand
oneself as someone who pursues purposes in the world.
CHAPTER 10
320
This practical self-understanding leads to certain normative commitments. One can
only intelligibly pursue purposes if one accepts the principle of instrumental rationality, i.e.
the principle that one ought to take the necessary means to one’s ends or give up the end.
The instrumental principle, combined with the observation that there are not only means
that are contingent on pursuing a particular end but also means that are necessary to
pursue any purpose whatsoever, leads to the categorical commitment that you ought to take
the means that are necessary to engage in any action whatsoever, and not just the necessary
means to one’s particular purposes. This, I have argued, ultimately amounts to saying that
one ought to claim a right to the necessary means of agency. This right is universalized
through ‘the argument from the sufficiency of agency’, which states that agency is a
sufficient condition of having a right to the necessary means of agency and that one
therefore necessarily has to accept that any agent has a right to the necessary means of
agency.
A transcendental argument justifies, first and foremost, a supreme principle of morality
which grants an equal right to the necessary means of agency to all agents. Substantive
morality should, subsequently, be understood in terms of the realization and protection of
the necessary means of agency. The identification of the necessary means of agency requires
both the incorporation of anthropology, i.e. general and contingent anthropological
considerations about human nature and, more importantly, the specification of what
exactly it means to understand oneself as an agent.
The reason that I am only moderately optimistic about the Kantian constructivist
project has to do with the question of substantive morality and more specifically with the
identification of the necessary means of agency. Although I have argued that the strong
emptiness objection to Kantian constructivism fails (the objection that no single substantive
norm of action can be justified), much more work needs to be done in order to justify a
wide range of substantive norms of action on the basis of a supreme principle of morality.
More specifically, I have argued that the conception of agency in Kantian constructivism is
underexplored and underspecified. This is where the action should be in the debate about
Kantian constructivism.
The main argumentative strategy that I have pursued throughout this thesis in order to
reach the above conclusions could be understood as an argument by elimination. I have
first of all criticized the most dominant approach to moral justification in ethics, the
method of reflective equilibrium, or at least insofar as it is used to justify objective moral
conclusions (chapter 2). In the remainder of the thesis, I have evaluated to what extent
GENERAL CONCLUSIONS
321
existing transcendental arguments can succeed where reflective equilibrium fails. First, I
have argued that transcendental arguments cannot be used to justify moral realist
conclusions and that insofar as transcendental arguments are understood against the
background of constructivism, transcendental arguments in ethics do not face the kinds of
problems that are typically associated with transcendental arguments in theoretical
philosophy (chapter 3).
Subsequently, I have evaluated existing transcendental arguments for constructivist
moral objectivity in light of what I have called the potential general dilemma for Kantian
constructivists. The dilemma is that either the starting point of a transcendental argument
is inescapable, but in that case no interesting moral conclusions can follow from the
starting point, or Kantian constructivism succeeds in justifying interesting moral
conclusions – but in that case the starting point turns out to be merely optional. I have
divided these ‘interesting moral conclusions’ into three categories: success in justifying any
categorical commitments at all (chapter 4), success in justifying categorical and universal
commitments, i.e. principles of interpersonal morality (chapters 5, 6 and 7), and success in
justifying substantive moral commitments (chapters 8 and 9). Although several existing
transcendental arguments are susceptible to one or both horns of the general dilemma, an
approach developed along Gewirthian lines is a lot more promising.
2. Confronting my Interlocutors
This thesis moved from abstract discussions about moral methodology in meta-ethics to
fundamental principles of morality and ultimately to the question of the possibility of
justifying substantive moral norms of action. As a result of this wide scope, this thesis
engages with a wide range of interlocutors, both directly and indirectly. By way of
conclusion, I will therefore try to summarize some of the main points of the thesis by
briefly distinguishing the most important interlocutors in this thesis. Although this way of
summarizing will inevitably lack some nuance, I hope that this lack of nuance will be
compensated for by the simplicity of this way of summarizing and by the more nuanced
discussion of different positions in the body of this thesis. In addition, I think this might be
a helpful way to illustrate both the achievements and the limits of this thesis project, while
abstracting from the detailed discussion in the body of this thesis. I end with some
suggestions for future research.
One of the most important (direct) interlocutors in this thesis is the proponent of
CHAPTER 10
322
reflective equilibrium, whether working in applied ethics, normative ethics or meta-ethics. I
have tried to give reasons to the proponent of reflective equilibrium for being sceptical
about the idea that the method of reflective equilibrium would be able to justify objective
moral judgements, as well as the idea that reflective equilibrium would be the only game in
town, even if one assumes the failure of epistemological intuitionism. In addition, I hope to
have seduced proponents of reflective equilibrium to take transcendental argumentation
more seriously, primarily because the epistemological commitments of transcendental
argumentation are not fundamentally different from the commitments of transcendental
argumentation and because a transcendental argument can succeed in justifying objective
moral claims.
The second main interlocutor is the person who is sceptical about Kantian
constructivism, who thinks that Kantian constructivism is susceptible to one or both horns
of the general dilemma for Kantian constructivism. Certainly, those who are sceptical about
Kantian constructivism constitute a very diverse group of interlocutors (some more specific
interlocutors are mentioned below). What they share is the basic thought that Kantian
constructivism either relies on optional starting points or it cannot justify any interesting
normative or moral conclusions. It is impossible to give an adequate summary of my reply
to this interlocutor, because my reply basically covers the whole thesis. But, very crudely,
my reply is that although critics are right that some forms of Kantian constructivism are
susceptible to one or more horns of the dilemma, not all Kantian constructivists are, and
hence, Kantian constructivism as such does not need to be. In addition, I have argued that
although Kantian constructivists can justify a supreme principle of morality, the general
dilemma has most force in the context of discussions regarding substantive morality. Pace
the critics of Kantian constructivism, however, I do not think that the dilemma undermines
the substantive ambitions of Kantian constructivism, but that it does show that Kantian
constructivists have to pay more attention to the specification of their conception of
agency.
More indirectly, I have also tried to address both the moral relativist and the moral
realist in this thesis. Against the moral relativist, I have claimed that moral relativism
should only be accepted if we think that objective moral judgements cannot be justified.
Moral relativism is first and foremost a negative thesis which denies that there is an
objective moral standard. In other words, the argument against moral relativism just is the
argument for Kantian constructivism. Things are slightly different in the case of moral
realism. I have not tried to show that moral realism should be rejected. Indeed, I have
GENERAL CONCLUSIONS
323
argued that Kantian constructivist views are stronger if they don’t start from such a
rejection, but that there is a way to justify moral judgements without relying on the
existence of moral facts. This project should be of interest for moral realists, because even
realists themselves typically acknowledge that realism has certain metaphysical and
epistemological problems (even if they ultimately think that these problems can be solved).
So if these problems could be avoided altogether, then that approach should be preferred.
In response to those who are generally sceptical about transcendental argumentation, i.e.
philosophers who think that transcendental arguments cannot work independent of their
usage in theoretical philosophy or in ethics, I have argued that there are important
differences between transcendental arguments in theoretical philosophy and in ethics and
that the most common objections to transcendental arguments in theoretical philosophy do
not apply to transcendental arguments in ethics, at least when they are understood against a
constructivist background. Although this implies that transcendental arguments cannot be
used to justify moral realist conclusions, this is not a problem because it is not clear that
moral realism is the most desirable moral metaphysics in the first place. The main point,
though, is that there is no reason to be sceptical about the Kantian constructivist project on
the basis of scepticism about transcendental arguments in theoretical philosophy.
Let me now turn to the intramural disagreements among Kantian constructivists. I
have argued that what I have called transcendental arguments from the second person are
susceptible to the general dilemma for Kantian constructivism. These arguments, of which
different varieties are put forward by Karl-Otto Apel, Onora O’Neill, Christine Korsgaard
and Stephen Darwall, either fail to justify a categorical and universal principle of
interpersonal morality or they start from an inescapable form of interaction (such as
argumentation, communication or a second-person relationship).
Christine Korsgaard deserves a special mention at this point as one of my specific
Kantian constructivist interlocutors. I have distinguished, reconstructed and criticized
several arguments she puts forward for a categorical and universal principle of
interpersonal morality, i.e. the universal value of humanity. Both her argument(s) from the
second person and her argument(s) from the first person fall prey to one or both horns of
the general dilemma for Kantian constructivism. In the literature on Kantian
constructivism, Korsgaard is often considered to be the most prominent Kantian
constructivist, or even the only Kantian constructivist worth discussing. If my criticisms of
Korsgaard’s work are valid, I think there are good reasons to broaden the focus of this
discussion.
CHAPTER 10
324
The second person who deserves special mention is Alan Gewirth (and the further
development of Gewirth’s theory by Deryck Beyleveld). I believe, at least on Beyleveld’s
reconstruction of Gewirth’s theory, that Gewirth succeeds in justifying a supreme principle
of morality. This is a surprising conclusion insofar as Gewirth is hardly ever seriously
discussed in the mainstream ligature on Kantian constructivism, and because the critical
reception of Gewirth’s theory has been predominantly negative.
I am, however, more critical about the potential of a Gewirthian application of the
supreme principle of morality. The final interlocutors, then, are Kantian constructivists
who have attempted to apply a supreme principle of morality in order to justify substantive
norms of action; more specifically, my interlocutors have been Onora O’Neill and Alan
Gewirth himself. My response to what one might call the substantive Kantian constructivist
is that although I agree that Kantian constructivism can justify at least some substantive
norms of action, more work needs to be done in order to justify the wide range of norms of
action which Kantian constructivist typically propose.
The identification of these interlocutors highlights at the same time the limits of this
thesis, because obviously there are many potentially relevant interlocutors whom I have not
directly or indirectly addressed in this thesis. One of these interlocutors deserves special
mention: the particularist who denies that abstract principles or norms can guide us in
making moral judgements about specific cases. For reasons of space, my discussion of
substantive morality has been limited to the justification of norms of action types. The
question of how both a supreme principle and a norm of action can guide someone in
making judgements about specific actions or cases, ‘the problem of moral judgement’, is a
topic which I think has not received sufficient attention and which I hope to work on in the
future.
In addition, I have said very little in response to those who are sceptical about
instrumental rationality. In the Gewirthian transcendental argument defended in this
thesis, the principle of instrumental rationality and the transcendental justification of this
principle proved to be fundamental in the justification of morality. Although I have briefly
discussed the justification of the instrumental principle in chapter 7, I have not provided a
comprehensive defence of this principle. I think the (transcendental) justification of the
instrumental principle and the relation between the instrumental principle and morality are
in need of further elaboration and defence.
Finally, it should be clear that another important lacuna, as I have already mentioned
several times throughout the thesis, is that I have not discussed the question of how Kantian
GENERAL CONCLUSIONS
325
constructivism relates to Kant’s own views. A comprehensive defence of Kantian
constructivism, however, will have to engage with the question of the extent to which the
position is truly Kantian or not. Many of the authors discussed in this book not only put
forward a systematic Kantian constructivist position in ethics but also claim that their
theory provides the best interpretation or reconstruction of Kant’s ethics. I have tried to
steer away from these exegetical controversies in this thesis in order to try to make
systematic progress.
I am sure there are many more potentially relevant interlocutors which I have failed to
consider or to mention in this thesis. But I hope to have shown that even though many
questions remain unanswered, Kantian constructivism, including transcendental
argumentation, is a more convincing position than is often assumed.
Although much of the discussion in this thesis has been critical of many Kantian
constructivist theories, and even though I do not unequivocally declare victory for Kantian
constructivism, I like to think that ultimately this thesis contains a hopeful message. The
hopeful message is that there really might be a rational basis for morality. And even though
it is not immediately clear which concrete norms of action can be justified through the
application of a supreme principle of morality, it has at least become clearer, or so I hope,
what this project of application amounts to, and hence what the landscape of future
research looks like.
Let me finish by mentioning two lines of future research, in addition to the justification
of the instrumental principle and the question of moral judgement, which I have already
mentioned above. Some of the crucial questions arising from this thesis are the question of
how exactly to specify agency and, more fundamentally, the methodological question of
how to determine whether a specific description of our practical self-understanding is
‘phenomenologically inescapable’ or not, and how, exactly, such inescapability must be
understood. These questions lie at the heart both of transcendental argumentation as a
method of (moral) justification (insofar as transcendental arguments start from a
phenomenologically inescapable starting point) and of the evaluation of the substantive
moral potential of Kantian constructivism.
Another fundamental issue that requires more research is the relation between metaethics and normative ethics. This thesis understands the relation as one of interdependence:
Kantian constructivism is not just a meta-ethical theory but also a normative ethical theory.
However, I have not tried to argue for this interdependency thesis directly, but instead I
have tried to show how fundamental questions about moral methodology, the justification
326
CHAPTER 10
of fundamental moral principles and the justification of substantive norms of action are
crucially interrelated. This raises fundamental questions about the unity of ethics, which I
could not address in this thesis but which I also hope to work on in the future.
Bibliography
Alexy, Robert. 2006. “Effects of Defects-Action or Argument? Thoughts about Deryck
Beyleveld and Roger Brownsword’s Law as a Moral Judgment.” Ratio Juris 19 (2):
169–79.
Altehenger, Hannah, Simon Gaus, and Andreas Leonhard Menges. 2015. “Being Realistic
about Reflective Equilibrium.” Analysis 75 (3): 514–22.
Ameriks, Karl. 1978. “Kant’s Transcendental Deduction as a Regressive Argument.” KantStudien 69 (1–4): 273–87.
———. 2003. Interpreting Kant’s Critiques. Oxford: Oxford University Press.
Anscombe, G. E. M. 2000. Intention. Cambridge, MA: Harvard University Press.
Apel, Karl-Otto. 1975. “The Problem of Philosophical Fundamental-Grounding in Light of
a Transcendental Pragmatic of Language.” Man and World 8 (3): 239–75.
———. 1980. “The a Priori of the Communication Community and the Foundations of
Ethics: The Problem of a Rational Foundation of Ethics in the Scientific Age.” In
Towards a Transformation of Philosophy, 225–300. Milwaukee: Marquette
University Press.
———. 2005. “Regarding the Relationship of Morality, Law and Democracy: on
Habermas’s ‘Philosophy of Law.’” Invenio 8 (15): 11–24.
Arras, John D. 2007. “The Way We Reason Now: Reflective Equilibrium in Bioethics.” In
The Oxford Handbook of Bioethics, edited by Bonnie Steinbock, 46–71. Oxford:
Oxford University Press.
Audi, Robert. 2005. The Good in the Right: A Theory of Intuition and Intrinsic Value.
Princeton: Princeton University Press.
Bagnoli, Carla. 2014. “Constructivism in Metaethics.” In The Stanford Encyclopedia of
Philosophy, edited by Edward N. Zalta.
http://plato.stanford.edu/archives/spr2014/entries/constructivism-metaethics/.
Barry, Melissa. 2013. “Constructivist Practical Reasoning and Objectivity.” In Reading
SAMENVATTING
328
Onora O’Neill, edited by David Archard, Monique Deveaux, Neil Manson, and
Daniel Weinstock, 17–36. Abingdon: Routledge.
Bauhn, Per. 2016. Gewirthian Perspectives on Human Rights. Abingdon: Routledge.
Bealer, George. 1996. “A Priori Knowledge and the Scope of Philosophy.” Philosophical
Studies 81 (2–3): 121–142.
Beall, J.C., and Michael Glanzberg. 2014. “Liar Paradox.” In The Stanford Encyclopedia of
Philosophy, edited by Edward N. Zalta.
http://plato.stanford.edu/archives/fall2014/entries/liar-paradox/.
Besch, Thomas M. 2008. “Constructing Practical Reason: O’Neill on the Grounds of
Kantian Constructivism.” Journal of Value Inquiry 42 (1): 55–76.
Beyleveld, Deryck. forthcoming. “What Is Gewirth and What Is Beyleveld? A Retrospect
with Comments on the Contributions.” In Ethical Rationalism and the Law, edited
by Patrick Capps and Shaun D. Pattinson. Oxford: Hart Publishing.
———. 1991. The Dialectical Necessity of Morality. An Analysis and Defence of Alan
Gewirth’s Argument to the Principle of Generic Consistency. Chicago: The
University of Chicago Press.
———. 2011. “The Principle of Generic Consistency as the Supreme Principle of Human
Rights.” Human Rights Review 13 (1): 1–18.
———. 2013. “Williams’ False Dilemma: How to Give Categorically Binding Impartial
Reasons to Real Agents.” Journal of Moral Philosophy 10 (2): 204–26.
———. 2015. “Korsgaard v. Gewirth on Universalization: Why Gewirthians Are Kantians
and Kantians Ought to Be Gewirthians.” Journal of Moral Philosophy 12 (5): 573–
97.
———. 2016. “Gewirth Versus Kant on Kant’s Maxim of Reason: Towards a Gewirthian
Philosophical Anthropology.” In Gewirthian Perspectives on Human Rights, edited
by Per Bauhn, 13–29. Abingdon: Routledge.
Beyleveld, Deryck, and Gerhard Bos. 2009. “The Foundational Role of the Principle of
Instrumental Reason in Gewirth’s Argument for the Principle of Generic
Consistency: A Response to Andrew Chitty.” King’s Law Journal 20 (1): 1–20.
Beyleveld, Deryck, and Roger Brownsword. 2001. Human Dignity in Bioethics and Biolaw.
Oxford: Oxford University Press.
———. 2007. Consent in the Law. Oxford: Hart Publishing.
Boylan, Michael. 1999. Gewirth: Critical Essays on Action, Rationality, and Community.
Lanham: Rowman & Littlefield.
SAMENVATTING
329
Brandt, Richard. 1979. A Theory of the Good and the Right. Oxford: Oxford University
Press.
Brink, David O. 1989. Moral Realism and the Foundations of Ethics. Cambridge: Cambridge
University Press.
———. 2014. “Principles and Intuitions in Ethics: Historical and Contemporary
Perspectives.” Ethics 124 (4): 665–94.
Brun, Georg. 2014. “Reflective Equilibrium Without Intuitions?” Ethical Theory and Moral
Practice 17 (2): 237–52.
Budde, Kerstin. 2009. “Constructivism All the Way down – Can O’Neill Succeed Where
Rawls Failed?” Contemporary Political Theory 8 (2): 199–223.
Caney, Simon. 2006. Justice Beyond Borders. Oxford: Oxford University Press.
Cassam, Quassim. 1987. “Transcendental Arguments, Transcendental Synthesis and
Transcendental Idealism.” Philosophical Quarterly 37 (149): 355–78.
———. 1996. “Self-Reference, Self-Knowledge and the Problem of Misconception.”
European Journal of Philosophy 4 (3): 276–95.
———. 2003. “Self-Directed Transcendental Arguments.” In Transcendental Arguments:
Problems and Prospects, edited by Robert Stern. Oxford: Clarendon Press.
———. 2014. Self-Knowledge for Humans. Oxford: Oxford University Press.
Chitty, Andrew. 2008. “Protagonist and Subject in Gewirth’s Argument for Human Rights.”
King’s Law Journal 19 (1): 1–26.
Claassen, Rutger. forthcoming. “An Agency-Based Capability Theory of Justice.” European
Journal of Philosophy.
Claassen, Rutger, and Marcus Düwell. 2013. “The Foundations of Capability Theory:
Comparing Nussbaum and Gewirth.” Ethical Theory and Moral Practice 16 (3):
493–510.
Cline, Brendan. 2016. “Against Deliberative Indispensability as an Independent Guide to
What There Is.” Philosophical Studies, 1–20.
Cohen, G.A. 2003. “Facts and Principles.” Philosophy & Public Affairs 31 (3): 211–45.
———. 2008. Rescuing Justice and Equality. Cambridge, MA: Harvard University Press.
Coleman, Mary Clayton. 2005. “Public Reasons and Practical Solipsism.” Southern Journal
of Philosophy 43 (3): 317–36.
Copp, David. 2013. “Is Constructivism an Alternative to Moral Realism?” In Constructivism
in Ethics, edited by Carla Bagnoli, 108–32. Cambridge: Cambridge University
Press.
SAMENVATTING
330
Dancy, Jonathan. 2004. Ethics Without Principles. Oxford: Oxford University Press.
———. 2013. “Moral Particularism.” In The Stanford Encyclopedia of Philosophy, edited by
Edward
N.
Zalta.
http://plato.stanford.edu/archives/fall2013/entries/moral-
particularism/.
Daniels, Norman. 1979a. “Moral Theory and the Plasticity of Persons.” The Monist 62 (3):
265–87.
———. 1979b. “Wide Reflective Equilibrium and Theory Acceptance in Ethics.” Journal of
Philosophy 76 (5): 256–282.
———. 1980. “Reflective Equilibrium and Archimedean Points.” Canadian Journal of
Philosophy 10 (1): 83–103.
Danto, Arthur C. 1984. “Comment on Gewirth Constructing an Epistemology of Human
Rights: A Pseudo Problem?” Social Philosophy and Policy 1 (2): 25–30.
Darwall, Stephen. 2006. The Second person Standpoint: Morality, Respect, and
Accountability. Cambridge, MA: Harvard University Press.
———. 2007. “Reply to Korsgaard, Wallace, and Watson.” Ethics 118 (1): 52–69.
———. 2010. “Precis: The Second person Standpoint.” Philosophy and Phenomenological
Research 81 (1): 216–28.
DePaul, Michael R. 1986. “Reflective Equilibrium and Foundationalism.” American
Philosophical Quarterly 23 (1): 59–69.
———. 2007. “Intuitions in Moral Inquiry.” In The Oxford Handbook of Ethical Theory,
edited by David Copp. Oxford: Oxford University Press.
Dorsey, Dale. 2016. “Moral Distinctiveness and Moral Inquiry.” Ethics 126 (3): 747–73.
Düring, Dascha, and Marcus Düwell. 2015. “Towards a Kantian Theory of Judgment: The
Power of Judgment in Its Practical and Aesthetic Employment.” Ethical Theory
and Moral Practice 18 (5): 943–56.
Düwell, Marcus. 2013. Bioethics: Methods, Theories, Domains. Abingdon: Routledge.
Dworkin, Ronald. 1996. “Objectivity and Truth: You’d Better Believe It.” Philosophy and
Public Affairs 25 (2): 87–139.
———. 2011. Justice for Hedgehogs. Cambridge, MA: Belknap Press of Harvard University
Press.
Ebertz, Roger P. 1993. “Is Reflective Equilibrium a Coherentist Model?” Canadian Journal
of Philosophy 23 (2): 193–214.
Enoch, David. 2006. “Agency, Shmagency: Why Normativity Won’t Come From What Is
Constitutive of Action.” Philosophical Review 115 (2): 169–198.
SAMENVATTING
331
———. 2010. “How Objectivity Matters.” In Oxford Studies in Metaethics, edited by Russ
Shafer-Landau, 5:111–52. Oxford: Oxford University Press.
———. 2011a. “Shmagency Revisited.” In New Waves in Metaethics. London: Palgrave
Macmillan.
———. 2011b. Taking Morality Seriously: A Defense of Robust Realism. Oxford: Oxford
University Press.
Estlund, David. 2011. “Human Nature and the Limits (If Any) of Political Philosophy.”
Philosophy & Public Affairs 39 (3): 207–37.
Ferrero, Luca. 2009. “Constitutivism and the Inescapability of Agency.” In Oxford Studies in
Metaethics, edited by Russ Shafer-Landau, 4:303–333. Oxford: Oxford University
Press.
Flikschuh, Katrin. forthcoming. “The Arc of Personhood: Menkiti and Kant on Becoming
and Being a Person.” Journal of the American Philosophical Association, 1–38.
———. 2000. Kant and Modern Political Philosophy. Cambridge: Cambridge University
Press.
———. 2014. “The Idea of Philosophical Fieldwork: Global Justice, Moral Ignorance, and
Intellectual Attitudes.” Journal of Political Philosophy 22 (1): 1–26.
Floyd, Jonathan. forthcoming. “Rawls’ Methodological Blueprint.” European Journal of
Political Theory, 1–15.
Formosa, Paul. 2013. “The Role of Vulnerability in Kantian Ethics.” In Vulnerability: New
Essays in Ethics and Feminist Philosophy, edited by Catriona Mackenzie, Wendy
Rogers, and Susan Dodds, 88–109. New York: Oxford University Press.
Freyenhagen, Fabian. 2012. “The Empty Formalism Objection Revisited: Section 135R and
Recent Kantian Responses.” In Hegel’s Philosophy of Right, edited by Thom
Brooks, 43–72. Oxford: Blackwell Publishing Ltd.
Gert,
Joshua.
2002.
“Korsgaard’s
Private-Reasons
Argument.”
Philosophy
and
Phenomenological Research 64 (2): 303–24.
Gewirth, Alan. 1960. “Meta-Ethics and Normative Ethics.” Mind 69 (274): 187–205.
———. 1978. Reason and Morality. Chicago: University of Chicago Press.
———. 1984. “Reply to Danto.” Social Philosophy and Policy 1 (2): 31–34.
———. 1985. “Why There Are Human Rights.” Social Theory and Practice 11 (2): 235–48.
———. 1988. “The Justification of Morality.” Philosophical Studies 53 (2): 245–262.
———. 1992. “Human Dignity as the Basis of Rights.” In The Constitution of Rights :
Human Dignity and American Values, edited by Michael J Meyer and William
SAMENVATTING
332
Parent, 10–28. Ithaca: Cornell University Press.
———. 1993. “The Constitutive Metaphysics of Ethics.” Revue de Métaphysique et de
Morale 98 (4): 489–504.
———. 1996. The Community of Rights. Chicago: The University of Chicago Press.
———. 1997. “‘Ought’ and Reasons for Action.” The Southern Journal of Philosophy 35 (2):
171–77.
———. 1998. Self-Fulfillment. Princeton: Princeton University Press.
———. 1999. “Reply to My Colleagues.” In Gewirth. Critical Essays on Action, Rationality,
and Community, edited by Michael Boylan, 191–213. Lanham: Rowman &
Littlefield.
Gilabert, Pablo. 2005a. “A Substantivist Construal of Discourse Ethics.” International
Journal of Philosophical Studies 13 (3): 405 – 437.
———. 2005b. “The Substantive Dimension of Deliberative Practical Rationality.”
Philosophy and Social Criticism 31 (2): 185–210.
———. 2006. “Considerations on the Notion of Moral Validity in the Moral Theories of
Kant and Habermas.” Kant-Studien 97 (2): 210–27.
Goodman, Nelson. 1955. Fact, Fiction and Forecast. Cambridge, MA: Harvard University
Press.
Gould, Carol C. 2004. Globalizing Democracy and Human Rights. Cambridge: Cambridge
University Press.
———. 2015. “A Social Ontology of Human Rights.” In Philosophical Foundations of
Human Rights, edited by Rowan Cruft, S. Matthew Liao, and Massimo Renzo,
177–95. Oxford: Oxford University Press.
Guyer, Paul. 2013. “Constructivism and Self-Constitution.” In Kant on Practical
Justification: Interpretive Essays, edited by Mark Timmons and Sorin Baiasu, 176–
200. Oxford: Oxford University Press.
Habermas, Jürgen. 1990a. “Discourse Ethics: Notes on a Program of Philosophical
Justification.” In Moral Consciousness and Communicative Action, 43–115.
Cambridge, MA: MIT Press.
———. 1990b. Moral Consciousness and Communicative Action. Cambridge, MA: MIT
Press.
———. 1994. Justification and Application: Remarks on Discourse Ethics. Cambridge, MA:
MIT Press.
SAMENVATTING
333
Hare, R. M. 1973. “Rawls’ Theory of Justice—I.” The Philosophical Quarterly 23 (91): 144–
55.
Harman, Gilbert. 1975. “Moral Relativism Defended.” The Philosophical Review 84 (1): 3.
———. 1977. The Nature of Morality: An Introduction to Ethics. Oxford: Oxford
University Press.
———. 1983. “Human Flourishing, Ethics, and Liberty.” Philosophy and Public Affairs 12
(4): 307–22.
———. 2015. “Moral Relativism Is Moral Realism.” Philosophical Studies 172 (4): 855–63.
Harman, Gilbert, and Judith Thomson. 1996. Moral Relativism and Moral Objectivity.
Cambridge, MA: Wiley-Blackwell.
Heath, Joseph. 2003. Communicative Action and Rational Choice. Cambridge, MA: MIT
Press.
Hegel, Georg Wilhelm Fredrich. 1991. Hegel: Elements of the Philosophy of Right. Edited by
Allen W. Wood. Translated by H. B. Nisbet. Cambridge: Cambridge University
Press.
———. 1999. Hegel: Political Writings. Edited by Lawrence Dickey and H. B. Nisbet.
Translated by H. B. Nisbet. Cambridge: Cambridge University Press.
Herman, Barbara. 1984. “Mutual Aid and Respect for Persons.” Ethics 94 (4): 577–602.
———. 1985. “The Practice of Moral Judgment.” The Journal of Philosophy 82 (8): 414.
———. 1996. “Leaving Deontology Behind.” In The Practice of Moral Judgment, 208–42.
Cambridge, MA: Harvard University Press.
———. 2007. Moral Literacy. Cambridge, MA: Harvard University Press.
Hocutt, Max. 2000. Grounded Ethics the Empirical Bases of Normative Judgements. New
Brunswick: Transaction Publishers.
Hoffe, Otfried. 2013. “Anthropology and Metaphysics in Kant’s Categorical Imperative of
Law.” In Kant on Practical Justification: Interpretive Essays, edited by Mark
Timmons and Sorin Baiasu, 110–24. Oxford: Oxford University Press.
Holmgren, Margaret. 1987. “Wide Reflective Equilibrium and Objective Moral Truth.”
Metaphilosophy 18 (2): 108–24.
Illies, Christian. 2003. The Grounds of Ethical Judgement: New Transcendental Arguments
in Moral Philosophy. Oxford: Oxford University Press.
Iwasa, Noriaki. 2013. “Reason Alone Cannot Identify Moral Laws.” The Journal of Value
Inquiry 47 (1–2): 67–85.
Jones, Karen. 2005. “Moral Epistemology.” In The Oxford Handbook of Contemporary
SAMENVATTING
334
Philosophy, edited by Frank Jackson and Michael Smith, 63–85. Oxford: Oxford
University Press.
Jongepier, Fleur. 2016. “Towards a Constitutive Account of Implicit Narrativity.”
Phenomenology and the Cognitive Sciences 15 (1): 51–66.
Joyce, Richard. 2015. “Moral Anti-Realism.” In The Stanford Encyclopedia of Philosophy,
edited by Edward N. Zalta.
http://plato.stanford.edu/archives/fall2015/entries/moral-anti-realism/.
Kant, Immanuel. 1996. The Metaphysics of Morals. Edited and translated by Mary J. Gregor.
Cambridge: Cambridge University Press.
———. 1998. Groundwork of the Metaphysics of Morals. Translated by Mary J. Gregor.
Cambridge: Cambridge University Press.
Katsafanas, Paul. 2013. Agency and the Foundations of Ethics: Nietzschean Constitutivism.
Oxford: Oxford University Press.
Kelly, Thomas, and Sarah McGrath. 2010. “Is Reflective Equilibrium Enough?”
Philosophical Perspectives 24 (1): 325–359.
Kerstein, Samuel J. 2001. “Korsgaard’s Kantian Arguments for the Value of Humanity.”
Canadian Journal of Philosophy 31 (1): 23–52.
Kolodny, Niko, and John Brunero. 2016. “Instrumental Rationality.” In The Stanford
Encyclopedia of Philosophy, edited by Edward N. Zalta.
http://plato.stanford.edu/archives/spr2016/entries/rationality-instrumental/.
Korsgaard, Christine M. forthcoming. “Valuing Our Humanity.” Downloaded from
http://www.people.fas.harvard.edu/~korsgaar/CMK.Valuing.Our.Humanity.pdf
July 17 2016.
———. 1985. “Kant’s Formula of Universal Law.” Pacific Philosophical Quarterly 66 (1–2):
24–47.
———. 1986. “Kant’s Formula of Humanity.” Kant-Studien 77 (2): 183–202.
———. 1989. “Personal Identity and the Unity of Agency: A Kantian Response to Parfit.”
Philosophy & Public Affairs 18 (2): 101–32.
———. 1996. The Sources of Normativity. Cambridge: Cambridge University Press.
———. 1997. “The Normativity of Instrumental Reason.” In Ethics and Practical Reason,
edited by Garrett Cullity and Berys Gaut. Oxford: Clarendon Press.
———. 2007. “Autonomy and the Second Person Within: A Commentary on Stephen
Darwall’s The Second person Standpoint.” Ethics 118 (1): 8–23.
———. 2008a. “Realism and Constructivism in Twentieth-Century Moral Philosophy.” In
SAMENVATTING
335
The Constitution of Agency, 302–26. Oxford: Oxford University Press.
———. 2008b. The Constitution of Agency. Oxford: Oxford University Press.
———. 2009. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University
Press.
Kramer, Matthew H., and Nigel E. Simmonds. 1996. “Reason Without Reasons: A Critique
of Alan Gewirth’s Moral Philosophy.” Southern Journal of Philosophy 34 (3): 301–
315.
———. 1998. “No Better Reasons: A Reply to Alan Gewirth.” Southern Journal of
Philosophy 36 (1): 131–139.
Laden, Anthony Simon. 2003. “The House That Jack Built: Thirty Years of Reading Rawls.”
Ethics 113 (2): 367–90.
Langton, Rae. 2007. “Objective and Unconditioned Value.” Philosophical Review 116 (2):
157–85.
Lazari-Radek, Katarzyna de, and Peter Singer. 2014. The Point of View of the Universe:
Sidgwick and Contemporary Ethics. Oxford: Oxford University Press.
Lear, Jonathan. 1998. “Transcendental Anthropology.” In Open Minded: Working Out the
Logic of the Soul, 247–81. Cambridge, MA: Harvard University Press.
LeBar, Mark. 2001. “Korsgaard, Wittgenstein, and the Mafioso.” Southern Journal of
Philosophy 39 (2): 261–71.
Lenman, James. 2013. “Deliberation, Schmeliberation: Enoch’s Indispensability Argument.”
Philosophical Studies 168 (3): 835–42.
Louden, Robert B. 2000. Kant’s Impure Ethics: From Rational Beings to Human Beings.
Oxford: Oxford University Press.
Maagt, Sem de. 2014. “In Defence of Fact-Dependency.” Canadian Journal of Philosophy 44
(3–4): 443–62.
———. 2016. “Reflective Equilibrium and Moral Objectivity.” Inquiry, 1–27.
MacIntyre, Alasdair. 1985. After Virtue. Vol. 2nd. London: Duckworth.
McDowell, John. 1998. “Deliberation and Moral Development.” In Aristotle, Kant, and the
Stoics: Rethinking Happiness and Duty, edited by Stephen Engstrom and Jennifer
Whiting, 19–35. Cambridge: Cambridge University Press.
McMahan, Jeff. 2000. “Moral Intuition.” In The Blackwell Guide to Ethical Theory, edited by
LaFollette Hugh, 92–110. Oxford: Blackwell Publishers.
McMahon, Christopher. 1986. “Gewirth’s Justification of Morality.” Philosophical Studies
50 (2): 261–281.
SAMENVATTING
336
McPherson, Tristram. 2012. “Unifying Moral Methodology.” Pacific Philosophical
Quarterly 93 (4): 523–49.
Miller, David. 2012. Justice for Earthlings: Essays in Political Philosophy. Cambridge:
Cambridge University Press.
Nagel, Thomas. 1979. The Possibility of Altruism. Princeton: Princeton University Press.
Nance, Michael. 2015. “Recognition, Freedom, and the Self in Fichte’s Foundations of
Natural Right.” European Journal of Philosophy 23 (3): 608–32.
Narveson, Jan. 1980. “Gewirth’s Reason and Morality – A Study in the Hazards of
Universalizability in Ethics.” Dialogue 19 (4): 651–74.
Nyholm, Sven. 2016. “On Kant’s Idea of Humanity as an End in Itself.” European Journal of
Philosophy 24 (2): 358–74.
O’Neill, Onora. 1986. “The Public Use of Reason.” Political Theory 14 (4): 523–51.
———. 1989a. “Consistency in Action.” In Constructions of Reason: Explorations of Kant’s
Practical Philosophy, 81–104. Cambridge: Cambridge University Press.
———. 1989b. Constructions of Reason: Explorations of Kant’s Practical Philosophy.
Cambridge: Cambridge University Press.
———. 1989c. “Reason and Politics in the Kantian Enterprise.” In Constructions of Reason:
Explorations of Kant’s Practical Philosophy, 3–27. Cambridge: Cambridge
University Press.
———. 1992. “Vindicating Reason.” In The Cambridge Companion to Kant, edited by Paul
Guyer, 280–308. Cambridge University Press.
———. 1996. Towards Justice and Virtue: A Constructive Account of Practical Reasoning.
Cambridge: Cambridge University Press.
———. 2000a. “Distant Strangers, Moral Standing and Porous Boundaries.” In Bounds of
Justice, 186–202. Cambridge: Cambridge University Press.
———. 2000b. “Four Models of Practical Reasoning.” In Bounds of Justice, 11–28.
Cambridge: Cambridge University Press.
———. 2000c. “Principles, Practical Judgement and Institutions.” In Bounds of Justice, 50–
64. Cambridge: Cambridge University Press.
———. 2000d. Bounds of Justice. Cambridge: Cambridge University Press.
———. 2001. “Practical Principles & Practical Judgment.” Hastings Center Report 31 (4):
15–23.
———. 2002. “Constructivism in Rawls and Kant.” In The Cambridge Companion to Rawls,
edited by Samuel Freeman, 347–67. Cambridge: Cambridge University Press.
SAMENVATTING
337
———. 2007a. “Experts, Practitioners, and Practical Judgement.” Journal of Moral
Philosophy 4 (2): 154–66.
———. 2007b. “Normativity and Practical Judgement.” Journal of Moral Philosophy 4 (3):
393–405.
———. 2011. “Orientation in Thinking: Geographical Problems, Political Solutions.” In
Reading Kant’s Geography, edited by Stuart Elden and Eduardo Mendieta, 540–87.
New York: SUNY Press.
———. 2013. “Responses.” In Reading Onora O’Neill, edited by David Archard, Monique
Deveaux, Neil Manson, and Daniel Weinstock, 219–43. Abingdon: Routledge.
Parfit, Derek. 1973. “Later Selves and Moral Principles.” In Philosophy and Personal
Relations, edited by A. Montefiore, 137–69. Abingdon: Routledge.
———. 1984. Reasons and Persons. Oxford: Oxford University Press.
Rawls, John. 1974. “The Independence of Moral Theory.” Proceedings and Addresses of the
American Philosophical Association 48: 5–22.
———. 1980. “Kantian Constructivism in Moral Theory.” Journal of Philosophy 77 (9):
515–72.
———. 1985. “Justice as Fairness: Political Not Metaphysical.” Philosophy & Public Affairs
14 (3): 223–51.
———. 1995. “Political Liberalism: Reply to Habermas.” Journal of Philosophy 92 (3): 132–
80.
———. 1999a. “Themes in Kant’s Moral Philosophy.” In Collected Papers, edited by Samuel
Freeman, 497–528. Cambridge, MA: Harvard University Press.
———. 1999b. A Theory of Justice: Revised Edition. Cambridge, MA: Belknap Press.
———. 2000. The Law of Peoples: With “the Idea of Public Reason Revisited.” Cambridge,
MA: Harvard University Press.
———. 2003. Justice as Fairness: A Restatement. Cambridge, MA: Harvard University Press.
———. 2005. Political Liberalism: Expanded Edition. New York: Columbia University
Press.
Regan, Donald H. 1999. “Gewirth on Necessary Goods: What Is the Agent Committed to
Valuing?” In Gewirth. Critical Essays on Action, Rationality, and Community,
edited by Michael Boylan, 45–70. Lanham: Rowman & Littlefield.
———. 2002. “The Value of Rational Nature.” Ethics 112 (2): 267–91.
Regis, Edward. 1984. Gewirth’s Ethical Rationalism: Critical Essays with a Reply by Alan
Gewirth. Vol. 96. Chicago: University of Chicago Press.
SAMENVATTING
338
Richardson, Henry S. 1999. “The Problem with Simply Promoting Agency: Comment on
Cummiskey.” In Gewirth: Critical Essays on Action, Rationality, and Community,
edited by Michael Boylan, 141–44. Lanham: Rowman & Littlefield.
Robeyns, Ingrid. 2011. “The Capability Approach.” In The Stanford Encyclopedia of
Philosophy, edited by Edward N. Zalta.
http://plato.stanford.edu/archives/sum2011/entries/capability-approach/.
Ronzoni, Miriam, and Laura Valentini. 2008. “On the Meta-Ethical Status of
Constructivism: Reflections on G.A. Cohen’s ‘Facts and Principles.’” Politics,
Philosophy & Economics 7 (4): 403–22.
Rorty, Richard. 1989. Contingency, Irony, and Solidarity. Cambridge: Cambridge University
Press.
Ross, David. 2002. The Right and the Good. Edited by Philip Stratton-Lake. Oxford: Oxford
University Press.
Rovane, Carol. 2008. “Did Williams Find the Truth in Relativism?” In Reading Bernard
Williams, edited by Daniel Callcut, 43–69. Abingdon: Routledge.
———. 2013. The Metaphysics and Ethics of Relativism. Cambridge, MA: Harvard
University Press.
Sayre-McCord, Geoff. 2014. “Metaethics.” In The Stanford Encyclopedia of Philosophy,
edited by Edward N. Zalta.
http://plato.stanford.edu/archives/sum2014/entries/metaethics/.
Scanlon, T. M. 2003. “Rawls on Justification.” In The Cambridge Companion to Rawls,
edited by Samuel Freeman, 139–67. Cambridge: Cambridge University Press.
———. 2012. “The Appeal and Limits of Constructivism.” In Constructivism in Practical
Philosophy, edited by Jimmy Lenman and Yonatan Shemmer. Oxford: Oxford
University Press.
———. 2014. Being Realistic about Reasons. Oxford: Oxford University Press.
Schafer, Karl. 2015a. “Realism and Constructivism in Kantian Metaethics (1): Realism and
Constructivism in a Kantian Context.” Philosophy Compass 10 (10): 690–701.
———. 2015b. “Realism and Constructivism in Kantian Metaethics (2): The Kantian
Conception of Rationality and Rationalist Constructivism.” Philosophy Compass
10 (10): 702–13.
Scheffler, Samuel. 1982. “Ethics, Personal Identity, and Ideals of the Person.” Canadian
Journal of Philosophy 12 (2): 229–46.
Scheuermann, James. 1987. “Gewirth’s Concept of Prudential Rights.” The Philosophical
SAMENVATTING
339
Quarterly 37 (148): 291–304.
Schuppert, Fabian. 2013. “Distinguishing Basic Needs and Fundamental Interests.” Critical
Review of International Social and Political Philosophy 16 (1): 24–44.
Sedgwick, Sally. 2011. “‘Letting the Phenomena In’: On How Herman’s Kantianism Does
and Does Not Answer the Empty Formalism Critique.” Kantian Review 16 (01):
33–47.
Shafer-Landau, Russ. 2003. Moral Realism: A Defence. Oxford: Oxford University Press.
Sidgwick, Henry. 1981. The Methods of Ethics. 7th edition. Indianapolis: Hackett Publishing
Company.
Silverstein, Matthew. 2012. “Inescapability and Normativity.” Journal of Ethics and Social
Philosophy 6 (3): 1–27.
———. 2015. “The Shmagency Question.” Philosophical Studies 172 (5): 1127–42.
Singer, Peter. 1974. “Sidgwick and Reflective Equilibrium.” Monist 58 (3): 490–517.
———. 1993. Practical Ethics. Cambridge: Cambridge University Press.
———. 2005. “Ethics and Intuitions.” Journal of Ethics 9 (3–4): 331–52.
Skidmore, James. 2002. “Skepticism about Practical Reason: Transcendental Arguments
and Their Limits.” Philosophical Studies 109 (2): 121–41.
Smit, Houston, and Mark Timmons. 2013. “Kant’s Grounding Project in The Doctrine of
Virtue.” In Kant on Practical Justification: Interpretive Essays, edited by Mark
Timmons and Sorin Baiasu, 229–68. Oxford: Oxford University Press.
Steigleder, Klaus. 1999. Grundlegung der normativen Ethik: Der Ansatz von Alan Gewirth.
Freiburg: Verlag Karl Alber.
Stern, Robert. 1999. Transcendental Arguments: Problems and Prospects. Oxford: Oxford
University Press.
———. 2000. Transcendental Arguments and Scepticism: Answering the Question of
Justification. Oxford: Oxford University Press.
———. 2011. “The Value of Humanity: Reflections on Korsgaard’s Transcendental
Argument.” In Transcendental Philosophy and Naturalism, edited by J. Smith and
P. Sullivan, 74–95. Oxford: Oxford University Press.
———. 2012a. “Constructivism and the Argument from Autonomy.” In Constructivism in
Practical Philosophy, edited by Jimmy Lenman and Yonatan Shemmer, 119–37.
Oxford: Oxford University Press.
———. 2012b. “On Hegel’s Critique of Kant’s Ethics: Beyond the Empty Formalism
Objection.” In Hegel’s Philosophy of Right, edited by Thom Brooks, 73–100.
SAMENVATTING
340
Oxford: Blackwell Publishing Ltd.
———. 2013a. “Transcendental Arguments.” In The Stanford Encyclopedia of Philosophy,
edited by Edward N. Zalta.
http://plato.stanford.edu/archives/sum2013/entriesranscendental-arguments/.
———. 2013b. “Moral Skepticism, Constructivism, and the Value of Humanity.” In
Constructivism in Ethics, edited by Carla Bagnoli, 22–40. Cambridge: Cambridge
University Press.
———. 2015. “Round Kant or Through Him? On James’s Arguments for Freedom, and
Their Relation to Kant’s.” In Pragmatism, Kant, and Transcendental Philosophy,
edited by Gabriele Gava and Robert Stern, 152–76. Abingdon: Routledge.
Strawson, Galen. 2004. “Against Narrativity.” Ratio 17 (4): 428–52.
Strawson, Peter. 1975. The Bounds of Sense: An Essay on Kant’s Critique of Pure Reason.
Abingdon: Routledge.
Street, Sharon. 2006. “A Darwinian Dilemma for Realist Theories of Value.” Philosophical
Studies 127 (1): 109–66.
———. 2008. “Constructivism about Reasons.” In Oxford Studies in Metaethics, edited by
Russ Shafer-Landau, 3:207–45. Oxford: Oxford University Press.
———. 2009. “In Defense of Future Tuesday Indifference: Ideally Coherent Eccentrics and
the Contingency of What Matters.” Philosophical Issues 19 (1): 273–98.
———. 2010. “What Is Constructivism in Ethics and Metaethics?” Philosophy Compass 5
(5): 363–84.
———. 2012. “Coming to Terms with Contingency : Humean Constructivism about
Practical Reason.” In Constructivism in Practical Philosophy, edited by Jimmy
Lenman and Yonatan Shemmer, 40–59. Oxford: Oxford University Press.
———. 2016. “I—Constructivism in Ethics and the Problem of Attachment and Loss.”
Aristotelian Society Supplementary Volume 90 (1): 161–89.
Stroud, Barry. 1968. “Transcendental Arguments.” Journal of Philosophy 65 (9): 241–256.
———. 1994. “Kantian Argument, Conceptual Capacities, and Invulnerability.” In Kant
and Contemporary Epistemology, edited by Paolo Parrini, 231–51. 54. Dordrecht:
Springer Netherlands.
Sussman, David. 2003. “The Authority of Humanity.” Ethics 113 (2): 350–66.
Taylor, Charles. 1978. “The Validity of Transcendental Arguments.” Proceedings of the
Aristotelian Society 79: 151–65.
———. 1985. “What Is Human Agency?” In Philosophical Papers: Volume 1, Human
SAMENVATTING
341
Agency and Language, 15–44. Cambridge: Cambridge University Press.
Tersman, Folke. 2008. “The Reliability of Moral Intuitions: A Challenge from
Neuroscience.” Australasian Journal of Philosophy 86 (3): 389–405.
Tiffany, Evan. 2012. “Why Be an Agent?” Australasian Journal of Philosophy 90 (2): 223–33.
Timmons, Mark. 1987. “Foundationalism and the Structure of Ethical Justification.” Ethics
97 (3): 595–609.
Tubert, Ariela. 2010. “Constitutive Arguments.” Philosophy Compass 5 (8): 656–666.
Vallentyne, Peter, Hillel Steiner, and Michael Otsuka. 2005. “Why Left-Libertarianism Is
Not Incoherent, Indeterminate, or Irrelevant: A Reply to Fried.” Philosophy &
Public Affairs 33 (2): 201–215.
Velleman, J. David. 2013. Foundations for Moral Relativism. Open Book Publishers.
http://www.openbookpublishers.com/product/181.
Walden, Kenneth. 2013. “In Defense of Reflective Equilibrium.” Philosophical Studies 166
(2): 243–56.
Walker, Rebecca L., and P. J. Ivanhoe. 2007. Working Virtue: Virtue Ethics and
Contemporary Moral Problems. Vol. 69. Oxford: Oxford University Press.
Wallace, R. Jay. 2009. “The Publicity of Reasons.” Philosophical Perspectives 23 (1): 471–97.
Williams, Bernard. 1973. “A Critique of Utilitarianism.” In Utilitarianism: For and against,
edited by John Jamieson Carswell Smart and Bernard Williams, 77–150.
Cambridge: Cambridge University Press.
———. 1979. “Internal and External Reasons.” In Rational Action, edited by Ross Harrison,
101–13. Cambridge: Cambridge University Press.
———. 1986. Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University
Press.
———. 1995. “Replies.” In World, Mind, and Ethics: Essays on the Ethical Philosophy of
Bernard Williams, edited by J.E.J. Altham and Ross Harrison, 185–224.
Cambridge: Cambridge University Press.
Wittgenstein, Ludwig. 2009. Philosophical Investigations. Edited by P. M. S. Hacker and
Joachim Schulte. Translated by G. E. M. Anscombe. Hoboken: Wiley-Blackwell.
Wood, Allen W. 1990. Hegel’s Ethical Thought. Cambridge: Cambridge University Press.
Wittgenstein, Ludwig. 2009. Philosophical Investigations. Edited by P. M. S. Hacker and
Joachim Schulte. Translated by G. E. M. Anscombe. Hoboken: Wiley-Blackwell.
Wood, Allen W. 1990. Hegel’s Ethical Thought. Cambridge: Cambridge University Press.
Samenvatting
Is een uitspraak als ‘we hebben de morele plicht vluchtelingen op te vangen’ of ‘we hebben
de morele plicht om iets te doen aan sociaal-economische ongelijkheid’ correct of
incorrect? En waar is de correctheid van een dergelijke morele uitspraak – een uitspraak
over hoe we moeten handelen – van afhankelijk? Is dit een slechts een mening, iets dat
afhankelijk is van een specifieke (sub)cultuur, of bestaan er wellicht objectieve morele
uitspraken, dat wil zeggen uitspraken die iedereen zou moeten accepteren? En hoe zouden
we de correctheid van een morele uitspraak kunnen bepalen? Dit soort vragen liggen ten
grondslag aan dit proefschrift over de fundering van moraal.
In dit proefschrift probeer ik niet zozeer een specifiek antwoord te geven op concrete
morele vragen (zoals vragen over vluchtelingen of ongelijkheid), maar onderzoek ik de
vraag welke standaard we zouden kunnen gebruiken om te bepalen of morele uitspraken
correct of incorrect zijn, en of er een ‘objectieve standaard’ bestaat om dit de beoordelen. Ik
verdedig een ‘Kantiaans constructivistische’ theorie van morele rechtvaardiging en morele
objectiviteit. Ik definieer Kantiaans constructivisme als de positie die (1) een
rechtvaardiging probeert te geven voor objectieve morele principes (2) zonder zich te
beroepen op het bestaan van morele feiten (3) door middel van een transcendentaal
argument.
In het inleidende hoofdstuk 1 werk ik de definitie van het Kantiaans constructivisme uit.
Volgens het Kantiaans constructivisme bestaan er geen morele feiten in de wereld, zoals er
bijvoorbeeld empirische feiten bestaan over stoelen en tafels, maar ligt de objectiviteit van
moraal besloten in hoe we onszelf begrijpen, een zelfbegrip dat we, ondanks soms grote
onderlinge verschillen, met elkaar delen. Dit zelfbegrip is, bijvoorbeeld, het begrip van
onszelf als een handelend wezen, een wezen dat doelen nastreeft in de wereld, of als een
wezen dat deelneemt aan bepaalde fundamentele vormen van interactie zoals
communicatie. Het idee van Kantiaans constructivisme is dat een bepaalde vorm van
SAMENVATTING
344
zelfbegrip onvermijdelijk is en dat bepaalde fundamentele morele principes onderdeel zijn
van dit zelfbegrip. Binnen het Kantiaans constructivisme wordt morele objectiviteit dus niet
begrepen in termen van het bestaan van morele feiten, maar in termen van normatieve
overtuigingen die iedereen moet accepteren onafhankelijk van zijn of haar specifieke
voorkeuren of overtuigingen (‘categoriciteit’) omdat deze onlosmakelijk verbonden zijn
met een bepaald type zelfbegrip.
De Kantiaans constructivist argumenteert vanuit dit onvermijdelijke zelfbegrip naar
morele principes door middel van een transcendentaal argument. Het is lastig om een
beknopte definitie te geven van een transcendentaal argument, deze definitie is zelf
onderwerp van dit proefschrift. Grofweg kan een transcendentaal argument begrepen
worden als een argument dat probeert te laten zien dat iedereen die zichzelf op een
bepaalde manier begrijpt (bijvoorbeeld als een handelend wezen), gecommitteerd is aan de
noodzakelijke mogelijkheidsvoorwaarde(n) van dit zelfbegrip. Een bekend voorbeeld in de
theoretische filosofie is Descartes’ ‘cogito-argument’. Het uitgangspunt van Descartes is dat
we onszelf begrijpen als denkende wezens. Vervolgens stelt hij dat bestaan een
noodzakelijke mogelijkheidsvoorwaarde is voor denken (dat wil zeggen: dat je alleen kunt
denken als je bestaat). Descartes concludeert hieruit dat je moet accepteren dat je bestaat.
Transcendentale argumenten in de ethiek proberen te laten zien dat er bepaalde morele
principes onderdeel zijn van de noodzakelijke mogelijkheidsvoorwaarden van een
onvermijdelijk onderdeel van ons zelfbegrip. Het startpunt van een transcendentaal
argument is dus het zelfbegrip van een persoon, de conclusie van een transcendentaal
argument in de ethiek is een moreel principe.
Transcendentale argumenten zijn niet oncontroversieel, en het succes van het
Kantiaans constructivisme is met name afhankelijk van de plausibiliteit van de
transcendentale methodologie. De focus van het proefschrift ligt dan ook op
transcendentale argumentatie.
Het proefschrift is opgebouwd als een argument door eliminatie. In hoofdstuk 2 bekritiseer
ik de meest dominante methode voor morele rechtvaardiging, de methode van het
‘reflectief equilibrium’ die geïntroduceerd is in de ethiek door John Rawls. Ik
beargumenteer dat de methode van het reflectief equilibrium niet kan leiden tot de
rechtvaardiging van objectieve morele uitspraken en dat voorstanders hiervan te snel
concluderen dat het de enige mogelijke methode van morele rechtvaardiging is. Deze
kritiek op de methode van het reflectief equilibrium motiveert de bespreking van een
SAMENVATTING
345
alternatieve methode van morele rechtvaardiging: transcendentale argumentatie. In het
vervolg van het proefschrift evalueer ik in hoeverre bestaande transcendentale argumenten
succesvoller zijn dan de methode van het reflectief equilibrium.
In hoofdstuk 3 introduceer ik allereerst transcendentale argumentatie als een algemene
argumentatiestrategie die zowel in de ethiek als in de theoretische filosofie wordt gebruikt.
Daarnaast beargumenteer ik in dit hoofdstuk dat er belangrijke verschillen zijn tussen
transcendentale argumenten in de theoretische filosofie en transcendentale argumenten in
de ethiek. Het belangrijkste verschil is dat transcendentale argumenten in de theoretische
filosofie (zoals Descartes’ cogito-argument) iets willen zeggen over hoe de wereld is, terwijl
transcendentale argumenten in de ethiek iets proberen te zeggen over hoe de wereld zou
moeten zijn. Dit onderscheid is belangrijk omdat er veel scepticisme over transcendentale
argumentatie bestaat die terug te leiden is tot scepticisme over een aantal specifieke
transcendentale argumenten in de theoretische filosofie die sterke, ontologische claims
maken over de wereld. Door het verschil tussen transcendentale argumenten in de
theoretische filosofie en de ethiek te benadrukken, hoop ik één mogelijke bron van
scepticisme over het Kantiaans constructivistische project weg te nemen.
In de hoofdstukken die daarop volgen, evalueer ik bestaande transcendentale
argumenten in de ethiek aan de hand van wat ik het ‘potentiele dilemma’ voor het
Kantiaans constructivisme noem. Dit dilemma luidt als volgt: oftewel er bestaat inderdaad
een onvermijdelijke vorm van zelfbegrip die we als startpunt kunnen nemen van een
transcendentaal argument, maar in dat geval is het onwaarschijnlijk dat er interessante
morele conclusies uit volgen. Oftewel vallen met het Kantiaans constructivisme interessante
morele conclusies te verdedigen, maar in dat geval is het zelfbegrip dat het startpunt vormt
van het argument slechts optioneel en dus niet onvermijdelijk. In het laatste geval zijn de
morele conclusies dus niet voor iedereen bindend, maar alleen voor diegene die toevallig
het contingente zelfbegrip van de Kantiaans constructivist deelt.
Ik onderscheid drie mogelijke ‘interessante morele conclusies:’ (1) categorische
normatieve uitspraken; (2) categorische en universele normatieve uitspraken; (3)
substantiële, categorische en universele normatieve uitspraken. Een categorische
normatieve uitspraak (1) is een normatieve uitspraak die elk persoon moet accepteren,
onafhankelijk van zijn of haar contingente overtuigingen of voorkeuren. Een categorische
en universele categorische uitspraak (2) is een categorische uitspraak die alle mensen als
object heeft. Een substantiële categorische en universele normatieve uitspraak (3), is een
categorische en universele uitspraak die tot substantiële restricties op handelen leidt, dat wil
SAMENVATTING
346
zeggen een uitspraak die ons vertelt hoe wel/niet moeten handelen.
In hoofdstuk 4 bespreek ik de kritiek dat het Kantiaans constructivisme geen enkele
categorische normatieve uitspraak kan rechtvaardigen (1) omdat er geen enkel
onvermijdelijk zelfbegrip zou bestaan, of omdat er geen enkele normatieve conclusie kan
volgen uit een transcendentaal argument. Ik argumenteer dat deze tegenwerpingen falen als
algemene tegenwerpingen tegen het Kantiaans constructivisme, maar dat dit openlaat dat er
specifieke startpunten zijn die vermijdelijk zijn of dat er specifieke transcendentale
argumenten zijn die niet tot normatieve uitspraken kunnen leiden.
In hoofdstuk 6 tot en met 8 bespreek ik de tegenwerping dat het Kantiaans
constructivisme geen universele normatieve uitspraken kan rechtvaardigen (2). In deze
context maak ik een onderscheid tussen transcendentale argumenten die beginnen vanuit
het eerstepersoonsperspectief van een handelend wezen en transcendentale argumenten die
beginnen vanuit het tweedepersoonsperspectief van iemand die deelneemt aan bepaalde
vormen van interactie zoals communicatie. In hoofdstuk 5 beargumenteer ik dat
transcendentale argumenten die de tweede persoon als startpunt hebben, vatbaar zijn voor
het algemene dilemma voor Kantiaans constructivisme. Dus: oftewel interactie is
onvermijdelijk, maar uit deze vorm van interactie volgen geen universele morele principes,
of er volgen universele principes uit interactie, maar in dat geval is de vorm van interactie
die het startpunt van het argument volgt slechts optioneel.
In hoofdstuk 6 bekritiseer ik het transcendentale argument vanuit de eerste persoon
van Christine Korsgaard voor de conclusie dat menselijkheid een universele waarde is. Ik
onderscheid en bespreek verschillende reconstructies van haar argument en ik
beargumenteer dat Korsgaard’s argument ook vatbaar is voor het algemene dilemma voor
het Kantiaans constructivisme.
In hoofdstuk 7 verdedig ik het transcendentale argument vanuit de eerste persoon van
Alan Gewirth. Ik beargumenteer dat het startpunt van zijn argument onvermijdelijk is en
dat hij er in slaagt om een universeel moreel principe te rechtvaardigen op basis van dit
startpunt. Volgens dit principe heeft elk handelend wezen een recht op de noodzakelijke
middelen om een handelend wezen te zijn.
In hoofdstuk 8 en 9 bespreek ik ten slotte de vraag in hoeverre Kantiaans
constructivisme ook tot substantiële handelingsnormen kan leiden. Dit is cruciaal, want
zelfs als je ervan overtuigd bent dat een transcendentaal argument kan leiden tot een
categorisch en universeel principe (zoals beweerd in hoofdstuk 7), dan zou het nog steeds
SAMENVATTING
347
kunnen dat dit principe te abstract of te formeel is om te leiden tot concrete, substantiële
morele normen.
Daarom introduceer ik in hoofdstuk 8 de tegenwerping dat het Kantiaans
constructivisme alleen ‘lege’ principes kan rechtvaardigen – een objectie die teruggaat op
Hegel’s objectie tegen Kant’s ethiek. Ik bespreek de algemeen Kantiaans constructivistische
strategie om substantiële normen te rechtvaardigen door gebruik te maken van
antropologische feiten in de rechtvaardiging van moraliteit. Ik argumenteer dat
‘antropologische feiten’ inderdaad noodzakelijk zijn voor het rechtvaardigen van
substantiële normen, en dat het gebruik van deze feiten binnen het Kantiaans
constructivisme legitiem is.
In hoofdstuk 9 bespreek ik de vraag of het gebruik van deze feiten ook voldoende is
voor het rechtvaardigen van substantiële normen. Ik doe dit door het evalueren van twee
specifieke Kantiaans constructivistische theorieën, namelijk Onora O’Neill’s theorie en
Alan Gewirth’s theorie. Ik argumenteer dat hoewel geen van beide theorieën zogezegd ‘leeg’
is, ze geen overtuigende argumenten geven voor de rijke set van normen die ze zelf
verdedigen. Het probleem is dat het specifieke idee van wat het betekent om jezelf als een
handelend wezen te begrijpen ondergespecificeerd is, en dat er een bepaalde neiging is om
te verschuiven van een ‘dun’ begrip van dit zelfbegrip in de rechtvaardiging van een
abstract moreel principe, naar een ‘dikker’ begrip in de toepassing van dit principe voor de
rechtvaardiging van substantiële normen.
Mijn conclusie met betrekking tot het succes van het Kantiaans constructivisme is dus
gematigd optimistisch, of gematigd pessimistisch, afhankelijk van hoe je ernaar kijkt. In het
proefschrift heb ik laten zien dat het Kantiaans constructivisme een fundamenteel, objectief
principe van moraliteit kan rechtvaardigen, en dat dit principe meer is dan een leeg
formalisme. Ik heb echter kanttekeningen geplaatst bij de manier waarop Kantiaans
constructivisten een relatief uitgebreide set aan morele normen verdedigen. Dit betekent
niet dat het Kantiaans constructivisme faalt als een substantiële morele theorie, maar vooral
dat er meer werk nodig is op dit gebied, en meer specifiek betreffende de vraag wat het
precies betekent om jezelf als een handelend wezen te zien.
Curriculum Vitae
Sem de Maagt (1987) received his Bachelor’s degree from University College Maastricht
(2009, cum laude) and his Research Master’s degree in Philosophy from Utrecht University
(2011, cum laude). Sem wrote his PhD thesis at the Faculty of Philosophy of the Erasmus
University Rotterdam (2011-2014) and the Ethics Institute of Utrecht University (20142016). His thesis was supervised by Prof.dr. Ingrid Robeyns, Prof.dr. Marcus Düwell and
Dr. Rutger Claassen. During his PhD project, Sem spent a term at the Manchester Centre
for Political Theory (2014) and another term at the London School of Economics (2015).
Sem is also the co-author of a Dutch book on the philosophy of the self. At present, he is a
temporary lecturer at the Ethics Institute of Utrecht University.
Quaestiones Infinitae
PUBLICATIONS OF THE DEPARTMENT OF PHILOSOPHY
AND RELIGIOUS STUDIES
VOLUME 21.
VOLUME 22.
VOLUME 23.
VOLUME 24.
VOLUME 25.
VOLUME 26.
VOLUME 27.
VOLUME 28.
VOLUME 29.
VOLUME 30.
VOLUME 31.
VOLUME 32.
VOLUME 33.
VOLUME 34.
VOLUME 35.
VOLUME 36.
VOLUME 37.
VOLUME 37A.
VOLUME 38.
VOLUME 39.
VOLUME 40.
VOLUME 41.
D. VAN DALEN, Torens en Fundamenten (valedictory lecture), 1997.
J.A. BERGSTRA, W.J. FOKKINK, W.M.T. MENNEN, S.F.M. VAN VLIJMEN,
Spoorweglogica via EURIS, 1997.
I.M. CROESE, Simplicius on Continuous and Instantaneous Change
(dissertation), 1998.
M.J. HOLLENBERG, Logic and Bisimulation (dissertation), 1998.
C.H. LEIJENHORST, Hobbes and the Aristotelians (dissertation), 1998.
S.F.M. VAN VLIJMEN, Algebraic Specification in Action (dissertation), 1998.
M.F. VERWEIJ, Preventive Medicine Between Obligation and Aspiration
(dissertation), 1998.
J.A. BERGSTRA, S.F.M. VAN VLIJMEN, Theoretische Software-Engineering:
kenmerken, faseringen en classificaties, 1998.
A.G. WOUTERS, Explanation Without A Cause (dissertation), 1999.
M.M.S.K. SIE, Responsibility, Blameworthy Action & Normative
Disagreements (dissertation), 1999.
M.S.P.R. VAN ATTEN, Phenomenology of choice sequences (dissertation),
1999.
V.N. STEBLETSOVA, Algebras, Relations and Geometries (an equational
perspective) (dissertation), 2000.
A. VISSER, Het Tekst Continuüm (inaugural lecture), 2000.
H. ISHIGURO, Can we speak about what cannot be said? (public lecture),
2000.
W. HAAS, Haltlosigkeit; Zwischen Sprache und Erfahrung (dissertation),
2001.
R. POLI, ALWIS: Ontology for knowledge engineers (dissertation), 2001.
J. MANSFELD, Platonische Briefschrijverij (valedictory lecture), 2001.
E.J. BOS, The Correspondence between Descartes and Henricus Regius
(dissertation), 2002.
M. VAN OTEGEM, A Bibliography of the Works of Descartes (1637-1704)
(dissertation), 2002.
B.E.K.J. GOOSSENS, Edmund Husserl: Einleitung in die Philosophie:
Vorlesungen 1922/23 (dissertation), 2003.
H.J.M. BROEKHUIJSE, Het einde van de sociaaldemocratie (dissertation),
2002.
P. RAVALLI, Husserls Phänomenologie der Intersubjektivität in den
VOLUME 42.
VOLUME 43.
VOLUME 44.
VOLUME 45.
VOLUME 46.
VOLUME 47.
VOLUME 48.
VOLUME 49.
VOLUME 50.
VOLUME 51.
VOLUME 52.
VOLUME 53.
VOLUME 54.
VOLUME 55.
VOLUME 56.
VOLUME 57.
VOLUME 58.
VOLUME 59.
VOLUME 60.
VOLUME 61.
VOLUME 62.
VOLUME 63
VOLUME 64
VOLUME 65
VOLUME 66
Göttinger Jahren: Eine kritisch-historische Darstellung (dissertation), 2003.
B. ALMOND, The Midas Touch: Ethics, Science and our Human Future
(inaugural lecture), 2003.
M. DÜWELL, Morele kennis: over de mogelijkheden van toegepaste ethiek
(inaugural lecture), 2003.
R.D.A. HENDRIKS, Metamathematics in Coq (dissertation), 2003.
TH. VERBEEK, E.J. BOS, J.M.M. VAN DE VEN, The Correspondence of René
Descartes: 1643, 2003.
J.J.C. KUIPER, Ideas and Explorations: Brouwer’s Road to Intuitionism
(dissertation), 2004.
C.M. BEKKER, Rechtvaardigheid, Onpartijdigheid, Gender en Sociale
Diversiteit; Feministische filosofen over recht doen aan vrouwen en hun
onderlinge verschillen (dissertation), 2004.
A.A. LONG, Epictetus on understanding and managing emotions (public
lecture), 2004.
J.J. JOOSTEN, Interpretability formalized (dissertation), 2004.
J.G. SIJMONS, Phänomenologie und Idealismus: Analyse der Struktur und
Methode der Philosophie Rudolf Steiners (dissertation), 2005.
J.H. HOOGSTAD, Time tracks (dissertation), 2005.
M.A. VAN DEN HOVEN, A Claim for Reasonable Morality (dissertation),
2006.
C. VERMEULEN, René Descartes, Specimina philosophiae: Introduction and
Critical Edition (dissertation), 2007.
R.G. MILLIKAN, Learning Language without having a theory of mind
(inaugural lecture), 2007.
R.J.G. CLAASSEN, The Market’s Place in the Provision of Goods
(dissertation), 2008.
H.J.S. BRUGGINK, Equivalence of Reductions in Higher-Order Rewriting
(dissertation), 2008.
A. KALIS, Failures of agency (dissertation), 2009.
S. GRAUMANN, Assistierte Freiheit (dissertation), 2009.
M. AALDERINK, Philosophy, Scientific Knowledge, and Concept Formation
in Geulincx and Descartes (dissertation), 2010.
I.M. CONRADIE, Seneca in his cultural and literary context: Selected moral
letters on the body (dissertation), 2010.
C. VAN SIJL, Stoic Philosophy and the Exegesis of Myth (dissertation), 2010.
J.M.I.M. LEO, The Logical Structure of Relations (dissertation), 2010.
M.S.A. VAN HOUTE, Seneca’s theology in its philosophical context
(dissertation), 2010.
F.A. BAKKER, Three Studies in Epicurean Cosmology (dissertation), 2010.
T. FOSSEN, Political legitimacy and the pragmatic turn (dissertation), 2011.
T. VISAK, Killing happy animals. Explorations in utilitarian ethics.
(dissertation), 2011.
VOLUME 67
VOLUME 68
VOLUME 69
VOLUME 70
VOLUME 70A
VOLUME 71
VOLUME 72
VOLUME 73
VOLUME 74
VOLUME 75
VOLUME 76
VOLUME 77
VOLUME 78
VOLUME 79
VOLUME 80
VOLUME 81
VOLUME 82
VOLUME 83
VOLUME 84
VOLUME 85
VOLUME 86
VOLUME 86A
A. JOOSSE, Why we need others: Platonic and Stoic models of friendship and
self-understanding (dissertation), 2011.
N. M. NIJSINGH, Expanding newborn screening programmes and
strengthening informed consent (dissertation), 2012.
R. PEELS, Believing Responsibly: Intellectual Obligations and Doxastic
Excuses (dissertation), 2012.
S. LUTZ, Criteria of Empirical Significance (dissertation), 2012
G.H. BOS, Agential Self-consciousness, beyond conscious agency
(dissertation), 2013.
F.E. KALDEWAIJ, The animal in morality: Justifying duties to animals in
Kantian moral philosophy (dissertation), 2013.
R.O. BUNING, Henricus Reneri (1593-1639): Descartes’ Quartermaster in
Aristotelian Territory (dissertation), 2013.
I.S. LÖWISCH, Genealogy Composition in Response to Trauma: Gender and
Memory in 1 Chronicles 1-9 and the Documentary Film ‘My Life Part 2’
(dissertation), 2013.
A. EL KHAIRAT, Contesting Boundaries: Satire in Contemporary Morocco
(dissertation), 2013.
A. KROM, Not to be sneezed at. On the possibility of justifying infectious
disease control by appealing to a mid-level harm principle (dissertation),
2014.
Z. PALL, Salafism in Lebanon: local and transnational resources
(dissertation), 2014.
D. WAHID, Nurturing the Salafi Manhaj: A Study of Salafi Pesantrens in
Contemporary Indonesia (dissertation), 2014.
B.W.P VAN DEN BERG, Speelruimte voor dialoog en verbeelding.
Basisschoolleerlingen maken kennis met religieuze verhalen (dissertation),
2014.
J.T. BERGHUIJS, New Spirituality and Social Engagement (dissertation),
2014.
A. WETTER, Judging By Her. Reconfiguring Israel in Ruth, Esther and
Judith (dissertation), 2014.
J.M. MULDER, Conceptual Realism. The Structure of Metaphysical Thought
(dissertation), 2014.
L.W.C. VAN LIT, Eschatology and the World of Image in Suhrawardī and
His Commentators (dissertation), 2014.
P.L. LAMBERTZ, Divisive matters. Aesthetic difference and authority in a
Congolese spiritual movement ‘from Japan’ (dissertation), 2015.
J.P. GOUDSMIT, Intuitionistic Rules: Admissible Rules of Intermediate
Logics (dissertation), 2015.
E.T. FEIKEMA, Still not at Ease: Corruption and Conflict of Interest in
Hybrid Political Orders (dissertation), 2015.
N. VAN MILTENBURG, Freedom in Action (dissertation), 2015.
P. COPPENS, Seeing God in This world and the Otherworld: Crossing
VOLUME 87
VOLUME 88
VOLUME 89
VOLUME 90
VOLUME 91
VOLUME 92
VOLUME 93
VOLUME 94
VOLUME 95
VOLUME 96
VOLUME 97
VOLUME 98
VOLUME 99
Boundaries in Sufi Commentaries on the Qurʾān (dissertation), 2015.
D.H.J. JETHRO, Aesthetics of Power: Heritage Formation and the Senses in
Post-apartheid South Africa (dissertation), 2015.
C.E. HARNACKE, From Human Nature to Moral Judgement: Reframing
Debates about Disability and Enhancement (dissertation), 2015.
X. WANG, Human Rights and Internet Access: A Philosophical
Investigation (dissertation), 2016.
R. VAN BROEKHOVEN, De Bewakers Bewaakt: Journalistiek en leiderschap
in een gemediatiseerde democratie (dissertation), 2016.
A. SCHLATMANN, Shi‘i Muslim youth in the Netherlands: Negotiating Shi‘i
fatwas and rituals in the Dutch context (dissertation), 2016.
M.L. VAN WIJNGAARDEN, Schitterende getuigen. Nederlands luthers
avondmaalsgerei als indenteitsdrager van een godsdienstige minderheid
(dissertation), 2016.
S. COENRADIE, Vicarious substitution in the literary work of Shūsaku Endō.
On fools, animals, objects and doubles (dissertation), 2016.
J. RAJAIAH, Dalit Humanization. A quest based on M.M. Thomas’ theology
of salvation and humanization (dissertation), 2016.
D.L.A. OMETTO, Freedom & Self-knowledge (dissertation), 2016.
Y. YALDIZ, The Afterlife in Mind: Piety and Renunciatory Practice in the
2nd/8th- and early 3rd/9th-Century Books of Renunciation (Kutub alZuhd) (dissertation), 2016.
M.F. BYSKOV, Between experts and locals. Towards an inclusive framwork
for a development agenda (dissertation), 2016.
A. RUMBERG, Transitions toward a Semantics for Real Possibility
(dissertation), 2016.
S. DE MAAGT, Constructing Morality. Transcendental Arguments in Ethics
(dissertation), 2017.

Similar documents

×

Report this document