Naturalism and Realism in Kants Ethics

8/19/2019 Naturalism and Realism in Kants Ethics

1/274


2/274


3/274

Naturalism and Realism in Kant ’s Ethics

In this comprehensive assessment of Kant ’s metaethics, Frederick Rauscher

shows that Kant is a moral idealist rather than a moral realist and argues that

Kant ’s ethics does not require metaphysical commitments that go beyond

nature. Rauscher frames the argument in the context of Kant ’s nonnaturalistic

philosophical method and the character of practical reason as action-oriented.

Reason operates entirely within nature, and apparently nonnatural claims –

God, free choice, and value – are shown to be heuristic and to reect reason’s

ordering of nature. The book shows how Kant hesitates between a transcen-

dental moral idealism with an empirical moral realism and a complete moral

idealism. Examining every aspect of Kant ’s ethics, from the categorical impe-

rative to freedom and value, this volume argues that Kant ’s focus on human

moral agency explains morality as a part of nature. It will appeal to academic

researchers and advanced students of Kant, German idealism, and intellectual

history.

f r e d e ri c k r a u s c he r is Professor of Philosophy at Michigan State

University. He is the editor and co-translator of Kant: Lectures and Drafts

on Political Philosophy (with Kenneth R. Westphal, Cambridge, 2015),

co-translator of Notes and Fragments (with Paul Guyer and Curtis Bowman,

Cambridge, 2005), and editor of Kant in Brazil (2012).


4/274


5/274

Naturalism and Realism in

Kant ’s Ethics

Frederick Rauscher


6/274

University Printing House, Cambridge CB2 8BS, United Kingdom

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence.

www.cambridge.orgInformation on this title: www.cambridge.org/9781107088801

© Frederick Rauscher 2015

This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place without the writtenpermission of Cambridge University Press.

First published 2015 A catalogue record for this publication is available from the British Library

Library of Congress Cataloging in Publication dataRauscher, Frederick, 1961–Naturalism and realism in Kant ’s ethics / Frederick Rauscher. – 1 [edition].

pages cm Includes bibliographical references and index.ISBN 978-1-107-08880-1 (Hardback : alk. paper)1. Kant, Immanuel, 1724–1804. 2. Ethics. 3. Naturalism. 4. Realism. I. Title.B2799.E8R38 2015170.92–dc23 2015020982

ISBN 978-1-107-08880-1 Hardback

Cambridge University Press has no responsibility for the persistence or accuracyof URLs for external or third-party internet websites referred to in this publication,and does not guarantee that any content on such websites is, or will remain,accurate or appropriate.

http://www.cambridge.org/http://www.cambridge.org/9781107088801http://www.cambridge.org/http://www.cambridge.org/9781107088801


7/274


8/274

Acknowledgments

I thank rst and more than anyone else my spouse Delores and my children

Konrad, Bennett, and Audrey for their love and support as I spent long days

working on this book over too many years. Without their patience this book

would never have been nished. This book is for them.Numerous colleagues and students have helped me to shape my ideas and

I here want to particularly thank several people who most generously com-

mented on this book as it was being drafted. Darlei Dall’Agnol pushed and

prodded me into better explaining and defending my views both in conversa-

tions and through his published criticisms. I greatly beneted from conversa-

tions and exchanges with Oliver Sensen and Patrick Kain. All three of them,

along with Paul Guyer, also kindly read parts of the nal manuscript. Robert

Louden and an anonymous reviewer for Cambridge University Press provided

extensive comments that helped me to improve my presentation and argu-ments. I am deeply grateful to them all, as well as to those who helped me in

earlier years when some of this material was published in independent papers.

I am also indebted to Cambridge University Press editors Hilary Gaskin and

Rosemary Crawley for their guidance in this project.

I am grateful to Michigan State University for a research leave and Inter-

mural Research Grant that allowed initial work on the book, and a sabbatical

that allowed me to complete it. In between, I was grateful to have many oppor-

tunities to present my work, resulting in quite a different book than originally

conceived. The Federal University of Santa Catarina, the Federal University of Rio Grande do Norte, the Federal University of Pelotas, the Federal University

of Pernambuco, the Pontical Catholic University in Rio de Janeiro, the Ponti-

cal Catholic University of Parana, the State University of Campinas, and

the University of São Paulo in Brazil, Pisa University in Italy, the Philipps

University in Marburg and the University of Siegen in Germany, the Uni-

versity of Illinois, University of Pennsylvania, Brown University, Western

Michigan University, and Michigan State University in the United States all

provided opportunities either directly or by hosting conferences for me to

present this work in various stages of development. I am indebted to the many

vi


9/274

contributors to discussions at these events who have stimulated my thought

and caused me to deepen my interpretation.

I wish also to thank the original publishers of material reproduced here.

With the exception of “‘God’ Without God”, which forms the content of

Chapter 5 with little alteration, these papers have been excerpted and/or modied. In most cases, only parts of the paper are included in the book, even

scattered over several chapters.

Part of Chapter 4: “Freedom and Reason in Groundwork III” in

Kant ’ s Groundwork of the Metaphysics of Morals: A Critical

Guide, edited by Jens Timmermann (Cambridge University Press,

2009), pp. 203–23.

Part of Chapter 2: “Why Kant ’s Ethics is A Priori – and Why It

Matters”, in Recht und Frieden in der Philosophie Kants: Akten

des X Internationalen Kant-Kongresses, Band 3, hrsg. Valerio

Rohden, Ricardo Terra, Guido de Almeida, and Margit Ruf ng

(Berlin: Walter deGruyter, 2008), pp. 347–57.

Chapter 5: “‘God’ Without God: Kant ’s Postulate” Kant e-Prints

Série 2, v. 2, n. 1, jan.– jun., (2007), pp. 27–62 [www.cle.unicamp.

br/kant-e-prints/ ]

Part of Chapter 4: “Reason as a Natural Cause”, in Moralische Moti-

vation. Kant und die Alternativen, edited by Heiner F. Klemme,

Manfred Kühn, and Dieter Schönecker. Reihe Kant Forschungen

Band 16. (Hamburg: Felix Meiner Verlag, 2006), pp.97–110.

Parts of Chapters 1 and 3: “Kant ’s Moral Anti-Realism ”, Journal of

the History of Philosophy 40 (2002): 477–99.

Parts of Chapters 2 and 3: “Kant ’s Two Priorities of Practical

Reason”, British Journal for the History of Philosophy 6 (3)

(1998): 397–419.

Parts of Chapters 2 and 4 were originally published in Portuguese as

“Razão prática pura como uma faculdade natural” [“Pure Practical

Reason as a Natural Faculty”], translated by Milene Consenso

Tonetto, Ethic@ 5 (2006), pp. 173–192 [https://periodicos.ufsc.

br/index.php/ethic/issue/view/1453]

Part of Chapter 7 was originally published in Portuguese as “Os

limites externos da losoa prática e as limitações da Dedução na

Fundamentação III”, translated by Kariel Giarolo, Studia Kantiana

14 (2013): 127–41 and in German as “Die äußerste Grenze aller

praktischen Philosophie und die Einschränkungen der Deduktion in

Grundlegung III”, in Kants Begründung von Freiheit und Moral in

Grundlegung III: Neue Interpretationen edited by Dieter Schönecker,

(Münster: Mentis Verlag, 2015), pp. 217–31.

Acknowledgments vii

http://www.cle.unicamp.br/kant-e-prints/http://www.cle.unicamp.br/kant-e-prints/https://periodicos.ufsc.br/index.php/ethic/issue/view/1453https://periodicos.ufsc.br/index.php/ethic/issue/view/1453http://www.cle.unicamp.br/kant-e-prints/http://www.cle.unicamp.br/kant-e-prints/https://periodicos.ufsc.br/index.php/ethic/issue/view/1453https://periodicos.ufsc.br/index.php/ethic/issue/view/1453


10/274


11/274

Introduction

Among the rst conceptions I had of Kant ’s philosophy, provided second- or

third-hand when I was in college before I had studied him enough to be able to

judge it, was the tale that Kant had divided the world into two separate realms,

that of appearances in space and time, constituting nature, and that of things inthemselves eerily existing not in space and time, constituting morality. This

neat division seemed like a tidy way of dealing with the potential conict

between our moral lives and the scientic world view. As I learned more about

Kant, it became clear that the ontological division was not that simple and that

whatever it was, it did not map onto the nature/morality division. The proper

understanding of these issues became a lasting puzzle.

This book is an attempt to solve that puzzle by showing what morality is and

just how nature and morality relate to one another in Kant. I have two main

goals. The rst is to show that Kant ’s ethics is fully compatible with a

metaphysical naturalism, meaning that no property or entity outside of empir-

ically real nature in space and time is needed. The second is to determine the

extent to which Kant is a moral realist, which can be decided only through a

detailed look at the nature of Kant ’s ethics and its specic elements. I will

conclude that the most plausible interpretation is that Kant is a moral idealist

(the term I prefer to “anti-realist ” or “constructivist ”) rather than a realist,

although given the resources in his philosophy, he could have been a realist in

a limited sense.

Part One, “Laying the Ground,”

sets the stage for the detailed assessment byproviding denitions of metaphysical naturalism and moral realism and by

showing how ethics ts into Kant ’s philosophical project as a whole. This part

is crucial for the overall project because it assesses the inadequacy of some

ways of approaching realism and naturalism, and provides an explanation for

the particular approach I take. The review of Kant ’s philosophical project

importantly shows the way in which transcendental philosophy allows for a

metaphysical naturalism and the way in which the claims of practical reason

are prima facie not ontological. While some readers are familiar with these

basic positions, the details and conclusions I draw here bear specically on theparticular topics in later chapters and form the foundation of their arguments.

1


12/274

Chapter 1 provides a denition of realism in terms of the independence of some

moral principles, properties, or objects from the moral agent. I argue that this is a

better denition for use in assessing Kant than one focused on the truth of moral

claims because the real issue between realists and idealists is not whether Kant ’s

morality claims objective validity but how to understand Kant ’s a priori moral

law, the nature of practical reason, and autonomy in relation to moral agency.

I treat moral agency in two ways: as actual agents in nature (empirical) and as the

necessary conditions for the possibility of any moral agent at all (that I label

“transcendental” in reference to the transcendental method of justication).

This distinction allows for Kant to be both an idealist and a realist at different

levels, and provides the complexity necessary to resolve the multi-faceted issue.

I then turn to naturalism, which is divided into methodological and metaphysical

naturalism. The former would claim that the only proper methodologies for

nding knowledge are those of the natural sciences; Kant rejects this in light of his use of a priori concepts and his transcendental method. I explain metaphysical

naturalism, which claims that the only entities that exist are those determined by

the natural sciences, in relation to Kant ’s own conception of nature as consisting

of matter studied by physics (and less strictly also by chemistry and biology) and

thinking nature studied by empirical psychology (and related disciplines).

A metaphysically naturalistic Kantian ethics would hold that nothing beyond

the entities in space and time, physical and mental, is needed for morality. The

chapter concludes with a list of the eight elements of Kant ’s ethics that need to be

assessed as realist or idealist: particular ends, particular duties, absolute value, thehighest good and the postulates, moral obligation, the moral law itself, pure

practical reason, and free choice. Thus, the task of the remainder of the book is

to assess these elements in terms of the transcendental and empirical levels of

realism and the possible limitation to nature in space and time.

Chapter 2 examines the nature of Kant ’s philosophy overall in order to place

morality in its proper context and show how that framework affects realism

and naturalism. I note that Kant insists that philosophy aims at systematically

organized cognitions, with the overall aim at the essential ends of human

reason. Each part requires a domain and further subdivision into parts inaccordance with a system. I look at several broad divisions of philosophy in

Kant – critique/system, pure/empirical cognition, and theoretical/practical, and

note how they interrelate. I show how the method of transcendental argument

provides Kant with a general way to defend ethical claims without resorting to

a non-natural metaphysics. Transcendental argument justies the use of a

priori cognitions by cognizers in nature as part of their empirical cognitive

systems. In looking at various ways to construe the theoretical/practical dis-

tinction, I show that the domain for practical philosophy is free acts understood

from the perspective of agents engaged in deliberation. With these founda-tional issues settled, the detailed work can begin.

2 Introduction


13/274


14/274

when he denies that the moral law can be subject to a transcendental deduction

and when he limits his claim to the practical point of view. Kant would then be

an empirical moral idealist about reason and the status of the moral law.

The nal part of the book, “Morality Beyond Nature?,” looks at God, free

choice, and absolute value. These elements of Kant ’s ethics represent apparently

non-natural entities or properties and cannot be understood in the same meta-

physically naturalistic way that reason itself can, that is, they cannot be an actual

part of nature operating as the empirically real manifestation of the transcen-

dental conditions of moral agency. Instead, I show that God and free choice have

a role through the postulates that comprehends them as concepts created by

reason without reference to the purported non-natural object or property but

only with an immanent reference as aids to moral action. I argue that value for

Kant cannot be an independent property of objects either within or outside

nature, but is merely a part of the order imposed on nature by practical reason.I then link that order back to the status of practical reason itself.

Chapter 5 argues that Kant ’s conception of a postulate has much more in

common with the general nature of ideas of reason than it does with any claim

to existence. I trace Kant ’s development of the notion of a postulate along with

the similar notions of transcendental hypothesis, idea of reason, and belief.

I show that the idea of the highest good has no ontological implications of its

own and focus on the postulate of God, since the same solution will also hold

of the postulate of immortality. The postulate of God, I show, is supposed to

have immanent reference, that is, to empirical agents’

moral lives, rather thantranscendent reference, that is, to a being in itself. In the practical point of

view, empirical moral agents operate with the concept of God for certain

purposes but relate it only to the ought, not to the is.

Chapter 6 assesses the most dif cult problem for a naturalist interpretation

of Kant ethics: freedom of the will as the freedom of the power of choice in

making a decision uncaused by anything in nature. I have two strategies in this

chapter. I rst show that Kant insists that free choice is needed for two reasons

related to ought-implies-can and moral responsibility. Both reasons require

only one non-natural choice, not a multitude as some commentators prefer.This single timeless choice of the entirety of one’s phenomenal character is the

best interpretation of his claims about freedom, timelessness, distinct causality,

and the intelligible character in relation to free choice. While still unnatural, it

is at least the minimal non-natural interpretation. Second, I argue that stressing

the status of freedom as a postulate, which Kant himself does not, allows for a

naturalistic interpretation of this choice as a concept merely playing a heuristic

role in moral life.

The nal chapter moves from value in particular to the status of the practical

point of view, or moral experience more broadly, in general. I reject thepossibility of any non-natural, intrinsic value property and instead show that

4 Introduction


15/274

the value of humanity as an end in itself is nothing more than the highest rank

in the order of ends that reason imposes on nature through the categorical

imperative. The intelligible order of things is not an order of intelligible things

but a rational ordering of natural things. I draw together Kant ’s various

discussions of this direct application of the moral law to experience as a moralworld. This is the way that practical reason applies to nature within the

practical point of view. Culminating this chapter is a look at the limits to

practical reason that Kant reveals in Groundwork III, where he admits that

reason’s own structure that requires both systematic connection and uncondi-

tional explanation is responsible for the claim that there is a necessary moral

law, and holds, in language similar to that of the Third Antinomy, that reason,

the source of morality, is itself ultimately incomprehensible.

In a postscript, I review the particular assessments I made regarding the

eight elements of Kant ’s moral theory laid out in Chapter 1. I pull together the

features of my interpretation of Kant as a metaphysical naturalist. The various

claims about transcendental and empirical realism and idealism are arranged

into their basic sets and the core interpretive points that ground the main

disagreement set out. I have identied a Kantian transcendental moral idealism

that is also an empirical realism, thus dissolving some of the realist/construct-

ivist disagreement. I show, however, that Kant is himself hesitant to endorse

this transcendental validity for morality and, particularly in light of the priority

of the practical point of view as an agent-perspective rational ordering of

nature with no ontological claims of its own, that the more appropriateconclusion is that Kant was an empirical moral idealist.

A brief word about my methodological approach. I am not providing a

strictly exegetical work. There are passages in Kant ’s writings that are incon-

sistent with elements of my interpretation, but I would argue that the same is

true of all interpretations of Kant given his own inconsistent use of termin-

ology and the diverse contexts in which he applies the same terms. The

purpose of this book is not merely historical but is aimed at assessing Kant ’s

ethics in light of twenty-rst century concerns about naturalism and realism.

My work is reconstructive in that I pursue the philosophical implications of Kant ’s positions to sometimes make connections that Kant himself does not

explicitly make. I believe that all of my claims are consistent with Kant ’s

general philosophical aims, methods, arguments, and conclusions, and nearly

always with his particular stated positions. Given the novelty of Kant ’s

philosophy, the complexity of the issues he raises, the vast range of his project,

the transformations of some of his positions over time, and above all, the

requirement of the interpreter to go beyond merely quoting texts in order to

explain a great philosopher ’s work in ways relevant to the present, no compre-

hensive interpretation can offer more than that.

Introduction 5


16/274


17/274

Part I

Laying the ground


18/274


19/274

1 Moral realism and naturalism

My study aims at providing an interpretation of Kant ’s ethics that cuts across

the issues of realism and naturalism. I see these two issues as closely con-

nected. One cannot resolve the question of whether and to what extent Kant

was a moral realist without resolving questions about the metaphysical statusof the elements of Kant ’s ethics. These elements, such as the status of reason

itself, the nature of value, and freedom of the will, relate to the conception

Kant has of the limits of human experience and legitimate claims that go

beyond experience. This concern in turn raises the issue of nature as a limit for

experience. Whether ethics goes beyond nature is in this way tied to the issue

of realism. The reverse is also true: When asking about whether Kant ’s ethics

is compatible with naturalism, one has to start with a conception of nature in

Kant and then turn to the particular elements. One would expect that anything

in nature would count as real for Kant, and on the empirical level, that is true of objects and of mental states and faculties. Questions arise, however, about how

to place in nature the distinctive perspective of the practical and its distinctive

role in determining what ought to be rather than what is. And since Kant also

invokes a transcendental level of analysis, some elements of the experience

moral beings have in nature might be ideal in the same sense that space and

time and the categories are ideal. A detailed understanding of what might be

real and what ideal, and in what senses, will help to resolve the issue about

whether morality requires more than is available in nature.

Both the term “naturalism ”

and the term “moral realism ”

need to bespecied and adapted to the peculiarities of Kant ’s philosophy. Contemporary

philosophical work on these issues does not always approach these issues in a

way amenable to Kant ’s critical philosophy. Kant ’s different conceptions of

the sciences and his transcendental idealism inform his understanding of

nature. His use of the terms “realism ” and “idealism ” in both empirical and

transcendental senses, and the very nature of his critical philosophy as largely

subject-centered, do not easily map onto discussions of realism in terms of

objective facts or empirical properties of human nature.

My aim in this chapter is twofold. First, I will provide an analysis of denitions of moral realism and nonrealism with a focus on how well they

9


20/274

can account for the peculiarities of Kant ’s general philosophical approach.

I use the term “nonrealism ” to capture all alternatives to moral realism until

their particular characteristics are dened in this chapter. The analysis of

realism I adopt stresses the independence of elements of morality from moral

agency as such; I contrast realism with idealism. I examine the transcendentaland empirical levels of realism and idealism in Kant and show where my

approach differs from the typical emphases in constructivism.

Second, I will examine the meaning of “naturalism ”. I stress that my aim is

limited to showing that Kant ’s ethics is compatible with a metaphysical

naturalism, understood in Kant ’s own terms as including both the physical

and the mental. He rejects methodological naturalism. In fact, his methodo-

logical antinaturalism will play an important role in my interpretation.

Third, I will identify the particular elements of Kant ’s moral theory that can

be interpreted as real or ideal, natural or nonnatural. This taxonomy willilluminate the point that inquiring whether Kant is a moral realist or naturalist

is not a simple yes-or-no question. On different levels and about different

elements, Kant holds realist as well as idealist views. Further, by highlighting

particular issues, this taxonomy will push the debate beyond mere terminology

to the concrete differences among Kant interpreters.

Dening moral realism and moral idealism

Philosophy over the past three decades has included extensive discussion of the nature of moral realism and its alternatives.1 The issue entered the Kantian

literature with John Rawls’ John Dewey Lectures in April 1980, “Kantian

Constructivism in Moral Theory”,2 which also introduced the term “construct-

ivism ” into the literature on Kant ’s practical philosophy.3 Rawls’ intention was

not simply to provide an interpretation of Kant ’s ethics along constructivist

lines but mainly to present Kantian constructivism as a general moral theory

alongside utilitarianism, intuitionism, and perfectionism.4 Kant commentators

1 Some of the earliest contributions to the debate are collected in (Sayre-McCord 1988a ). My

discussion will invoke contributions to the debate about moral realism in general only whendoing so is useful for understanding the particular debate within Kant circles.

2 The Dewey Lectures were published in (Rawls 1980) and reprinted in (Rawls 1999b,pp. 303–58).

3 Larry Krasnoff traces the rst use of the term “constructive” in moral theory to a review byRonald Dworkin in 1973 (reprinted in [Dworkin 1978, pp. 150–83]) of John Rawls’ book

A Theory of Justice. Rawls did not use the term “constructivism ” in A Theory of Justice but soon adopted it to describe his own theory. See (Krasnoff 1999).

4 Since constructivism was introduced by Rawls, one might think that it can be applied in politicalphilosophy in addition to moral theory. In this book, I exclude directly political principles in

favor of moral ones, broadly speaking. For work on the political use of constructivism, whichdraws more directly on Rawls’ own work in A Theory of Justice and Political Liberalism rather

10 Laying the ground


21/274

soon took up this suggestion with greater or lesser attention to historical claims

about Kant ’s own theory.5 Since that time, Kant commentators have almost

exclusively characterized the discussion of the question of moral realism as a

choice between the alternatives of moral realism and moral constructivism,

effectively lumping together all possible nonrealist interpretations of Kant asconstructivist. While doing so, these same commentators have generally

bemoaned the lack of specicity of the term “constructionism ” and have

attempted to dene or rene it using various taxonomies they present. Rarely

do these taxonomies match precisely; still more rarely do they cross-reference

or incorporate one another. The result is that much philosophically interesting

work on the issue of moral realism in Kant is scattered in individual articles in

isolation from one another.

Moral realism is one of those issues about which philosophers spend much

of their time simply dening the terms of the debate. There is no clear consensus on what “realism ” actually means, and many acknowledge that tha t

the meaning of “realism ” has shifted along with the philosophical tide.6

According to this last view, some theories which would have counted as

antirealist fty or hundred years ago would count as realist today. Even this

admission seems too optimistic since it assumes that there is a general consen-

sus at any given time.

A popular de nition

Two problems beset the task of dening moral realism. One is that various

metaethical theories differ in their interpretation of key terms such as “truth”

and “validity”, “objective” and “subjective”, “obligatory” and “permissible”.

Disagreement about these key terms allows various different theories to claim

to present moral realist positions although they vary widely with one another.

The debate over moral realism also suffers from the connotation of one of the

central terms of the debate: “antirealism ”. As an “anti”, the latter connotes that

one is opposed to some positive claim rather than that one is giving a positive

claim of one’s own. In ethics, this tendency is exaggerated by the moral import

of the terms involved. To deny moral realism seems to imply a lessening of the

claims or values of morality itself. For this reason, I and others prefer the more

than his particular interpretation of Kant, see (O’Neill 2002), (O’Neill 2003), (Bagnoli 2014),and (Kaufman 2012).

5Two of the earliest Kant commentators to employ the language and method of constructivism areThomas Hill (Hill 1989) and Onora O’Neill (O’Neill 1989). They were soon followed by themost inuential constructivist after Rawls, Christine Korsgaard (Korsgaard 1996b). The intense

debate about moral realism in Kant began in the wake of Korsgaard’s book.6 See, for example, (Railton 1996).

Moral realism and naturalism 11


22/274

positive-sounding term “idealism ” and still other theorists, “constructivism ”

over “antirealism ”.

These two problems are highlighted by a famous illustration of this bias

toward supporting moral realism. Geoffrey Sayre-McCord suggests that one

might want to be able to say that “the Nazis were really bad”

, but might feeluna ble to make such a strong claim without a theory of moral realism behind

one.7 Moral idealism is then saddled with the burden of being implicitly

equated with a rejection of any kind of validity to moral claims. An analogous

claim regarding objects would illustrate the fallacy in this attitude:

A traditional early modern realist about objects might be one who holds to

the existence of objects independent of human perception of them; a traditional

idealist about objects might hold that objects are nothing more than collections

of sense data. Someone who rejected idealism about objects on the grounds

that it really cannot be about objects would simply be begging the questionabout the denition of “object ”, as if the idealist were denying the existence of

objects entirely rather than identifying them as collections of sense data. In the

same way, someone who thinks that saying that the Nazis were really bad

requires moral realism is begging the question about the denition of “bad”.

A moral idealist can easily make essentially the same claim as “the Nazis were

really bad” when the claim is stated in some non-question begging sense such

as “bad without qualication”.

Sayre-McCord himself offers a denition of moral realism that has gained

traction. He claims that “realism involves embracing just two theses: (1) theclaims in question, when literally construed, are literally true or false (cogni-

tivism), and (2) some are literally true”.8 Many recent Kant commentators

writing on the issue of moral rea lism either use his denition or assess one

much like it, generally critically.9 According to his denition, a theory is not

realist if it either (a) denies cognitivism, in which case it is some kind of non-

cognitivist theory, or (b) denies that there are any moral truths, in which case it

is some kind of error theory.

As these commentators note, Sayre-McCord’s denition does not adequately

capture the crux of the debate regarding Kant ’s metaethics. Prima facie, Kant is

neither a noncognitivist nor an error theorist. Constructivists and other non-

realists allow that in Kant there are moral claims that bear truth or falsity.

Under this denition, they would count as realist theories. Sayre-McCord’s

7 (Sayre-McCord 1988b, p.1) credits the illustration to Philip Gasper. Sayre-McCord does not takethe illustration to be more than a caricature and subsequently discusses ways that nonrealism can

capture moral condemnation. It is still useful as a caricature.8 (Sayre-McCord 1988b, p. 5).9 Paul Formosa (Formosa 2013) and I (Rauscher 2002) cite Sayre-McCord’s denition directly.

Others who also note that this sort of denition does not really settle the issue of realism andidealism include Jochen Bojanowski (Bojanowski 2012), and Sharon Street (Street 2010).



23/274

denition casts too wide a net. It makes every single possible moral theory

that allows for moral truth into a realist theory. It co-opts the use of truth claims

in a moral theory so that any kind of nonrealist theory is seen to deny validity

to morality. Using the analogy with realist and idealist denitions of objects

again, both the traditional realist and idealist hold that there are truths about objects, yet they have quite different conceptions of what those objects are.

Similarly, a moral realist and a moral nonrealist can hold that there are moral

truths, yet they can have quite different conceptions of what constitutes the

bearer of those truths.

Roughly speaking, the kind of denition Sayre-McCord provides equates

“real” with “true” and places the fault line between moral realism and moral

idealism precisely atop the fault line between acceptance and denial of moral

truth. Kantians should be skeptical that this perfect correspondence does

justice to Kant ’s theory, which, incorporates transcendental idealism yet

af rms moral truth.

Sharon Street phrases this point with regard to constructivism in general:

“if we understand realism this way, then metaethical constructivism counts as

a brand of realism – as indeed do an extremely wide range of views, including

even a simple subjectivism according to which what ’s good for a person is

whatever that person thinks is good”.10 She is right to stress the overextension

of the term “realism ” to any moral theory that offers some criterion for truth. In

particular, if Kant has a theory of moral truth that can be understood as

subject-dependent, perhaps even only dependent upon a certain kind of activityof subjects, then it would be peculiar to count his theory as realist although it

would contain a criterion for judging certain moral claims as true or false. The

basis of the truth or falsity would seem to make a great deal of difference in our

view of whether a theory should count as real.

This kind of denition of realism centered on truth claims is more suited to

consequentialist than Kantian moral theories. Because consequentialists hold

to the priority of the good over the right, they tend to take moral facts primarily

as facts about some good to be attained rather than some formal laws of right;

for a consequentialist, the content of laws that express right is derived from theconception of the good to be attained. One can picture how accepting the truth

value of claims about the good can make one a realist in a more metaphysical

sense by noting that goods are taken to be independent of the belief that they

are good; those who deny this independence are expressivists and those who

deny that there are any goods are error theorists. Right, on the other hand, has a

more ambiguous connection to belief. It is possible to conceive of moral agents

who are so constituted that, in a roughly Kantian way, their practical cognitive

10 (Street 2010, p. 370).



24/274

faculty both produces the moral law as rational and produces belief in it. The

truth value of the moral facts, here the criterion for right, would not be

independent of rational belief in moral agents. This kind of Kantian view

would open up space between the denial of the truth of some moral facts

(expressivism and error theory) and the af rmation of moral facts on someindependent basis. Thus, consequentialists might be more inclined to be

satised with the kind of denition that Sayre-McCord offered while Kantians

ought to be skeptical of its value.

A better de nition

In order to move beyond this problem, a conception of the proper division

between realism and nonrealism more appropriate for Kantians should include

a metaphysical dimension, as some Kant interpreters recognize.11

I offer thefollowing denitions that are more appropriate for Kantian ethics:

Moral realism: The moral principles, properties, or objects of the world

are independent of the transcendental or empirical moral agent.

Moral nonrealism: The moral principles, properties, or objects of the

world a re dependent upon the transcendental or empirical moral

agent.12

11 Paul Formosa labels realism that excludes only noncognitivists and error theorists a “weak senseof moral realism ” that “amounts to nothing more than a claim about the truth of some moral judgments” in contrast to a “strong realism ” that holds that what makes the moral judgments trueor false is “an independent moral order ” (Formosa 2013, p. 172). The weak sense of moralrealism accords with the kind I have just criticized. Formosa is correct in noting the need for a stronger, more meaningful sense of moral realism. This “strong sense” requires ontologicalindependence. Richard Boyd offers a denition that identies a requirement for independencefrom “our moral opinions, theories, etc”. and also claims that ordinary moral reasoning is a reliable method for obtaining moral knowledge (Boyd 1988, p. 182). The dif culty with Boyd’sdenition is the vagueness included in his “etc.”. Allen Wood alludes to Boyd’s denition in hisrst book on Kant ’s ethics as “the most agreed-upon sense” of the term “moral realism ” (Wood1999, pp. 157, 374, n. 4) but claims (quite emphatically) in his second book on Kant ’s ethics that

he does not endorse Boyd’s own type of “Cornell Realism ” as the proper way to construerealism in conformity with that denition (Wood 2008, p. 295, n. 8).

12I offered a similar set of denitions in (Rauscher 2002, p. 482). There I used the term “moralidealism ” while here I use “moral nonrealism ”; the latter term broadens the denition’s scope toinclude anyone who rejects moral truth such as error theorists and expressivists. I will focus myargument on idealism as I proceed. I also used the term “moral characteristic” while here I am clearer in using “moral principle, property, or object ”, the original intent of the former term;there is no change in meaning accompanying this clarication in terminology. Finally,I formerly used the term “human mind” while here I say “transcendental or empirical moralagent ”; this change is intended to allow for the different levels of realism or nonrealism at theempirical and transcendental levels and reects a shift in my analysis over the past decade. I am very grateful to Darlei Dall’Agnol for his criticisms of my earlier approach and for many

discussions in which we have attempted to nd the best Kantian position on moral realism.Some of his concerns are in (Dall’Agnol 2012b) and are discussed in my reply (Rauscher 2012).



25/274

These denitions dene the difference as metaphysical since both the term

“moral principle, property, or object ” and the relation of “independence” are

metaphysical. Very roughly, they are intended to put the onus on the moral

nonrealist to prove that there are no essentially moral principles, properties, or

objects, or if one wants to phrase it differently, no moral truth that wouldcorrespond to any principle, property, or object, such that the principle,

property, or object would be a part of the cosmos were moral agents qua

moral subjects not to exist. By “nonrealism ” I mean all possible alternatives to

moral realism, including expressivism, error theory, the constructivist views

that are not realist, and idealism, to be explained later in this chapter.

A moral principle, property, or object is not the same as a principle,

property, or object required for morality. The difference between a moral

principle, property, or object and a principle, property, or object required for

morality is that the latter consists of principles, properties, or objects which arenot solely moral, the former of principles, properties, or objects which are

solely moral. Moral principles, properties, or objects are to be understood as

exclusively involving moral normativity or value. For example, the human

mind may be an object required for morality, but because it can play a role in

other areas not pertaining to moral normativity or value, such as theories of

qualia, it is not a moral object. An individual’s being good, however, is a moral

property, since this characterization can play no role except in situations

involving moral normativity or value or those derivative upon it such as

descriptions of morally good persons. For denitional purposes, moral proper-ties and objects are not limited to a particular metaphysics. As examples of

moral principles, properties, or objects, consider the following: Good, evil,

bad, rightness, wrongness, justice, value, moral law. As examples of prin-

ciples, properties, or objects required for morality, consider the following: The

existence of minds, pragmatic or technical means–ends principles, and so on.

Of course, particular moral theories might differ regarding which specic

principles, properties, or objects are solely moral rather than more general

principles, properties, or objects, and whether the solely moral principles,

properties, or objects are separable from the more general principles, proper-ties, or objects. In Kant, we might legitimately question whether freedom of

the will, agent causality, or the existence of God count as moral properties or

objects or as simply properties or objects required for morality. I will treat

them as moral properties or objects and include them in my arguments in this

book.

The term “independent ” invokes a metaphysical separation between the

moral agent qua agent and whatever moral principles, properties, or objects

are at issue. I do not intend to hold that the moral principles, properties, or

objects are to be independent of all moral agents in existence but rather independent of the individual moral agent or particular type of moral agent



26/274

(to be specied) as a moral subject . There is no quick route to idealism through

a claim that morality depends upon the existence of any moral beings at all.

Thus, if moral agents themselves were to have independent intrinsic value as

objects of consideration for a moral agent who is the subject facing a moral

decision, that value would be real because it would be independent of themoral agent qua subject . Similarly, if moral value were to reside in satisfaction

of desires for moral agents in general and that moral value had a justication

beyond simply moral agents considering satisfaction of their own desires to be

good, the value of desire satisfaction would be real. Beca use moral agents are

not passive, dependence on mental activity is included.13 This emphasis on

moral agents qua subjects rather than objects is extremely important for a

Kantian theory that places much weight on the practical point of view or the

structure of moral agency.

A divine command theory or any theory that grounded moral laws onsomething independent of moral agents would be realist on this account, as

would any theory that held that there are values independent of moral agents’

acts of valuation or capacity of valuation. A moral theory that depended solely

on contingent facts about particular persons, such as a theory referenced in the

quote from Street earlier that the good for a person is merely whatever that

person happens to think is good for her, would not count as realist. An error

theory would be nonrealist because it denies that there are any moral prin-

ciples, properties, or objects.

I use the phrase “transcendental or empirical moral agent ”

in order to invoketwo possible levels of assessment of moral agency. The transcendental moral

agent would be the subject of an assessment of the conditions for moral agency

as such. In Kant, such an assessment would be transcendental, in other kinds of

theories conceptual. The description of the transcendental moral agent would

be valid for all possible moral agents. The empirical moral agent would be

the subject of an assessment of the particular mental structure of human

and similar beings as moral agents. This approach, which looks not at a

13 Jochen Bojanowski cites my earlier denition but claims that his “moral idealism ”, which he

takes to be a third alternative between moral realism and moral antirealism in Kant, does not utilize the sense of dependence he sees in my denition”. The idealism I want to ascribe to Kant holds not that the good depends on the human mind, but that its existence depends on self-affection in human cognizers” (Bojanowski 2012, p. 4). I hope that my clarication hereadequately shows that, as I had always intended and as is clear from the position I detail inmy earlier paper, the kind of dependence on the human mind can include mental activity such asdecision-making, reection, and self-affection. Bojanowski hints in the claim quoted here that

he might mean that the meaning of the term ‘good’ is dependent upon something besides thehuman mind while the existence of good depends upon self-affection; because he does not utilize semantic terminology, this attribution is speculative. Later in his article, he holds that

practical cognition precedes normative facts, which implies that the meaning of the term “good”is also the result of practical activity (Bojanowski 2012, p. 13).



27/274

transcendental moral agent but at a humanlike agent in particular, is empiric-

ally informed, including facts about the existence or types of desires, the actual

capabilities of the agent such as a capacity for free choice amid various

possible determinants of action, the access the mind has to any moral facts,

and so on. This approach is both empirical and conceptual in Kant in theapplication of practical principles and ideas that are not derived empirically. It

is not the equivalent of anthropology. The description of the empirical moral

agent in Kant would be valid only for moral agents sharing the particular moral

characterization at issue. Of course, at the empirical level, moral agents may

instantiate the transcendental moral agent, but whether they do, what the latter

consists of, and precisely what the relationship between them is depends upon

the nature of Kant ’s transcendental/empirical distinction (to be discussed later

on) and the success of particular philosophical arguments.

The scope of the “empirical”

agent in my argument is essentially restrictedto “humanlike” and should be taken to refer to the relevant characterization of

the moral nature of human beings, not to similarity to human beings in morally

irrelevant ways. In Kant, a humanlike moral agent could be dened roughly as

a nitely rational being capable of a pursuit of happiness, conscious of the

categorical imperative, and capable of choosing between following that

imperative or not. From now on, I will use the terms “human moral agent ”

and “empirical moral agent ” to refer to all humanlike beings. I stress the

human moral agent in order to emphasize that dependence upon some particu-

lar type of existing moral agent is at issue. A comparison wit h more generalmental features is useful. Some features of the human mind14 such as the

ability to sense only a particular portion of the electromagnetic spectrum are

certainly contingent and not a necessary feature of universal sensibility; other

organisms can see ultraviolet light or respond to magnetic elds. Kant holds

that space and time must be considered dependent in some way on the human

mind and are not conditions for experience for all possible minds. Arguably, a

similar variation is possible in theory with regard to mental processes more

closely related to morality. Supposing that various other types of mental

organization were differently instantiated in different species – and here onemust imagine something like a mind without memories or one determined

immediately by emotion – the particular parts of that mental organization that

were unique to one species might, on some theories, ground moral truth for

14Human moral agents would have human minds. When I use the term “mind”, I do not intend toinvoke any Cartesian substance; rather, I intend to be agnostic about the true nature of minds. At the same time, I limit this to the human mind as we generally experience it and conceive of it. If

it were to turn out that the true nature of human minds were that they are parts of God ’s mind, a resulting morality could be realist. I take this possibility as incompatible with Kant ’s theory.



28/274

them.15 Were there some general theory that showed that only one type of

mental organization could possibly exist, or that morality could be grounded

only in some particular core of mental functions that all possible moral agents

must exhibit, then those moral principles, properties, or objects would be

characteristic of both the empirical human and the transcendental moral agent.To prove this identity is Kant ’s goal.

The aforementioned discussion makes clear an error in equating realism with

objectivity. Objectivity can coincide with nonrealism provided that the minds in

question are all structured in such a way that they necessarily produce the same

ideas. An analogy can be made with the question of color perception. Simply

because our color qualia may be ideal and phenomenal colors may not inhere in

objects independent of the human mind, does not mean that the color qualia we

perceive are not objective. There might be some necessary causal connection

between our physical sensations and our mental perceptions. We might have a particular mental faculty which provides for the objectivity. This objectivity is

nonetheless not universal for all possible sensible beings but restricted to beings

with the particular kind of physical and mental constitution we have. This

objectivity lacks a transcendental foundation. For morality, objectivity can be

preserved even when nonrealism is accepted, provided that the theory explains

how the structure of the human moral agent dictates that all humans share the

same moral properties and objects at the empirical level or how the structure of

15 As an example of this kind of variation, consider Sharon Street ’s ants (Street 2012, pp. 53–54).She supposes for the sake of argument that a species of intelligent, conscious ants could exist.Because of the complex genetic relationships among ants, all female ants share seventy-vepercent of their genetic code with one another while only one, the queen, is able to reproduce.There are relatively few male ants. Survival of the colony relies on individual ants sacricingthemselves to protect the queen to ensure the existence of future generations. Street imaginesthat a female worker ant would exhibit a value system to reect these facts, valuing the survivalof the queen above her own survival and not seeing herself as intrinsically valuable. Street offers this as a counterexample to the Korsgaardian/Kantian claim that all reective beings whoare capable of valuing anything must value themselves as ends in themselves. I take thisexample to illustrate my point that some kind of organism with a suf ciently different kindof nature could have a different kind of moral system. I would like to make two observations

about this thought-experiment. First, it is extremely unlikely that such intelligent ants, or similarly genetically related social beings with the same behavioral patterns, would evolve.Such complex intelligence in animals requires a great deal of investment in the rearing of offspring and a great deal of relative mass devoted to the brain. Given the costs of producingand maintaining a functioning intelligent adult, it would be nearly impossible for a species tomass-produce them to such an extent that they could be easily expendable in such menial waysas providing a live bridge for others of her kind. Still, as Street notes, the point is not theplausibility of such a species actually evolving but the fact that the conceivability of the speciesprovides a counterexample to Korsgaard’s argument that all creatures capable of valuing must value themselves the most. Second, it is interesting to note that human beings sometimes think the same way that Street ’s ants do. Soldiers or others who willingly sacrice themselves for their nation are a case in point. This puzzling kind of altruism, which cannot be explained either

as inclusive tness (near genetic relations such as offspring or cousins) or as reciprocal altruism appears to have no genetic basis. For an attempt to explain such behavior, see (Kitcher 1993).



29/274

the transcendental moral agent dictates that all possible moral agents share the

same moral properties and objects at the transcendental level.

Transcendental and empirical realism

The term “realism ” in Kant has a vital twofold signicance largely overlooked

in the realism debate: Kant ’s distinction between, on the one hand, transcen-

dental idealism or realism and on the other hand, empirical idealism or realism.

This distinction does not appear in the contemporary general debate about

realism in moral theory because it applies to Kantian but not consequentialist

theories. Most of contemporary Kantian ethics does not even employ Kant ’s

transcendental/empirical distinction, presumably taking it to be an ontological

claim about the real versus apparent nature of objects that is not relevant to

ethics beyond a discredited theory of free will. These Kantian moral theoristsappropriate Kant ’s discussion of a practical point of view without linking it to

transcendental idealism. Interpreters of Kant ’s own moral theory who operate

without the distinction have incomplete models of Kant ’s moral realism or

idealism.16

In the Critique of Pure Reason, Kant differentiates between the transcen-

dental and empirical senses of realism and idealism. The four resulting possi-

bilities – transcendental idealism, transcendental realism, empirical idealism,

and empirical realism – work roughly as follows in relation to space: Tran-

scendental idealism is the claim that all objects in space are mere appearancesand not things as they are in themselves because the space in which they exist

is only a form of our intuition. Transcendental idealism allows objects to be

empirically real, that is, directly knowable by the empirical subject in an

objective order in space. Transcendental realism holds that space is something

given in itself independent of human intuition and hence, that things as they

are in themselves are spatial. But this transcendental realism leads to empirical

idealism, that is, that empirical objects in space are not directly accessible to

the empirical mind because space exists independently of human intuition

(B69–

71, A369–

70). Kant, of course, is a transcendental idealist with regardto time as well as space. The transcendental ideality of space and time means

that objects as they are in themselves are not spatiotemporal and cannot be

experienced by human beings, who are limited to experiences provided in

intuition. Human beings are able to know only appearances, that is, empirically

real objects. But the possibility that other objects exist that are not spatiotem-

poral remains.

16 This criticism holds for my earlier work as well (Rauscher 2002), but I do approach the

transcendental/empirical distinction without a comprehensive treatment in (Rauscher 2006b)and with a better but still incomplete treatment in (Rauscher 2006a ).



30/274

Kant ’s argument for the transcendental ideality of space is given through a

transcendental exposition of the concept of space as a basis for the a priori

truths of geometry (B40–41). To be able to function as the basis for a priori

truth in geometry, space as an outer intuition “has its seat merely in the

subject ”

. Similar reasoning holds for the transcendental deduction of the pureconcepts of the understanding: They are held to have their transcendental

ground as necessary conditions for the synthetic unity of apperception

(B159–60). Transcendental idealism thus takes these crucial formal features

of experience to depend upon the subject in a transcendental but not empirical

sense.

The transcendental/empirical distinction is relevant for morality because we

can take certain aspects of experience like objects in space to be dependent

upon the transcendental subject although they are in some sense independent

of the empirical subject. There is use for this distinction in practical philoso-phy. Just as objects in space are really independent of the subject in an

empirical sense but dependent on the subject in a transcendental sense, there

might be moral principles, properties, or objects that are really independent of

the moral agent in an empirical sense but dependent on the moral agent in a

transcendental sense, making Kant an empirical moral realist but a transcen-

dental moral idealist. Or some moral principles, properties, or objects might be

transcendentally real, that is, entirely independent of the transcendent al moral

agent, in which case, Kant would be a transcendental moral realist.17

For this overall transcendental/empirical distinction to work in practicalphilosophy, the concept of empirical experience, or the everyday experience

of human beings, must be understood to have a practical dimension in addition

to the theoretical dimension provided in the analytic of the rst Critique. Given

Kant ’s interest in anthropology as a natural level at which we can understand

human beings as moral beings, a practical dimension of experience in Kant is

at hand. Kant discusses our consciousness of ourselves as moral in relation to

our moral predisposition in the Anthropology (7:324). Even in the less obvi-

ously empirical works, Kant discusses everyday moral experience. He begins

the Groundwork by invoking common moral cognition or beliefs actually held

17One might be further tempted to conclude that a transcendental moral realist would have to bean empirical moral idealist, in which case, some moral principles, properties, or objects wouldbe not directly known in experience. As with Kant ’s claim that with regard to objects in space, a transcendental realist ends up as an empirical “skeptical” idealist, this translates into anempirical moral skepticism. Suppose that the moral value of a rational agency were understoodto be transcendentally real, that is, independent of the transcendental moral agent qua agent. Theempirical moral agent might have no transcendental justication for a recognition of this moralvalue, since it is independent of that moral agent considered transcendentally. Hence, on theempirical level, an empirical moral agent might need another argument that would show the

actuality of those moral values. For an argument along these lines, see (Sensen 2011,pp. 18–20).



31/274

by people (4:393); in the second Critique, he takes his project not be a

replacement of an old principle of morality with a new one but a philosophical

justication of it through a new formulation (5:8, ftn.). In other words, he is not

denying that the principle of morality is part of the actual experience of human

beings. Moral duties and values are experienced by human beings in nature,that is, human beings as appearances. Given the possibility of empirical moral

experience, the controversial interpretive issue relevant to current purposes is

whether morality would be empirically real or empirically ideal, that is,

independent of or dependent on the empirical moral agent.

In order for this empirical/transcendental distinction to apply to ethics, we

have to be able to identify transcendental arguments for the necessary condi-

tions for a moral agent. Kant attempts to provide a deduction of the moral law

in Groundwork III and later discusses a deduction in the Critique of Practical

Reason. Although Kant does not identify them as such, these deductions havebeen treated by commentators as transcendental.18 If it is proper to consider

them as transcendental deductions, then they might be understood as the basis

for claims that the moral properties and objects are dependent on the transcen-

dental moral agent. Further, one might be tempted to treat Kant ’s derivation of

the value of humanity as a k ind of transcendental deduction (as Korsgaard does

without using that term 19), even though Kant never calls it a deduction.

Understanding these issues requires a look at the nature of Kant ’s philosophy

in general and practical philosophy in particular. Chapter 2 will examine these

issues in detail.The importance of this discussion of the empirical/transcendental distinction

is that there can be two levels of possible moral realism or idealism. At the

transcendental level, one might ask whether moral principles, properties, or

objects are dependent or independent of some possible transcendentally iden-

tied moral agent, that is, an idealist might argue that autonomy of the will as

pure practical reason might be a transcendentally ideal ground for the categor-

ical imperative, or a realist might argue that humanity as an end in itself should

be understood as a transcendentally real value independent of the transcenden-

tal moral agent. There is a third option besides moral transcendental realism and idealism: There might be no transcendental justication for morality at all.

If human moral experience is dependent upon some contingent features of the

human moral agent and cannot nd any transcendental basis, Kant ’s ethics

would be empirically ideal but would have no transcendental status, in the

18 See, for example, the articles by John Rawls, Henry Allison, and Barbara Herman in (Förster 1989).

19 See (Korsgaard 1996a , p. 123). Recently, Robert Stern has endorsed a version of her transcen-dental argument (Stern 2011a ).



32/274

same way that color sensations are empirically ideal and have no transcenden-

tal status (B44–45).

At the empirical level, an empirical realist might argue that the transcenden-

tally ideal ground for the law might be compatible with the empirical reality of

duties such that those duties are not dependent upon the empirical moral agent but only the transcendental moral agent. This distinction might prove useful in

dissolving some strong disagreements about whether the categorical impera-

tive is independent of rational moral agents. The effect of the transcendental/

empirical distinction on the realism/nonrealism debate can come only after a

look at the nature of philosophy in Kant in Chapter 2.

Allow me to briey note the importance of the distinctions made so far.

A standard denition of moral realism has been shown to be inadequate in

capturing the elements peculiar to Kantian ethics. A denition that stresses

the role of the metaphysical independence of moral principles, properties, or objects from either the transcendental or empirical moral agent serves better.

Kant ’s moral theory – the target of discussion in this book – will be

subjected to a more appropriate denition that illuminates the genuine fault

line between Kantian realism and Kantian nonrealism in terms of independ-

ence or dependence on the moral agent. The independence or dependence

appears at two possible levels when Kant ’s transcendental/empirical distinc-

tion is applied. If there is a transcendental dependence of all moral prin-

ciples, properties, and objects upon the transcendental moral agent, Kant

would be a transcendental moral idealist; if the transcendental justication of morality requires that some moral principles, properties, or objects be inde-

pendent of the transcendental moral agent, then Kant would be a transcen-

dental moral realist. A similar distinction would operate at the empirical level

of everyday moral experience with reference to the empirical moral agent.

The conception of the human moral agent and the status of Kant ’s justica-

tory arguments for morality will have to be determined in detail in order to

resolve this dispute.

This fourfold division crossing empirical/transcendental and real/ideal is

still too stark a choice for interpreting Kant. There can be interpretations that hold Kant to be a transcendental realist regarding some moral principles,

properties, or objects and a transcendental idealist regarding others. No one

takes Kant to be a realist about every moral principle, property, or object and

few take him to be a complete idealist. Philosophically, it is more fruitful to

avoid slapping a blanket identication on Kant as one or the other and instead

to assess all aspects of Kant ’s moral theory in light of the issue. I will devote a

later section of this chapter to listing the various kinds of moral principles,

properties, or objects that are raised in the literature. Before that, however,

I will evaluate various conceptions of constructivism to show how they t intothe realism/nonrealism debate I have just characterized.



33/274

Comparison with constructivism

No single denition of “constructivism ” garners universal or even wide

acceptance. Kantian moral realists tend to tr eat all opponents as constructivists,

but some identify themselves otherwise.20

The positions are further compli-cated because some constructivists label themselves moral realists while others

seem to reject realism.21 And “constructivism ” itself is an unfortunate term to

use when discussing Kant ’s philosophy because he had a particular meaning

for the term in his theoretical philosophy that conicts with its use by contem-

porary Kantians in ethical theory.22 In this section, I will rst review two

possible general types of construction, showing which ought to be taken as

20 Robert Stern labels the debate about Kant ’s moral realism “the constructivist/realist contro-

versy” (Stern 2012, p. 119). Paul Formosa ’s very title makes the assumption: “Is Kant a MoralConstructivist or Moral Realist?” (Formosa 2013). Although I have described my position asidealist and never as constructivist, I have been lumped with others as a constructivist severaltimes (Kain 2006, Stern 2011b), (but see Dall’Agnol 2012b for a notable exception). KarlAmeriks, in contrast, is to be praised for his argument against two “nonrealist ” interpretations of Kant, those of J.B. Schneewind and Charles Larmore, in which he does not see them primarilyas constructivist (Ameriks 2003).

21 In part, this is due to their differing conceptions of moral realism. The most famous of the self-proclaimed constructivist realists is Christine Korsgaard, who distinguishes Kant ’s “proceduralmoral realism ” from metaphysical realism, which she calls “substantive moral realism ”(Korsgaard 1996b, p.35) or “objective realism ” (Korsgaard 1996a , pp. 278–282). She claimsthat “procedural moral realism is the view that there are answers to moral questions; that is, that

there are right and wrong ways to answer them ”. Substantive (or objective) moral realism further holds that there are moral facts or truths which account for the right answer to thosequestions. This claim implies that any method of answering a question about right and wrongcounts as a procedure. She seems to conrm this a few pages later when she claims that substantive realism, which holds that moral facts are true in virtue of something independent of the moral agent, is a version of procedural realism. This understanding of procedure would beitself too broad. The universality of the term “procedure” when taken to include any method of answering questions makes it useless when trying to nd something distinctive about proce-duralism. Her description would encompass a procedure of logical deduction from a set of statements, which is of course one way to try to nd right or wrong answers. She later characterizes constructivism as a form of problem solving, making it appear that all attemptsto nd answers to practical problems are constructivist, although without invoking the term

“procedure” (Korsgaard 2003). In these statements, Korsgaard effectively erases any meaning-ful distinction between nding an answer and creating an answer. Any attempt to answer moralquestions by reference to independently existing moral facts would still count as constructivist.Hence, Korsgaard is not a good source for an account of the distinctiveness of a procedure for construction.

22 In the Critique of Pure Reason, he distinguishes philosophy from mathematics when noting that concepts are constructed in mathematics but not constructed in philosophy (A713–16/ B741–44). To construct a concept is to “exhibit a priori the intuition corresponding to it ”. Onlymathematics can construct concepts because philosophical concepts cannot be represented in

intuition. A mathematically constructed concept is a specication of the universal, for example,the properties of triangles in general, in a particular, say, a specic triangle drawn on paper or inthe imagination, because the concept already contains a pure intuition that needs to be exhibited.

Philosophical concepts, in contrast, are not constructed because they do not contain any pureintuition but only “nothing but the synthesis of possible intuitions which are not given a priori”



34/274

paradigm, before showing how my approach is situated in relation to con-

structivist theories.

At its most general level, a constructivist theory claims that some particular

moral principles, properties, or objects such as norms, values, or moral facts,

are derived from some other particular principles, properties, or objects,whether purely moral, for example, a highest value, or not, say, rationality

itself, through a procedure of some sort.23 Rawls initially used the term

“procedure” in A Theory of Justice as a way to understand the process by

which individuals would determine principles of justice by means of the

original position.24 Constructivists who had studied with Rawls adopted the

term, and even Onora O’Neill, who sharply differentiates her constructivism

from Rawls’, sometimes characterizes her position this way.25 Yet the specic

procedures they invoke are not of the same type. I see two general types of

constructivism in the literature. I will suggest that the one that is closest to theapproach I have to moral idealism ought not to be considered to construct

anything.

Thomas Hill offers a Rawlsian understanding of the nature of a procedure in

what he calls a “procedure of construction”. In this type, moral principles (for

Hill, the target of construction) are valid “ just in case and because they would

be endorsed by all members of an appropriately dened initial choice

and hence can relate to objects in intuition only synthetically and discursively (A719–20/

B747–48). In the Critique of Practical Reason, Kant makes an analogous claim with regardto the practical in a discussion of how to separate the doctrine of happiness from the doctrine of morals. He suggests that making this distinction requires the care and precision of a geometer.

“A philosopher, however, has greater dif culties to contend with here (as always in rationalcognition through mere concepts without construction of them) because he cannot put anyintuition (a pure noumenon) as its basis” (5:92). Presumably, Kant is making an analogybetween the pure intuition used in geometry with any pure conceptual space (noumena) that would be used in practical philosophy. Practical philosophy cannot construct its concepts in a

pure conceptual space but must show how they affect actual human beings in their actualdecisions. To resolve this, Kant suggests that practical philosophy can use a method similar to a chemist who adds alkali to a certain solution in order to separate the acid from the remainder of the solution. The experiment that Kant suggests is that in a particular person awareness of the

moral law will similarly separate considerations of advantage from rational considerations in theperson’s mind, allowing for a focus on moral motivation. For a thorough examination of thisparticular analogy, see (Rohden 2012). Like other claims stemming from reason, for example,the concept of a soul, the moral law is not constructed in Kant ’s sense of the term.

23 Sharon Street believes that inclusion of a procedure is not a necessary element of constructivist views (Street 2010, p. 366). Constructivism is better characterized as a metaethical theory that stresses the practical point of view. Ultimately, I agree that theories that stress the practical point of view are superior theories, but I hesitate to label them as “constructivist ” and instead use theterm “idealist ”. Street denes a constructivist position in terms of the justication of some moral judgments by reference to another set of moral judgments (Street 2008, p. 208).

24 (Rawls 1999a , p. 17).25 O’Neill does not generally describe her position as proceduralist but has used the term (O’Neill

1989, p. 216, n. 10). Korsgaard uses the term quite prominently when she characterizes her position as “procedural realism ” (Korsgaard 1996b, p. 35).



35/274

situation”.26 I will call this the “decision procedure” type of constructivism.

The crucial element of this denition is the requirement that a decision in an

initial situation, one that provides some material the agents may use while

deciding, is used to determine the result. This kind of procedure could produce

various kinds of moral principles, properties, or objects, not merely principles,from some other set of moral or nonmoral principles, properties, or objects.

O’Neill’s discussion of the construction of principles by agents seeking agree-

ment is another example.27 An exemplary mainstream denition that stresses

the role of choice stemming from an idealized process of deliberation is given

by Carla Bagnoli.

As a “metaethical account ” – an account of whether there are any normative truths and,

if so, what they are like – constructivism holds that there are normative truths. These

truths are not xed by facts that are independent of the practical standpoint, however

characterized; rather, they are constituted by what agents would agree to under somespecied conditions of choice.

28

The crucial characteristic of the decision procedure is that it entails some

ability to m ak e a decision, whether real or hypothetical, on the part of the

participants.29 A procedure of this kind differs from a procedure in which the

creation of the moral principles (or norms or values) conceptually precedes

any capability for agents to make decisions. In other words, this type of model

of a construction procedure requires that the agents in the hypothetical or real

situation can be conceived as capable of making a decision to endorse or reject

the outcomes of the procedure. Neither agency itself nor any prerequisites for

agency can be constructed through a decision procedure because one needs

agents in order to make decisions in the rst place.

26 (Hill 2012, p. 78). 27 (O’Neill 1989) and (O’Neill 2002). 28 (Bagnoli 2011).

29 Paul Formosa calls actual constructivism “all the way down” because it involves actual willing

by individuals either collectively as a culture or individually. Hypothetical constructivism is

“not all the way down” because both the content and the authority of the procedure itself are

“laid out ” rather than the subject of some actual choice (Formosa 2013, pp. 173–74). Formosa does not provide any example of an “all the way down” constructivist but suggests that a model

would be someone who takes moral norms to be embodied in cultural practices but nonethelesssubject to the “individual or collective” act of will. This characterization is problematic because

if the norms are embodied in the cultural practices, they are not the result of any direct acts of will at all. If he means that the acts of will are those of an agent independently endorsing thecultural norms, then his characterization would not capture the spirit of constructivism inparticular but would be applicable to any acts of an agent who faces the question of whether to subject himself or herself to moral norms. Formosa does not discuss actual procedures in theremainder of his article but defends a “not all the way down” reading of Kant.An example of an “all the way down” constructivism might be what Andrews Reath calls the

“Principle of Individual Sovereignty” in which each particular will is subject only to laws it actually legislates for itself. Reath and Patrick Kain cite problems with this view on its owngrounds and as an interpretation of Kant and the two philosophers to whom they attribute this

view, Robert Paul Wolff and Rüdiger Bittner, do not themselves take it to be Kant ’s view. See(Reath 2006, pp. 97–98), (Kain 2004, pp. 262–65), (Wolff 1974), and (Bittner 1989).



36/274

A kind of constructivism that alleges to construct moral principles, proper-

ties, or objects on the basis of agency itself is thus the second kind of

constructivism in the literature. What I will call the “nature of agency proced-

ure” allows for the generation of norms, laws, or values by means of the

expression of the nature of moral agents conceptually prior to their ability tochoose anything. This procedure does not involve conscious endorsement,

whether real or hypothetical, by agents but is simply the product or character-

istics of a particular kind of active being. Korsgaard’s attempt to ground the

categorical imperative on rational beings’ very ability to make reective

decisions is one example of this kind of procedure. Although she has claimed

to construct agency itself, the actual procedure she offers is one in which she

takes rational agency as the starting point from which she then draws out the

categorical imperative and moral value.30 The main question asked is about

which moral norms are intrinsically tied to the nature of free agency itself.Certainly, the agents might later reect on their situation and endorse the

results of the activity, but this reective endorsement is not the source of the

construction.

30 Korsgaard’s position is more ambiguous than stated here. She sometimes describes her positionas one of a constitution of agency, but when giving details, she shifts to a constitution of agents

as particular individuals who use their ability as agents to make choices that dene who they are(Korsgaard 2008, p. 109). It seems clear that she does not construct agency as such but usesagency to construct other elements of her moral theory.

She also sometimes argues that the categorical imperative is a result of a decision by an agent tobe a certain kind of person, for example, one committed to the kingdom of ends or to egoism (Korsgaard 1996b, p. 101). At other times, she insists that acceptance of the categoricalimperative is a requirement stemming from the “reective structure of human consciousness”and thus not a result of a decision by such an agent (Korsgaard 1996b, pp. 103–04). The former approach is what I call “decision procedure” and the latter “nature of agency procedure”.Some opponents of constructivism claim that Korsgaard’s constructivism is not only not a genuine alternative to moral realism, it is also not even a metaethical claim. Nadeem J.Z.Hussain and Nishi Shah argue that Korsgaard’s constructivism makes no metaethical claims but is better described as making claims in normative ethics or moral psychology (Hussain andShah 2006). Allen Wood hints that “no distinctive metaethics or metaphysics of value” isimplied by the procedural account of practical reason but does not offer details (Wood 2008,

pp. 282–83, n.3). I nd these charges dif cult to accept. Korsgaard and other constructivistsmight not have offered a clear and internally consistent metaethics, but they have made somemetaethical claims. The constr

Documents

Naturalism and Realism in Kants Ethics