Professional Documents
Culture Documents
Contents
Acknowledgments
Citations of Kants writings
page vi
viii
Introduction
Part I
Part II
Part III
50
79
81
107
147
149
180
209
241
Works cited
Index
250
258
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 commented on this book as it was being drafted. Darlei DallAgnol pushed and
prodded me into better explaining and defending my views both in conversations and through his published criticisms. I greatly beneted from conversations 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 arguments. 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 Intermural 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 opportunities 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 Pontical Catholic University of Parana, the State University of Campinas, and
the University of So Paulo in Brazil, Pisa University in Italy, the Philipps
University in Marburg and the University of Siegen in Germany, the University 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
Acknowledgments
vii
Quotations from Kant use the Cambridge Edition of the Works of Immanuel
Kant when available. I have modied the translation in some cases. When the
material is not included in the Cambridge Edition, the translations are my own.
All references will be given parenthetically in the text. References to Kants
writings except the Critique of Pure Reason will be to their volume and page
number in Kants gesammelte Schriften, 29 Vols. (Berlin: Walter de Gruyter
[and predecessors], 1902). References to Critique of Pure Reason are taken
from the edition Kritik der reinen Vernunft edited by Jens Timmermann
(Hamburg: Felix Meiner, 1998) and identied by page numbers for the rst
and/or second editions, cited as A and B. These page numbers are given in
the margins of the Cambridge translation.
viii
Introduction
Introduction
Introduction
Introduction
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 Kants ethics represent apparently
non-natural entities or properties and cannot be understood in the same metaphysically naturalistic way that reason itself can, that is, they cannot be an actual
part of nature operating as the empirically real manifestation of the transcendental 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 Kants 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 Kants 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 than
transcendent 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 difcult 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 ones 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 the
possibility of any non-natural, intrinsic value property and instead show that
Introduction
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 Kants various
discussions of this direct application of the moral law to experience as a moral
world. 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
reasons own structure that requires both systematic connection and unconditional 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 Kants 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/constructivist 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 appropriate
conclusion 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 Kants writings that are inconsistent 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 terminology 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 Kants
ethics in light of twenty-rst century concerns about naturalism and realism.
My work is reconstructive in that I pursue the philosophical implications of
Kants positions to sometimes make connections that Kant himself does not
explicitly make. I believe that all of my claims are consistent with Kants
general philosophical aims, methods, arguments, and conclusions, and nearly
always with his particular stated positions. Given the novelty of Kants
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 philosophers work in ways relevant to the present, no comprehensive interpretation can offer more than that.
Part I
10
2
3
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 when
doing so is useful for understanding the particular debate within Kant circles.
The Dewey Lectures were published in (Rawls 1980) and reprinted in (Rawls 1999b,
pp. 30358).
Larry Krasnoff traces the rst use of the term constructive in moral theory to a review by
Ronald Dworkin in 1973 (reprinted in [Dworkin 1978, pp. 15083]) 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).
Since constructivism was introduced by Rawls, one might think that it can be applied in political
philosophy 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, which
draws more directly on Rawls own work in A Theory of Justice and Political Liberalism rather
11
soon took up this suggestion with greater or lesser attention to historical claims
about Kants 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 as
constructivist. 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 that
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 consensus at any given time.
A popular denition
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 ones 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 (ONeill 2002), (ONeill 2003), (Bagnoli 2014),
and (Kaufman 2012).
Two of the earliest Kant commentators to employ the language and method of constructivism are
Thomas Hill (Hill 1989) and Onora ONeill (ONeill 1989). They were soon followed by the
most inuential constructivist after Rawls, Christine Korsgaard (Korsgaard 1996b). The intense
debate about moral realism in Kant began in the wake of Korsgaards book.
See, for example, (Railton 1996).
12
8
9
(Sayre-McCord 1988b, p.1) credits the illustration to Philip Gasper. Sayre-McCord does not take
the illustration to be more than a caricature and subsequently discusses ways that nonrealism can
capture moral condemnation. It is still useful as a caricature.
(Sayre-McCord 1988b, p. 5).
Paul Formosa (Formosa 2013) and I (Rauscher 2002) cite Sayre-McCords denition directly.
Others who also note that this sort of denition does not really settle the issue of realism and
idealism include Jochen Bojanowski (Bojanowski 2012), and Sharon Street (Street 2010).
13
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 Kants theory, which, incorporates transcendental idealism yet
afrms 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 whats 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 activity
of 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 the
conception 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
14
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 afrmation of moral facts on some
independent 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 denition
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 the
following 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 are dependent upon the transcendental or empirical moral
agent.12
11
12
Paul Formosa labels realism that excludes only noncognitivists and error theorists a weak sense
of 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 true
or false is an independent moral order (Formosa 2013, p. 172). The weak sense of moral
realism 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 ontological
independence. Richard Boyd offers a denition that identies a requirement for independence
from 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 difculty with Boyds
denition is the vagueness included in his etc.. Allen Wood alludes to Boyds denition in his
rst book on Kants ethics as the most agreed-upon sense of the term moral realism (Wood
1999, pp. 157, 374, n. 4) but claims (quite emphatically) in his second book on Kants ethics that
he does not endorse Boyds own type of Cornell Realism as the proper way to construe
realism in conformity with that denition (Wood 2008, p. 295, n. 8).
I offered a similar set of denitions in (Rauscher 2002, p. 482). There I used the term moral
idealism while here I use moral nonrealism; the latter term broadens the denitions scope to
include anyone who rejects moral truth such as error theorists and expressivists. I will focus my
argument 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 moral
agent; this change is intended to allow for the different levels of realism or nonrealism at the
empirical and transcendental levels and reects a shift in my analysis over the past decade. I am
very grateful to Darlei DallAgnol 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 (DallAgnol 2012b) and are discussed in my reply (Rauscher 2012).
15
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 would
correspond 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 are
not 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 individuals 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 properties 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 principles, properties, or objects required for morality, consider the following: The
existence of minds, pragmatic or technical meansends 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, properties, 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
16
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 selfaffection in human cognizers (Bojanowski 2012, p. 4). I hope that my clarication here
adequately shows that, as I had always intended and as is clear from the position I detail in
my earlier paper, the kind of dependence on the human mind can include mental activity such as
decision-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 the
human 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).
17
transcendental moral agent but at a humanlike agent in particular, is empirically 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 the
application 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 Kants 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 restricted
to 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 particular type of existing moral agent is at issue. A comparison with more general
mental 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 one
must 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
14
Human moral agents would have human minds. When I use the term mind, I do not intend to
invoke 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 Gods mind, a
resulting morality could be realist. I take this possibility as incompatible with Kants theory.
18
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 Kants 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 Streets ants (Street 2012, pp. 5354).
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-ve
percent 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 sacricing
themselves to protect the queen to ensure the existence of future generations. Street imagines
that a female worker ant would exhibit a value system to reect these facts, valuing the survival
of 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 who
are capable of valuing anything must value themselves as ends in themselves. I take this
example to illustrate my point that some kind of organism with a sufciently different kind
of 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 producing
and maintaining a functioning intelligent adult, it would be nearly impossible for a species to
mass-produce them to such an extent that they could be easily expendable in such menial ways
as providing a live bridge for others of her kind. Still, as Street notes, the point is not the
plausibility of such a species actually evolving but the fact that the conceivability of the species
provides a counterexample to Korsgaards 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 Streets 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).
19
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: Kants distinction between, on the one hand, transcendental 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 Kants
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 theorists
appropriate Kants discussion of a practical point of view without linking it to
transcendental idealism. Interpreters of Kants own moral theory who operate
without the distinction have incomplete models of Kants moral realism or
idealism.16
In the Critique of Pure Reason, Kant differentiates between the transcendental and empirical senses of realism and idealism. The four resulting possibilities transcendental idealism, transcendental realism, empirical idealism,
and empirical realism work roughly as follows in relation to space: Transcendental idealism is the claim that all objects in space are mere appearances
and 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
(B6971, A36970). Kant, of course, is a transcendental idealist with regard
to 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 spatiotemporal 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).
20
One might be further tempted to conclude that a transcendental moral realist would have to be
an empirical moral idealist, in which case, some moral principles, properties, or objects would
be not directly known in experience. As with Kants claim that with regard to objects in space, a
transcendental realist ends up as an empirical skeptical idealist, this translates into an
empirical moral skepticism. Suppose that the moral value of a rational agency were understood
to be transcendentally real, that is, independent of the transcendental moral agent qua agent. The
empirical moral agent might have no transcendental justication for a recognition of this moral
value, since it is independent of that moral agent considered transcendentally. Hence, on the
empirical 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. 1820).
21
18
19
See, for example, the articles by John Rawls, Henry Allison, and Barbara Herman in (Frster
1989).
See (Korsgaard 1996a, p. 123). Recently, Robert Stern has endorsed a version of her transcendental argument (Stern 2011a).
22
same way that color sensations are empirically ideal and have no transcendental status (B4445).
At the empirical level, an empirical realist might argue that the transcendentally 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 imperative 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.
Kants 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 independence or dependence on the moral agent. The independence or dependence
appears at two possible levels when Kants transcendental/empirical distinction is applied. If there is a transcendental dependence of all moral principles, 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 independent of the transcendental moral agent, then Kant would be a transcendental 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 Kants justicatory 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 Kants 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 into
the realism/nonrealism debate I have just characterized.
23
21
22
Robert Stern labels the debate about Kants moral realism the constructivist/realist controversy (Stern 2012, p. 119). Paul Formosas very title makes the assumption: Is Kant a Moral
Constructivist or Moral Realist? (Formosa 2013). Although I have described my position as
idealist and never as constructivist, I have been lumped with others as a constructivist several
times (Kain 2006, Stern 2011b), (but see DallAgnol 2012b for a notable exception). Karl
Ameriks, 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 primarily
as constructivist (Ameriks 2003).
In part, this is due to their differing conceptions of moral realism. The most famous of the selfproclaimed constructivist realists is Christine Korsgaard, who distinguishes Kants procedural
moral realism from metaphysical realism, which she calls substantive moral realism
(Korsgaard 1996b, p.35) or objective realism (Korsgaard 1996a, pp. 278282). She claims
that 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 those
questions. This claim implies that any method of answering a question about right and wrong
counts 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 be
itself 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 proceduralism. 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 attempts
to nd answers to practical problems are constructivist, although without invoking the term
procedure (Korsgaard 2003). In these statements, Korsgaard effectively erases any meaningful distinction between nding an answer and creating an answer. Any attempt to answer moral
questions 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.
In the Critique of Pure Reason, he distinguishes philosophy from mathematics when noting that
concepts are constructed in mathematics but not constructed in philosophy (A71316/
B74144). To construct a concept is to exhibit a priori the intuition corresponding to it. Only
mathematics 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 in
the 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 pure
intuition but only nothing but the synthesis of possible intuitions which are not given a priori
24
23
24
25
and hence can relate to objects in intuition only synthetically and discursively (A71920/
B74748). In the Critique of Practical Reason, Kant makes an analogous claim with regard
to 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 difculties to contend with here (as always in rational
cognition through mere concepts without construction of them) because he cannot put any
intuition (a pure noumenon) as its basis (5:92). Presumably, Kant is making an analogy
between 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 actual
decisions. 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 the
persons mind, allowing for a focus on moral motivation. For a thorough examination of this
particular analogy, see (Rohden 2012). Like other claims stemming from reason, for example,
the concept of a soul, the moral law is not constructed in Kants sense of the term.
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 the
term 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).
(Rawls 1999a, p. 17).
ONeill does not generally describe her position as proceduralist but has used the term (ONeill
1989, p. 216, n. 10). Korsgaard uses the term quite prominently when she characterizes her
position as procedural realism (Korsgaard 1996b, p. 35).
25
27
28
(Hill 2012, p. 78).
(ONeill 1989) and (ONeill 2002).
(Bagnoli 2011).
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. 17374). 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 nonetheless
subject 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 the
cultural norms, then his characterization would not capture the spirit of constructivism in
particular 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 the
remainder 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 own
grounds and as an interpretation of Kant and the two philosophers to whom they attribute this
view, Robert Paul Wolff and Rdiger Bittner, do not themselves take it to be Kants view. See
(Reath 2006, pp. 9798), (Kain 2004, pp. 26265), (Wolff 1974), and (Bittner 1989).
26
A kind of constructivism that alleges to construct moral principles, properties, 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 procedure 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 to
choose anything. This procedure does not involve conscious endorsement,
whether real or hypothetical, by agents but is simply the product or characteristics of a particular kind of active being. Korsgaards 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
Korsgaards position is more ambiguous than stated here. She sometimes describes her position
as 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 uses
agency 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 to
be 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 categorical
imperative 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. 10304). The former
approach is what I call decision procedure and the latter nature of agency procedure.
Some opponents of constructivism claim that Korsgaards 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 Korsgaards constructivism makes no metaethical claims but
is better described as making claims in normative ethics or moral psychology (Hussain and
Shah 2006). Allen Wood hints that no distinctive metaethics or metaphysics of value is
implied by the procedural account of practical reason but does not offer details (Wood 2008,
pp. 28283, n.3). I nd these charges difcult to accept. Korsgaard and other constructivists
might not have offered a clear and internally consistent metaethics, but they have made some
metaethical claims. The constructivist insistence, for example, that value is itself not a real
property of objects or persons but results merely from a particular practice guided by reason is a
claim about the nature of value itself rather than about what particular things are of value, the
former being a metaethical and the latter, a normative claim. The difference between holding
that value is a real but nonnatural property of objects and holding that value is merely a
projection onto objects is metaethical since both can identify precisely the same objects as
being of value. And if expressivist theories that tie moral claims to emotions operate at a
metaethical level, why should not a theory that ties ethical claims to a particular conception of
moral agency or to the determinations of a particular mental faculty? Not only can the
metaphysical question differentiate constructivism from other metaethical theories, some have
also claimed that constructivism makes distinctive epistemological metaethical claims (Bagnoli
2012).
27
Kant rarely uses the terms naturalism or naturalist at all, and even less in ways which would
dene a philosophical program related to contemporary understandings of the term. The few
occurrences fall into three types, which correspond to metaphysical naturalism, methodological
naturalism in knowledge claims, and method in theology.
a.
The rst type metaphysical use of the term naturalism occurs most clearly in the Prolegomena in a discussion of the cosmological ideas: The cosmological ideas, through the manifest
inadequacy of all possible cognition of nature to satisfy reason in its rightful demands, serve to
deter us from naturalism, which would have it that nature is sufcient unto itself (4:363).The
idea of a naturalist as taking nature to be sufcient unto itself provides a picture which accepts
28
might seem to be anachronistic because of this lack of attention. But this issue
is important in that it highlights the lack of metaphysical commitments in
Kants moral theory. By framing my interpretation in naturalistic terms, I am
able to highlight what I take to be a central characteristic of Kants metaethics,
namely, the importance of the practical standpoint as an action-oriented facet
of experience that is largely indifferent to theoretical claims about ontology.
Kant is generally taken to require some strong metaphysical commitments, or
at least commitments that lead to metaphysical claims, in his moral theory:
Freedom of the will most prominently, God and immortality, moral value
claims, a conception of pure rational agency and the resulting need for a view
of the self as other than merely appearance. Many contemporary ethicists
inspired by Kant shy away from these claims, particularly the roles of God,
immortality, and the metaphysics of persons-in-themselves. These metaphysical commitments are largely, though certainly not fully, out of favor in
contemporary philosophy. Much contemporary work does stress the practical
no entities other than those required by the natural sciences. This meaning is also implied in
Kants use of the term in one of his earliest works. The terms naturalism and naturalist occur
in the 1755 Universal Natural History four times (1:222, 1:223, 1:224, 1:239) without denition, but in context, their meaning is one who considers nature sufciently explained without
recourse to divine causes.
b.
The naturalist of pure reason is mentioned in the Prolegomena and the Critique of Pure
Reason in a way related to method, but quite different from any contemporary denition. In the
Prolegomena, the naturalist of pure reason is dened as he who trusts himself, without any
science, to decide in matters of metaphysics (4:314). The discussion in the Critique of Pure
Reason accords with the Prolegomenas denition. In the very nal page of the Critique, Kants
discussion of the history of pure reason turns to the question of method: The naturalist of pure
reason takes as his principle that through common understanding without science (which he
calls sound reason) more may be accomplished with regard to the most sublime questions that
constitute the task of metaphysics than through speculation (A855/B883). Kant apparently
offered this denition of naturalist in his lectures on logic as well; the Dohna-Wundlacken
Logic of 1792 quotes him as saying A naturalist is one who acquires cognitions that do not
constitute a system. Metaphysics and morals are the hobby horses of such people (24:783). The
methodological naturalism Kant rejects the naturalist of pure reason is far removed from
contemporary denitions of methodological naturalism because this naturalist is said to reject
rather than accept science.
c.
A few of the references concern theology. In the 1788 essay On the Use of Teleological
Principles in Philosophy, Kant notes that his explanations follow natural science so carefully
that he is assumed to be a naturalist (presumably in the rst sense) by one reviewer but rejects
that label as more properly belonging to theological discussions (8:178). The term appears in the
Religion, where Kant denes a naturalist as one who denies the reality of any supernatural
divine revelation (6:154), and in the Conict of the Faculties, where naturalism is dened in
religion as ecclesiastical faith without the Bible (7:60).
Graham Bird (Bird 1995) also canvasses Kants use of the term, but lists only the occurrences of
naturalism and not naturalist as well. John Zammito (Zammito 2008) refers to Birds paper
for the list of Kants uses of the terms but provides a different summary of their meaning.
Zammito sees Kant as associating naturalism with two impulses: First, a claim that nature is
ontologically self-sufcient, and second, the use of common sense or healthy reason in
place of abstruse philosophical reasoning (Zammito 2008, p. 545).
29
standpoint, but there has not been a thorough argument that Kants ethics as he
himself presents it in full does not require any ontological commitments. The
clearest way to frame this issue is in terms of a metaphysical naturalism. The
relations between the claims about the practical standpoint and claims about
realism and idealism also make a discussion of naturalism a good topic to
pursue for the light it sheds on the nature of Kants idealism in ethics.
Naturalism is a relatively recent term in philosophy, although what the
term signies has been part of philosophy for millennia. My discussion of the
term here is limited to the goal of providing a conception of naturalism
appropriate for an assessment of whether Kant can be considered a naturalist
in his ethics. Many aspects of the contemporary debate about naturalism in
philosophy will be left out of this discussion.32 I leave it to others who come to
different conclusions about the denition or viability of naturalism to evaluate
Kant on their own terms; however, I believe that my denitions capture the
kind of naturalism that is of most concern to Kant.33
Naturalism prioritizes the sciences over philosophy or any other discipline,
and it does so in two ways, metaphysically and methodologically.34 By
metaphysical naturalism, I mean the type of naturalism which accepts as
real only entities studied by the sciences. By methodological naturalism,
I mean the type of naturalism which accepts as sound only the procedures for
acquiring knowledge used in the sciences.
Since Kant perhaps more than anyone in the history of philosophy since
Plato is responsible for the inclusion of the a priori as a central element of
philosophy, it is clear that Kant rejects methodological naturalism.35 The
transcendental method is also at odds with a methodological naturalism. My
32
33
34
35
I argue in detail that Kant can be conceived as a naturalist to a large extent, encompassing not
only physics and chemistry but also the social sciences, in (Rauscher 2009b).
For recent discussions of naturalism, see (Ritchie 2008) and (De Caro and Macarthur 2004 and
2010).
Joel Smith and Peter Sullivan see three types of naturalism: Ontological (i.e. metaphysical),
methodological, and epistemological, which holds that knowledge is a natural phenomenon, but
they admit that the last is closely related to the rst two and can be understood for the most part
in those ways (Smith and Sullivan 2011, pp. 1014). Mario de Caro and David Macarthur
initially follow the division into ontological (metaphysical) and methodological naturalism
before suggesting a third type of naturalism, semantic naturalism which claims that only
the concepts employed by the natural sciences are genuine concepts (De Caro and Macarthur
2004, pp. 37). I believe that this semantic naturalism can be understood in terms of methodological naturalism and will not pursue it as a distinct form of naturalism in relation to Kant. Kant
has been interpreted semantically by others, although not as a semantic naturalist. See (Loparic
2005) and (Hanna 2001).
Michael Friedman has characterized philosophical naturalism as a two-pronged thesis, both
prongs of which are methodological: First, the rejection of a priori knowledge, and second, the
placement of philosophy among the empirical natural sciences. Friedman devoted his
1996 Presidential Address to the American Philosophical Association to his argument that
naturalism so understood has reached the end of its useful life (Friedman 1997, p. 7).
30
36
Bird follows this remark with a claim that Kants approach is acceptable to a naturalist or
antinaturalist, although by those terms he is referring to the methodological approach as well in
which the scientic approach needs a philosophical supplement (Bird 1995, pp. 40608). Bird
is responding to a claim by John Skorupski that Kants transcendental idealism is an antinaturalism. Part of Birds complaint against Skorupski is that naturalism is ill-dened. Despite a
vague denition of naturalism, Skorupski sums up his own view that Kant cannot be one quite
well in this passage:
Whatever else is obscure about Kants transcendental idealism, one thing is clear it involves
the rejection of naturalism . . . The great cultural inuence of Kants Critique stems from this:
that it upstages the natural attitude the perspective of common sense and its outgrowth,
science. If Kant is right, the natural attitude cannot give the absolute truth about the world and
our situation in it. (Skorupski 1990, p. 7)
Skorupski clearly takes transcendental idealism to be a methodologically antinaturalistic thesis.
31
38
39
40
41
(Zammito 2008, pp. 540, 55152).The difference between Birds claim that Kant is compatible
with metaphysical naturalism and Zammitos that he is not, lies in their understandings of the
role of reason and transcendental methodology, a topic I will cover in Chapter 2.
Another discussion of naturalism in Kant that cuts across the metaphysical/methodological
divide is given by Sami Pihlstrm, who presents a naturalistic pragmatic account of transcendental philosophy in (Pihlstrm 2001).
(Pollok 2013).
The article was appropriate for its time since before the 1980s, much of the literature on Kants
ethics limited itself to the Groundwork even to the extent that Kants four examples of specic
duties provided in the explanation of the categorical imperative were misunderstood as actual
derivations (Dye 1978).
Guyers essay (Guyer 2007b) offers some textual support for his claim that Kants precritical
ethics considered this naturalistic approach, but as Henry Allison notes in his comments to the
essay, it does not reect Kants more consistent precritical position that does not rely on a
psychological foundation (Allison 2007).
32
42
43
44
DallAgnol (DallAgnol 2013) is responding to several papers of mine, in particular the paper
Kant as Metaphysical Naturalist (Rauscher 2009b) but also a paper on the postulates
(Rauscher 2007), which forms the basis of Chapter 5, and Pure Practical Reason as a Natural
Faculty which was published in Portuguese (Rauscher 2006a), parts of which appear in
Chapter 4. He identies the scope of my claim too broadly to encompass a complete metaphysical naturalism instead of one limited to ethics.
The quotations that follow are taken from (Wood 2007, pp. 47274).
This claim is puzzling in light of Woods rejection of supernatural ideas that are used in place of
natural explanations. He admits that there is some mystery in life and even allows that pious
naturalists ought to respect and tolerate supernatural ideas. He holds that philosophers ought to
33
45
46
be interested in them, at least to the extent of seeing which can be made intellectually
respectable. Yet at the same time, Wood denounces those who hold supernatural beliefs for
which there is no proof or evidence and even claims that people who take the mystery of life to
open the way for such ideas as guilty of intellectual self-contempt. This dichotomous
approach would need to be grounded on a criterion that can be used to distinguish intellectually
respectable from nonrespectable supernatural ideas; I doubt that any methodological naturalist
could provide such a criterion. There appears to be no naturalistic way for Wood to distinguish
philosophically acceptable explanations that operate without reference to nature and philosophically unacceptable beliefs that do the same (Wood 2007, pp. 47273).
I argue at length for this position in (Rauscher 2009b). Here, I provide a very short summary of
the main points.
There is a plausible interpretation of Kant to that effect given by Graham Bird in The
Revolutionary Kant and elsewhere. Bird offers a methodological understanding of
transcendental idealism in which Kants critical project is seen as opposed to traditional
philosophy by employing a critical method to evaluate knowledge claims made in nature. Bird
stresses the empirical reality of objects in space and time and holds that Kant does argue for the
34
47
existence of any transcendent objects. He cites many earlier commentators with similar revolutionary views that differ in their particular interpretations, among them Henry Allison
(Allison 2004), Arthur Collins (Collins 1999), and Paul Abela (Abela 2002a) (for his complete
list, see (Bird 2006,p. 776, n 18)). Allison does not see Kant as a metaphysical naturalist, at least
with regard to agent causation (see especially [Allison 1997]), but Abela does think that Kants
ontology need not extend at all beyond the empirically real to any things in themselves, at least
for theoretical purposes (Abela 2002a, pp. 29293). Collins takes Kant to be an empirical realist
and tries to dispel the belief that Kant is an idealist of any ontological kind. Birds book does not
use the term naturalism to a great extent but does embody metaphysical naturalism nonetheless (Bird 2006, pp. 8392). Two discussions that hold that Kant provides arguments against
metaphysical naturalism in the antinomies are (Falkenburg 2004) and (Pollok 2013). Otfried
Hffe also argues that Kant provides a critique of naturalism but limits his discussion to
methodological issues (Hffe 2010, pp. 188191).
Paul Abela interprets this passage as an admission by Kant that there can be hidden truths
truths which are inquiry transcendent that would remain outside the set of truths that would
emerge even at the end of inquiry (Abela, 2002a, p. 238). He must be operating with a crucial
assumption, namely that the laws of nature which we use to posit these unobservable entities are
not themselves part of the inquiry, even the ideal end of inquiry which for Kant would equal
the rules and concepts of the understanding ideally systematized by reason using its regulative
ideas. He states this assumption in the following way It relates instead to the formal structure
of the regress itself as we extend it, counterfactually, to the desired point (Abela, 2002a, p. 239,
emphasis mine). Abela is here advocating a position in which the laws of nature applied to our
perceptions are merely a subset of some greater set of possible perceptions which could, in
35
In Kants thought, the laws of physics cannot explain chemicals, organisms, human social systems, and minds. They are not susceptible to
mathematization, and some are also not susceptible to testing or mechanical
laws. All are, however, studied by some form of investigation according to
some of the other requirements for natural law such as systematization, the
use of the categories, and the inclusion of higher level concepts. Proper
scientic causal laws are limited to those of physics and other sciences that
could in the future be mathematizable such as chemistry (4:47071). The
entities of physics in space (moving in time) provide the ontological foundation for other sciences that view the properties or arrangements of matter
from different perspectives, for example, biology describes organisms using
teleological principles but without claiming that organisms are more than
material entities.
The discipline of history which can be used to exemplify the social
sciences also relies on teleological laws. This is conrmed in the Critique
of the Power of Judgment when Kant devotes part of the discussion of the
teleological power of judgment to the study of organisms which exhibit
internal purposiveness (biology) and part to the study of beings who exhibit
external purposiveness (history) (5:425f).That the social science disciplines
also include teleological reasoning does not affect the ontological status of
their objects in nature any more than it would for biology. Further, although
Kant denies them the status of a natural science (4:471), he does not deny that
the subject matter of these social sciences is both determinate and part of
empirical nature.
Empirical psychology and, more broadly, anthropology (for simplicity,
I will focus on empirical psychology), are particularly important for this
book since morality involves the mental faculties of will, choice, and reason.
Unlike the social sciences, which rely on outer behavior, empirical psychology describes that part of nature that is revealed only through inner
intuition. While Kant does not deny that mental states could in principle be
instantiated in physical states of the brain, he does not think that such
reductionism is useful or even possible in explanation (7:119, B420). Still,
mental states as such, like material states, are subject to deterministic relations even though we are unable to know any deterministic laws describing
them. This is clear regarding empirical psychology in several places. In the
rst Critique, Kant divides what he calls immanent physiology, or the
study of nature as the sum total of the objects of the senses, into physics
and psychology, the objects of outer and inner sense, respectively (A846/
B874), and adds that empirical psychology can use the a priori principles
formal terms, connect to the laws of nature but which in fact do not, at least from the perspective
of the end of human inquiry.
36
For a discussion of this relation between the transcendental psychology of the rst Critique and
empirical psychology, see (Schmidt 2008).
37
particular moral principles, properties, or objects in Kants theory that constitute the material for the realism/idealism debate; they will also serve as the
elements to be incorporated into a metaphysical naturalism. This analysis will
contribute two things to setting the stage for my later arguments: First, sharpen
the focus of the debate between realism and idealism on the one hand and
naturalism and antinaturalism on the other hand in order to unearth the specic
interpretive points in Kant at issue, and second, reveal some of the particular
grounds for both realist and nonrealist claims in order to begin assessing their
strength.
A.
B.
C.
D.
E.
F.
G.
H.
A. The value of particular chosen ends should be uncontroversially understood as idealist. Subjective ends such as drinking wine, listening to jazz,
running a marathon, and pursuing a career have value only because some
valuing beings confer value upon them. Kant lays this relation out in the
Groundwork when he says The ends that a rational being intends at his
discretion as effects of his actions (material ends) are all only relative; for
only their mere relation to a specially constituted faculty of desire on the
part of the subject gives them their value (4:427, modied). Moral realism
about the value of particular ends would require that the value is independent of the agent, whether it be the agents choice, set of beliefs, or nature.
Hence the value of these particular ends must be empirically idealist. At the
transcendental level, one must abstract from the particular content of the
chosen ends and simply treat them as possible contingent choices of agents
capable of setting and pursuing ends. The particular ends operate at the
empirical level and clearly fall within the ontology of nature as understood
through empirical psychology. The ability of agents freely to set and
pursue ends is a further matter treated in (H). This idealism is of little
direct import in the debate about Kant and moral realism since no Kant
interpreter understands the value of particular ends as real. But it does offer
a toehold for the idealist, who can point to Kants justication the relation
to something in the subject to emphasize the role of the subject in value.
B. Particular duties or norms are distinguished from the categorical
imperative. No one doubts that Kant intends for particular maxims to be
tested using the rst formula and subsequent formulas that include the term
38
maxims such as the formula of autonomy (at 4:432) and the formula of
the kingdom of ends (at 43536), both of which can be understood to carry
the emphasis on maxims over from the formula of universal law. The
formula of humanity can also provide a basis for the adoption of particular
duties or norms. They can be abstracted from their source in the categorical
imperative itself (F). The duties are distinct from the moral law or categorical imperative itself for a very good reason: They depend in part on
empirical information about the nature of human beings as a particular
kind of nite rational being.49 Each of the four examples Kant provides in
the Groundwork requires some empirical information about human nature:
Human beings are susceptible to death, cannot read one anothers minds,
have latent talents that require development and practice, and lack complete self-sufciency. Many particular principles and duties are generated
through an application of purely rational considerations from the moral law
combined with empirical information. Kant has room in his theory for
different particular kinds of nite rational beings to have different particular duties based upon different empirical natures.
A realist about these particular duties would have to admit that they
are not themselves a priori in that they rest on information that is itself
independent of the a priori categorical imperative.50 But because they rest
on human nature in general, the realist might argue that these duties are
empirically real even though transcendentally ideal. Thomas Hill and
Andrews Reath view these particular duties as constructed because they
can be said to result from a decision procedure to determine which particular duties human beings would themselves derive (hypothetically) using
the categorical imperative. Hill provides a constructivist account of the
49
50
Kants work in anthropology is in part aimed at determining these particular duties on both the
species and the cultural levels. Robert Louden and Allen Wood both devote signicant attention
to the application of anthropology to the derivation of particular duties in Kant (Wood 1999,
pp. 193320) (Louden 2000).
The Kantian moral realist Allen Wood acknowledges that our actual duties as provided in the
published Metaphysics of Morals are not entirely a priori but are derived from the categorical
imperative as applied to human nature in general (Wood 1999, pp. 19596). In order to act in
accordance with the value of humanity, for example, we would need to know what particular
kinds of human ends there are that might or might not accord with the more abstract value of
humanity. He thus allows a distinction between two levels of moral law: The pure moral law
itself and the set of particular duties that result from the application of that moral law to
empirical information about human nature. He does note that this position of Kants differs from
that provided in the Preface to the Groundwork, where a metaphysics of morals was limited to
the pure moral law, and all empirical information about human nature was part of anthropology.
See also (Wood 2008, p. 60), where he holds that the particular rules and duties result when
the pure law and values are interpreted in the light of human nature and even particular
cultural or historical circumstances and are not the result of a rigorous deductive procedure so
not contained in the moral law itself.
39
derivation of particular moral rules or norms from the categorical imperative treated as a given.51 Similarly, Reath offers an interpretation of
particular duties as constructed while the categorical imperative itself is
not.52 This is the decision procedure type of constructivism in which the
duties depend upon some particular, in this case, hypothetical activity of
empirical moral agents. It is wrong to understand these particular duties as
optional, as a decision procedure sometimes suggests. In both Hills and
Reaths accounts, the hypothetical agents are not given leeway in determining the particular principles or duties at stake once the particular information about human nature, historical or cultural setting, and the like are
introduced as the basis for the decision in addition to the categorical
imperative. Since those things are empirically real, the duties could be
empirically real as well. But since for both Hill and Reath they depend
upon a hypothetical decision of the agent, they would be empirically ideal.
Under the nature of agency interpretation that I take, these duties would be
empirically real because they would follow directly from the nature of
empirical moral agents. Their empirical reality, however, is crucially
dependent on the empirical reality of the moral law itself (F). If it were
to turn out that the moral law itself is empirically ideal, then these particular
duties would be as well because they are derived from it. Similarly, these
duties could be comprehended within a metaphysical naturalism if the
categorical imperative is as well.
C. The value of necessary ends, in contrast to particular contingent ends
discussed in (A), is a matter of great controversy in the debate about moral
realism in Kant. There are two kinds of necessary ends: The ends that are
also duties presented in the Doctrine of Virtue, namely the happiness of
others and ones own perfection (6:38586), and humanity as an end in
itself as rst explained in the Groundwork (4:4289).
The ends that are also duties of virtue can be understood in two ways.
On the one hand, to the extent that the necessary ends of virtue are specic
ways in which Kant eshes out the value of humanity as an end in itself,
they can be assumed to have the same realist or idealist status as the latter.
51
52
Hill details his constructivism in his chapter Kantian Constructivism as Normative Ethics
(Hill 2012, pp. 7192). His version of constructivism centers on Kants Doctrine of Virtue
rather than the Groundwork. He takes the Doctrine of Virtue to begin with a general conception
of the value of persons and to lay out principles that human beings would adopt in order to
create a society that incorporates that value. Only the principles, not the value of persons, is
constructed.
Reath presents his position in (Reath 1994). Patrick Kain correctly argues that any position such
as Reaths does not entail that the particular duties are in any sense optional to empirical agents
in (Kain 2004). The duties are based on human nature rather than on any kind of preferences
any particular agent or all particular agents may have within the decision procedure situation.
40
53
Formosa and Wood both take this option, for different reasons. Formosa classies these two
necessary ends of virtue along with the particular ends that human beings choose as their own,
calling all of them ends with conditional and nonabsolute worth in contrast to humanity as an
end in itself which has unconditional and absolute value (Formosa 2012, p.9). On his account,
these two necessary duties of virtue have the status of chosen ends as in (B) and would be
treated as constructed in the same way. His reason is that the necessary duties of virtue are wide
duties and thereby do not have to be the basis of every act. But this reason is insufcient to
make these ends conditional. Formosa is conating the happiness of others and ones own
perfection as necessary ends to adopt in general with the particular ends that we may adopt in
particular actions that will implement those necessary ends in various ways. This is the import
of calling them wide duties. The categorical imperative demands that the two ends be adopted
and to that extent they are of unconditional and absolute value. But Kant recognizes that many
particular maxims can honor these ends. A duty of wide obligation does not allow human
beings to reject the necessary end but only to act on some maxim or other that might further
ones own perfection or the happiness of others: a wide duty is not to be taken as permission to
make exceptions to the maxim of actions but only as permission to limit ones maxim of duty by
another (6:390).Thus these two necessary ends of virtue are not conditional and of nonabsolute
value. Wood holds that empirical information about human nature is used to determine which
ends we ought to set as moral beings and identies them as the ends which are also duties,
although elsewhere he allows that ones own perfection and the happiness of others are ends of
pure reason as opposed to simply ends of reason; Wood also connects the concept of perfection
to happiness in that part of perfection is being able to use ones developed talents (Wood 1999,
pp. 195, 32628). Wood is not clear about the status of the general ends that are also duties. He
appears to treat them as kinds of ends rather than generalized ends that we ought to adopt (Wood
2008, p. 167). In adopting any particular end that is, for example, of the kind that promotes the
happiness of others, we would need to use empirical information as Wood says. On Woods
account, the necessary ends of virtue would have the same status as particular duties that require
empirical information as in (A) and be subject to the kind of constructivism that Hill suggests.
But Kant holds in the second Critique that the concept of happiness not the conception, or
particular content of what things count as happiness, which is clearly empirical is an a priori
concept relevant for all nite rational beings and not dependent upon human nature (5:25).
Since the concept of happiness is a priori, the two necessary ends of virtue are also a priori, do
not rely on information about human nature, and can be understood as having the same status as
humanity as an end in itself.
41
assessment of the relation the value of humanity has to the debate about
realism and naturalism. Kantian realists generally hold that the value of
humanity as an end in itself is independent of and conceptually prior to the
categorical imperative and even serves as the ground of the categorical
imperative. Idealists and some constructivists in contrast hold that the value
of humanity is derived directly from the categorical imperative, in which
case, the reality or ideality of the value of humanity depends upon the
status of the moral law itself (F). Alternatively, some constructivists claim
that the value of humanity is dependent upon a conception of agency
necessarily adopted by agents as such; without the existence of that
particular kind of agency there would be no need for the value of humanity
or any value at all.54 This view sees a parallel between the construction of
moral value and the construction of the categorical imperative itself rather
than seeing the former as derived from the latter. Its success or failure
would be tied to the success or failure of the parallel claim about the
categorical imperative to be discussed in (E) later on.
Value is often understood as a property of objects; in the case of the
value of humanity, a property of nite rational beings, but more broadly
speaking, objects or mental states could possess intrinsic value. A value
property could be understood at either the transcendental or empirical level.
If value were known to be an intrinsic property of some beings or things-inthemselves independently of the nature of moral agency, then value would
be transcendentally real. An example would be if God were to have created
some beings with intrinsic value. This value would be a constraint on the
conception of rational agency itself. In this case, the value would ground
the very possibility of a moral law. Alternately, if absolute moral value
were known only as a requirement of moral agency, namely, that in order
for moral experience to be possible there must be something of value
independent of the particular moral agents, then it could be empirically
real but transcendentally ideal. This conception of value would be limited
to the kind of thing that is compatible with the demands of the categorical
imperative, namely, humanity. Rational beings possessing humanity would
have value as an intrinsic property. It would still, however, not t into a
naturalist framework because empirically there would be a value property
of some beings in nature that does not gure in Kants conception of
nature. A third option is a transcendental and empirical ideality of value
54
I say that particular kind of agency in order to allow that a constructivist could hold either of
two alternatives. First, a constructivist might hold that there is only one possible kind of agency,
free rational agency, as Korsgaard does (Korsgaard 1996b). Second, a constructivism might
hold that it is possible for other kinds of agencies to exist which might not require the same
moral values or the moral law.
42
that would be fully compatible with metaphysical naturalism. The empirical ideality of the value of humanity, I will argue, consists of an objective
ordering of ends by reason such that humanity is always to be treated as a
limiting end for maxims, whatever subjective contingent end the empirical
moral agent would be pursuing. This ordering by reason would be based on
reasons own needs and not on any intrinsic value property in human
beings and so would be compatible with a metaphysical naturalism and
an empirical ideality of value.
D. The highest good is an object of pure practical reason that, Kant argues,
requires the postulates of God and immortality for its realization. Kant
understands the highest good as an idea of reason that serves as an
archetype of a perfectly just and virtuous world to be attained. As such,
it is not an object that exists empirically or as a thing-in-itself (except to the
extent that it is actualized by moral agents through their actions). Realists
and idealists would agree that the idea of the highest good is dependent
upon the existence of nite, rational moral agents for whom the highest
good serves to satisfy both their obligation to be moral and their desire for
happiness. Perfectly rational agents are not in need of the highest good
because they have no independent desire for happiness. As an idea, the
highest good is dependent upon the moral agent. The realization of the
highest good itself would be brought about through the actions of moral
agents and of God. Actions of moral agents depend empirically upon them
and are thus empirically ideal.55 God would, in contrast, seem to have a
transcendentally real status.
Two of the three postulates of practical reason immortality of the soul
and the existence of God are postulated because they are needed for
realization of the highest good. (Freedom, the third postulate, underlies the
categorical imperative itself and has no relation to the highest good.) These
postulates are what I called earlier principles, properties, or objects required
for morality rather than purely moral principles, properties, or objects; but
because they play a role in Kants moral theory, they have been incorporated
into the realism/idealism debate, with discussion focusing on the existence
of God. The realist holds that the postulate requires that God exist, while an
idealist can hold that human beings are required to believe in God as a
distinct being but that God need not exist. The most extreme idealist would
hold that the postulate not only does not require that God exist, proper
comprehension of the function of the postulate shows that it does not involve
human beings believing that God as a distinct being exist but that the idea of
55
Treating the highest good as an object to be actualized by human beings is constructivist in the
different sense of bringing an idea to fulllment. For a discussion of the highest good as
something for human beings to create over the course of history, see (Reath 1988).
43
God be used for immanent purposes. This last is the position I argue for in
Chapter 5. A metaphysical naturalism, of course, would preclude any role in
morality for the immortality of the soul or the existence of God.
E. Moral obligation is the fth area of Kants ethics at issue. One can
distinguish the obligation from the other elements of a duty because one
might be able to identify a value or acknowledge a moral law without being
obligated to uphold or follow them. Two examples would be an amoralist
and a cultural relativist: Each might recognize that there are values or laws
without accepting any obligation. (I am not concerned with the plausibility
of these positions.) The realism or idealism of moral obligation in Kant
depends upon whether it is the result of some procedure or in some other way
can be separated from the moral law, value, or practical reason itself.56 If it
cannot be separated, then obligation is inherent in them and the realism or
idealism of obligation would depend upon the real or ideal status of the moral
law or practical reason.57 In particular, if the nite will can be conceived in a
way that allows it to obligate itself to the categorical imperative, then
obligation would be idealist. Some constructivists hold that obligation is a
matter of an autonomous act of will in giving the law to itself.
One might tie obligation to the nature of the nite rational will. This is
successful in presenting an idealist conception of obligation only if the
nature of the nite rational will is understood as idealist, since it grounds
obligation in the nature of our nite will. Realists and idealists can agree
that obligation is tied to the nature of the human nite rational will; the
dispute is only about whether a Kantian theory should take the nite
rational will as real or ideal. And this question cuts to the heart of the
nature of practical reason and practical philosophy itself. Whether practical
reason itself is realist or idealist is the deeper question and will be noted in a
few paragraphs under (G).
Obligation is not considered to be a natural relation, making it difcult
to see how a metaphysical naturalism would work. Unlike value as a
property that purported to exist within objects or rational agents, obligation
is a relation between an agent and a set of duties and ends. The duties and
ends themselves, both necessary and contingent, could lie within nature as
56
57
Robert Stern offers a way to distinguish obligation from the moral law. Stern, who takes a
realist position regarding value and law, argues that moral obligation is idealist because we
give the content of morality its obligatory form, in so far as this depends on our limitedness as
nite creatures. (Stern 2011, p. 90)
Patrick Kain argues that because obligation is tied to our rational will, it must be understood as
real rather than as constructed (Kain 2004). In Chapter 3, I argue that obligation is similarly a
product of reason, but in the guise of the commanding nature of the categorical imperative as a
fact of reason. Obligation for either of these views is tied to the status of the faculty of reason.
I will show in Chapters 4 and 7 that reason should be understood as not transcendentally real.
44
(Kain 2004).
It is also worth noting that there are some constructivist positions that give a different origin to
the moral law than the faculty of pure reason. One strand of Christine Korsgaards constructivism centers on the construction of the moral law itself. (Korsgaards position has shifted over
45
The categorical imperative in empirical consciousness could be understood as empirically real or ideal. As a fact of reason, the awareness of the
categorical imperative is in the empirical mind of the actual moral agent. If
there is also a transcendental justication for the categorical imperative,
then it would be empirically real because it is valid independent of particular moral agents. But if it has no transcendental justication, then it would
be empirically ideal. The subjective nature of obligation, the enumeration
of particular duties, and in general, the experience of morality would
remain the same. I will examine the reasons in favor and against this
interpretation in Chapters 3 and 7.
G. Practical reason itself, the source of the moral law, is perhaps the core
topic for this book. A realist about practical reason would hold that it
discloses to moral agents some facts about the world itself independent of
those moral agents. A transcendentally real reason would disclose something about the nature of things in general because things in general would
be fundamentally rational. Morality would then be transcendentally real
independent of the transcendental moral agent. But if, as I argue Kant
holds, practical reason itself is constitutive of and restricted to moral
agency, then a transcendental idealist position about pure practical reason
is warranted. This position limits the existence of the faculty of reason as
well as its scope to only those beings capable of moral agency. I will
examine the basis for moral agency throughout the book, but the discussion
of the nature of practical philosophy in Chapter 2 will show that the
domain of practical reason is free acts, so that practical reason is valid
only for those beings.
On the empirical level, practical reason would be empirically real, that
is, independent of the empirical moral agent, if the empirical agent can
know that she meets the conditions of transcendental moral agency. Kant
provides reasons to support this identication when talking about the
priority of the practical over theoretical concerns, the inevitability of the
fact of reason, and his arguments for the categorical imperative. But he also
hesitates to allow that a transcendental deduction for morality is possible,
time. I am referring here to her arguments in The Sources of Normativity (Korsgaard 1996b).)
Her approach, as discussed here, takes as its basis a view of agency. If an agent asks not merely
what she should do on the basis of her desires but the broader question of what she should do
overall, the agent is driven by a process of reection to nd the answer that she ought to act only
in ways that could count as laws, and hence she has formulated for herself the categorical
imperative. Korsgaard claims that rational agency itself requires the categorical imperative but
not that it stems directly from reason, as might, for example, the law of the excluded middle.
Without the active reection of agents, there would be no categorical imperative. In my
argument in the following chapters, I will show that the moral law is a direct product of
reason, so I reject Korsgaards type of construction of the moral law.
46
ONeill makes this argument particularly in (ONeill 1989, pp. 350) and (ONeill 2002).
Another aspect of her argument is that any kind of agreement at all that does not accept some
kind of preexistent authority is governed by the categorical imperative, making it the supreme
principle not merely of practical philosophy but also of theoretical philosophy. I take this to
mean that practical reason has primacy over theoretical reason not merely in being authorized to
accept certain claims that cannot be proven by theoretical reason (the postulates) but to trump
claims of theoretical reason in action, as discussed in Chapter 3.
47
each, but here it is important to note that a claim that practical reason is
constructed by the free actions of human beings requires the libertarian
conception. I will argue that a rst-person freedom of deliberation for Kant
is vital to the foundation of practical philosophy but that at the same time it
is not transcendentally justied. Kants conception of freedom is multifaceted, and the transcendental justication of freedom holds for reason and
rational action rather than for free choice, although he does allow for a
postulate of freedom related to free choice. Still, this freedom of choice
does not ground actual deliberation; empirically, human beings can consider themselves rational with a greater justication than they can attribute
to themselves a free choice.
To say that reason itself is ideal is to make a claim about the nature of
practical reason and, further, the nature of philosophy as a whole. It delves
into the heart of Kants claims about philosophical method itself. ONeills
interpretation, although incorrect as an account of the nature of reason,
identies the core of Kants philosophical project as a human-centered
endeavor, not merely in the familiar sense in which Kant is delineating the
basis of human knowledge and its limits but also in the broader sense in
which Kant comprehends that philosophical project itself as a part of the
active life of human beings who must confront the world from their
perspective. To be able to judge whether one can claim that reason itself
is constructed, or real, or ideal, one must examine the nature of philosophy
and reason in Kant.
H. Freedom of choice is the nal controversial element to be examined.
Kants moral theory emphasizes freedom. I differentiate the freedom of
the faculty of reason to provide the moral law from the freedom of the
faculty of choice to make decisions as the specic way in which I discuss
Kants Wille/Willkr distinction. The freedom of reason is a part of the
discussion of the nature of practical reason, but the freedom of the faculty
of choice is a distinct issue. The second kind of freedom is much more
directly at issue in metaphysical naturalism and it plays little part in the
realist/idealist debate. My Chapter 6 will examine the place of the freedom of the faculty of choice in Kants overall moral theory and provide a
metaphysically naturalistic interpretation that stresses its status as a
postulate. Whatever its specic metaphysical status, the freedom of choice
ought to be considered as ideal rather than real because it is a property of
moral agency itself. Moral idealism is perfectly compatible with a libertarian conception of agency. A realist might argue that actual freedom of
choice would have to be true of all moral agents and would thus fall
outside the boundaries of moral idealism understood as prioritizing the
moral agent as subject. A moral realist might further argue that this
freedom is the ground of the value of moral agents as objects that must
48
49
The work of the previous chapter on realism/idealism and metaphysical naturalism point toward the need to examine the nature and methodology of Kants
philosophy in order to see how it might determine the ontology required for
ethics. The empirical/transcendental distinction, the nature of moral agency
and the status of reason itself, and the ontological status of moral principles,
properties, and objects will all depend on their place in Kants philosophy as a
whole. This chapter will show that the very broad strokes of Kants philosophy the topic and structure of philosophy as a whole, the transcendental
method, and the practical/theoretical distinction provide a key basis for
resolving the particular issues about naturalism and realism.
The broadest questions that one might ask about Kants ethics (indeed,
about any ethical system) are these: what is ethics? Why does existence
have a normative component? What is the nature and scope of morality?
What kinds of beings partake of it? Why is there any such thing as ethics at
all? This last is a twofold question. It is not only asking about the existence
and nature of ethics, that is, whether there is good or right or value. The
question what is ethics also asks about the philosophical study of those
aspects of the world such as good and right and value. Why should the
study of ethics have a place within philosophy instead of some other
discipline? How does it differ from other elements of philosophy? Is there
a distinct methodology for ethics? To what kind of philosophical treatment
is it susceptible: justication, application, explanation? The broadest question about Kants philosophical ethics is really about what kind of inquiry
ethics is and whether such inquiry is possible. The particular subject matter
of ethics will in part determine and in part be determined by ethics as
philosophical inquiry.
To ask about the very possibility of a philosophical study of something is
itself the hallmark of Kantian philosophy. Kants Critique of Pure Reason is
devoted to determining whether and how metaphysics itself is even possible
(Axii); the Prolegomena is even clearer in making this question a centerpiece of the structure of the book, devoting the entire third part and the
conclusion to the question of how metaphysics is possible (4:327, 4:365).
50
51
52
The contents of the Architectonic chapter parallel the rst pages of Kants lecture on Philosophical Encyclopedia (29:545), with the exception of the denition of philosophy in relation to the
highest ends of human beings. This material does not appear in the textbook which Kant used for
the course (Feder 1769). Information on Kants textbooks is taken from (Naragon 2006).
I use the English term domain for various German terms that Kant uses with the same
meaning: Umfang in the rst Critique (A832/B860) and Gebiet in the third (5:17476). In the
discussion of the latter, Kant contrasts a domain determined by our cognitive faculties with a
broader territory (Boden) that includes all areas where cognition is possible for us, including
empirical cognition. Although Kant does not use the term idea here in establishing a domain,
he does claim that the domain is determined by the legislative power of the concepts and
corresponding faculty of cognition (reason or understanding).
Here, I follow Paula Manchester (Manchester 2008), who shows how Kants discussion of
architectonic is intended primarily to stress the active nature of reason itself as providing the
structure for its own investigations. This active organization replaces the passive scholastic
organization that stems from historical cognition. Manchester also provides a thorough discussion of previous views of the Architectonic that associate it more with merely providing an
architecture or structure within which reason would operate.
53
In this paragraph, Kant holds that the method for pure practical reason
concerns what is done with the pure practical principles in motivating human
beings to action and providing means for adoption of the correct maxims. But
the paragraph goes further than this to suggest that there is, in fact, no method
for collecting together practical principles, a claim which would entail that
there is no system of pure practical reason but only an aggregate of particular
moral principles. This further claim, however, does not reect Kants actual
practice, given, for example, that he devotes his Metaphysics of Morals to the
task of delimiting a system of ethics based upon the categorical imperative. As
this chapter develops, I will show that Kant does apply the requirements for a
systematic organization to ethics.
The domain of philosophy
Philosophy as a whole is a system of rational cognition from concepts, and
thus the systematic organization of rational cognition must be organized by
reason in such a way that it has both a domain and an internal structure. The
domain would be related to philosophys guiding idea, which has often, Kant
hints, been mistakenly neglected in place of mere systematic unity of rational
cognition as such; this is the scholastic concept of philosophy that has no
further purpose or end. The true ideal concept of philosophy, which Kant
calls the cosmopolitan concept, is the science of the relation of all cognition
to the essential ends of human reason of which the highest is the nal end
(Endzweck) or vocation of human beings (A83940/B86768).4
4
As an ideal, Kants denition of philosophy has a limitation: there is as yet no such complete
science. Kant claims that philosophy ought to be understood as the archetype against which
attempts to philosophize are compared. Actual philosophical practice ought to approach this
ideal philosophical cognition. (A838/B866)
54
This guiding thread for philosophy provides a broad domain for philosophy
insofar as virtually all cognitions can be related to the nal end of human
beings, given that human practical deliberation and its resulting actions can
benet from knowledge and skill related to any possible action. One might
think that it is problematic that some of these cognitions would be empirical
even though reason itself operates a priori. Kant does in fact allow for
reasoning based upon empirical principles as part of empirical philosophy
in contrast to cognition from pure reason (A840/B868). Cognition by reason
from empirical principles is still cognition by reason. The domain for philosophy is identied as all cognition related to the essential ends of human reason,
provided that cognition is derived from principles.
In only one other published work does Kant provide such an allencompassing denition of philosophy.5 The Jsche Logic, which was based
on Kants course lecture notes and not written personally by Kant, uses the
same language as the Architectonic in distinguishing the scholastic concept
from the cosmopolitan concept of philosophy, the latter of which is the
science of the nal ends of human reason (9:23), elaborated as the science
of the relation of all cognition and of all use of reason to the ultimate end of
human reason, to which, as the highest, all other ends are subordinated, and
in which they must all unite to form a unity (9:24).6 But since choice of ends
and thereby, the activity of reason are involved, Kant also calls philosophy
a science of the highest maxim for the use of our reason (9:24, see also
Metaphysic L2 28:533).
The ultimate end of human reason should not be confused with the ultimate
end (Endzweck) of the existence of the world, a topic that Kant discusses in
light of his teleology in the Critique of the Power of Judgment (5:43436). It is
possible for human reason to have one end and the existence of the world
another. Further, reason as an activity can possess its own ends or goals that
Kant does offer less sweeping denitions of philosophy. In the Introduction to the Critique of
the Power of Judgment, he denes philosophy excluding logic as principles of rational cognition of things by means of concepts (logic is said to consist of principles of the form of thinking
in general) (5:171); the First Introduction similarly calls it the system of rational cognition
through concepts(20:195). The Groundwork, accepting the ancient Greek division of philosophy into logic, physics, and ethics, simply treats philosophy as rational cognition (4:387). In
Kants lectures on ethics, he generally started with a simple division of philosophy into two parts
without dening philosophy as such (e.g., Collins 27:243), but the notes from his winter
1793 course do offer a denition similar to that of the Groundwork and Judgment (Vigilantius,
27:479).
Kants course lectures on logic generally contained the same distinctions, although without the
connection to the nal ends of human beings; the textbook Kant used matches the absence of
nal ends ((Meier 1752), which is available in Band 16 of Kants gesammelte Schriften). The
Vienna Logic (24:797800), for example, does not provide discussion of system or the denition
of philosophy.
55
are intrinsic to that activity7, whereas any nal end of the existence of the
world depends upon the supposition that the world has a supreme cause that
acts for ends. Without that kind of teleological view, there could be no ends for
things that are not themselves active as human reason is. A mere mechanical
explanation of the world does not invoke any ends at all. The activity of reason
is self-sufcient for producing ends while the existence of the world is not.
Finally, reason itself with its ends might provide human beings with the
conception of the ultimate end of the existence of the world for them to use
in interpreting their experience. The ultimate end would then be a product of
reason and not something in any way inherent in the world itself.
The identication of the domain of philosophy as the ultimate end of human
reason does not provide the content one might hope for, namely, an identication of the ultimate end of human reason. Kant identies this end in the Canon
of the Critique of Pure Reason as the determination of what is to be done
(A800/B828).8 The ultimate end of reason is thus moral. Since philosophy also
has the study of morality as one of its two main topical divisions, this is
another way of saying that the practical has primacy over the theoretical. I will
examine the nature of that priority in Chapter 3. Before that must come the
identication of precisely what counts as practical philosophy later in this
chapter.
The structure of philosophy
In addition to the domain, philosophy would have to have an internal structure
which also stems from the idea of a nal end for human beings. Kant offers
ve ways of dividing philosophy as a whole: logic and object-directed philosophy, critique and metaphysics, theoretical and practical, pure and empirical,
and the three (or four) main questions of reason.
A. Logic and object-directed philosophy: Logic is separated from the remainder of philosophy because, although it consists of cognition from principles, it has no object of its own in contrast to nature as the object of
theoretical metaphysics and freedom as the object of practical metaphysics.
Logic is instead a study of the rules of thought or the formal aspects of
cognition. The Preface to the Groundwork is clear in calling logic a science
7
The issue of any goals or interests of reason is more complicated than I am presenting it here.
Pauline Kleingeld argues that Kants claims about interests of reason cannot be taken merely
metaphorically (Pauline Kleingeld 1998). Axel Hutter stresses the centrality of interests in Kants
entire philosophical project (Hutter 2003).
Kant initially identies the ultimate end of human reason with three objects: freedom of the will,
the immortality of the soul, and the existence of God (A798/B826). He then subordinates them to
their more remote aim, which is the determination of duties. I will discuss the relation between
God and moral duties in Chapter 6 about the postulate of God.
56
For another examination of the Architectonic chapter that stresses the difference between
Critique and system, see (Kinnaman 2008).
Kants division is problematic in a few other aspects: the Groundwork appears to hover in a no
mans land between critique and systematic metaphysics. The titles of its sections indicate
transitions from and to various approaches to morality, including philosophical moral rational
cognition, metaphysics of morals, and critique of pure practical reason (4:393, 446).
Although commentators generally place the book as a foundational part of the metaphysics
rather than a critique, the third section is nonetheless often treated as if it were a kind of critique
in its discussion of the limits of practical philosophy and its inclusion of a deduction, which is
compared to the parts of the second critique that replace a deduction with the fact of reason.
Karl Ameriks, for example, features extensive discussion of the Groundwork in a line with the
Critiques in his book Interpreting Kants Critiques (Ameriks 2003). The Prolegomena presents
the results of the Critique of Pure Reason in an analytic rather than synthetic form and thus, has
a tenuous critical identity.
57
He offers four different characterizations of the difference between theoretical and practical philosophy. Because these characterizations reveal
important methodological considerations concerning the specic domain
for the practical, I will devote extensive space to them later in this chapter
by comparing them in detail.
D. Pure and empirical: Another division of philosophy is between cognition
from pure reason and rational cognition from principles derived empirically
(A840/B868). Philosophy had a wide scope in the centuries leading up to
Kant and included what today are considered natural and social sciences, so to
identify part of philosophy as based on empirical sources was not unusual.
To count as philosophy the sciences must still use principles that reason can
identify and work with. Empirical information from which no rational principles could be derived would not count as philosophy. Empirical philosophy,
then, is still rational philosophy rather than an exclusively empirical study.
E. The questions of reason: In two publications, Kant divides the interests of
reason into a set of questions: What can I know?; What ought I to do?;
What can I hope?; and What is the human being? In the Canon, Kant
includes only the rst three questions and then quickly assigns the rst to
speculative (theoretical) and the second to practical reason, with the third
consisting of both (A805/B833). In effect, then, this division reduces to
the theoretical/practical division. In the Jsche Logic two decades later, he
claims that the rst question belongs to metaphysics, the second to morals,
the third to religion, and the fourth to anthropology, which itself can also
be seen to subsume the rst three questions as well (9:25). He appears to
correlate the rst three questions with three tasks for the philosopher:
(a) determining the sources of knowledge, which would presumably correlate with theoretical metaphysics, (b) determining the use of knowledge,
which would correlate with morals, and (c) determining the boundaries of
reason, which might correlate with a philosophical approach to religion (the
term boundary (Grenze) invokes Kants Religion within the Boundaries
of Mere Reason11), or with the project of critique in general, or more
narrowly with the transcendental dialectic and teleology as they relate to
religious questions of the soul, free will, and God. One might also be
tempted to align the rst three questions with the three Critiques and the
last with the published Anthropology; however, the topics covered by the
questions extend to other works as well. Further, if anthropology is to
encompass the three prior questions, then it is not a distinct subdivision
within the structure of philosophy. The explanation of the four-question
division is in the end too vague to be helpful in dividing philosophys
domain, while the three-question division simply reduces to the theoretical/
11
In a letter to Carl Friedrich Studlin on May 4, 1793, Kant repeats these four questions and
explicitly identies the Religion as completing the answer to the third question (11:429).
58
practical dichotomy.12 In the end, despite its fame, this division of philosophy is unhelpful, and I will not discuss it further in this book.
These ve ways of providing an initial division of philosophy do not
correspond to one another. Leaving aside the questions that summarize philosophy, the others work variously in parallel or perpendicular to one another.
How do these various ways of distinguishing the tasks of philosophy provide
its internal structure, particularly in relation to ethics? The justication for this
structure rests on the requirements for eshing out the demands of morality.
First, the separation of logic from the remainder of philosophy provides the
basis on which any investigation by reason may progress. Second, the work of
critical philosophy is needed to develop the particular limits and foundations for
further systematic work. Third, the division into practical and theoretical is
necessary because the practical must be informed by the theoretical in order to
know how to implement practical ends. Finally, to determine the particular
moral duties that form the vocation of human beings, the pure must be separated
from the empirical so that reason can properly identify pure duties that will
ground all particular empirically-informed duties. These explanations of the
structure of philosophy, excepting the rst about logic, will receive more
attention in direct relation to the practical in each of the remaining two sections
of this chapter. One section will cover both the critique/system and pure/
empirical divisions, the second, the practical/theoretical.
The nature of the transcendental method and ontology
The divisions of philosophy into critique and system on the one hand and pure
and empirical on the other hand are related. Kant is particularly concerned
about the division between pure and empirical philosophy as this passage from
the Architectonic shows.
It is of the utmost importance to isolate cognitions that differ from one another in their
species and origin, and carefully to avoid mixing them together with others with which
they are usually connected in their use. What chemists do in analyzing materials, what
mathematicians do in their pure theory of magnitude, the philosopher is even more
obliged to do, so that he can securely determine the proper value and inuence of the
advantage that a special kind of cognition has over the aimless use of the understanding.
(A842/B870)
12
Axel Hutter suggests that another difference between the two sets of questions is that the task of
uniting the questions together belongs to the third question in the Critique of Pure Reason but
the fourth in the Logic. He claims that since the unication by anthropology in the Logic is
weaker than that of the Critique, Kant loses genuine systematic connection among the questions. I believe that Hutter is reading too much into the short presentation in the Logic and that
Kants fourth question does not exclude the third from combining the rst and second (Hutter
2003, pp. 8890).
59
The philosopher must be careful to distinguish the a priori from the a posteriori.
Kants rejection of any empirical basis for philosophy, any philosophy based
solely on a posteriori cognitions, is not a complete rejection of empirically
derived knowledge but a separation of empirically derived knowledge from a
priori. The study of nature and the study of morals will both have their pure and
empirical parts. The pure parts, metaphysics, are the distinctive contribution of
philosophy as the cognition from concepts which have an a priori origin.
The critical transcendental method is used in the determination of pure as
opposed to empirical cognitions. In what follows, I will summarize the transcendental method in a way that I hope is not controversial. Then I will show
how it can relate to a metaphysical naturalism and the empirical/transcendental
levels of moral realism.
Transcendental method in general
That philosophy requires a priori claims is central to Kants philosophical
framework and needs little elaboration. A priori cognition is dened by Kant
as that which occurs absolutely independently of all experience (B3) and
which involves necessity and strict universality (B4). A merely empirical
approach, one which draws all its cognitions from experience, could not
conclude that the cognitions are true necessarily and universally. But knowledge from concepts alone must be true necessarily and universally for the
domain governed by those concepts, in this case experience. When Kant
distinguished pure and empirical philosophy, he was distinguishing principles
known a priori from those known a posteriori. He makes this same kind of
distinction using different terminology when he contrasts historical and
rational cognitions, the former from what is given, the latter from principles
(A8356/B8634). Only rational cognition is such that it could have arisen
from reason alone. Kant puts it this way in the late essay What Real Progress
has Metaphysics Made in Germany?: If all knowledge is of empirical origin,
then . . . transcendental philosophy is itself an absurdity. (20:275). Critical
philosophical method must focus on a priori rather than a posteriori cognitions.
This role for the a priori lies at the basis of Kants critical project. Since
philosophy as a discipline is characterized by the use of reason, and any claims
by reason to knowledge of its own would be a priori, Kant holds that philosophy is in need of a critique in order to determine the possibility, validity, and
extent of reasons a priori claims. Critique exhibits a unique method: the
transcendental argument.13
13
The method of a transcendental argument is controversial inside and outside Kant scholarship.
I will not enter into the details of this work but will only offer a relatively non-controversial
account of its nature and role. Within Kant studies, see (Frster 1989). In contemporary
60
61
experience and reection on it, and therefore concerns not the lawfulness but
the fact from which the possession has arisen (A85/B117). Transcendental
philosophy abstracts from any empirical manifestation or source of a concept
to ask instead about its justication. The transcendental deduction of an a priori
cognition or principle, then, will show that it is necessary for experience yet
not derived from experience.
The method of transcendental deduction is not restricted to concepts necessary for the very possibility of experience in general. A transcendental deduction is one which assesses and justies the conditions necessary for the
possibility of some specic kind of experience, using the term broadly.
Although he does not identify them as transcendental, this is how Kant
understands a deduction when he discusses the role of deductions for the moral
law (4:447,454, 5:467), for aesthetic judgments (5:27980, 5:28990), and
for property (6:249250). The experience at issue in the rst Critiques transcendental deduction is the cognition of objects of experience, or as described
in the Prolegomena, the possibility of natural science (4:275 and 4:29496).
The range of other sorts of experiences whose conditions of possibility could
be the subject of a transcendental deduction is not explicitly limited by Kant. In
the rst Critique, a priori concepts which can relate to objects of experience are
specied, but this stricture is violated in the discussion of a transcendental
deduction of the moral law in the second Critique and other deductions. We
can assume that Kant restricts transcendental deductions to what can be known
absolutely a priori, whether it is a concept for an object, a moral law, a manner
of judgment, or a property relation.
Regardless of the kind of pure a priori cognition or principle, Kant is clear
that they cannot be derived empirically. A transcendental justication cannot
be based upon derivation from experience, but it can justify the validity of
cognitions or principles of the understanding or reason as they are used in
experience. Just as space and time are empirically real but transcendentally
ideal, a priori cognitions and principles can be understood as transcendentally
ideal (dependent upon the faculties of the subject) yet empirically real (formal
elements of experience governing empirical objects).
Take the pure concepts of the understanding, in particular causation, as
paradigmatic. Kant argues in the transcendental deduction that a necessary
condition for experience is the transcendental unity of apperception, that is,
that all intuitions be synthesized into one whole; cause and effect is one of the
ways in which this synthesis is exhibited. But the concept of cause and effect is
one that we can meet with in experience. The physical sciences make synthetic
a priori causal claims; to that extent, we are in possession of universal and
necessary principles such as causation (Prolegomena 4:295). Kants transcendental justication of the causal relation is intended to show how pure natural
science is possible. He does not deprive physics of these claims in order to shift
62
63
16
64
17
(Bird 1995). In contrast, David Bell understands transcendental idealism methodologically and
argues that it is a re-description of reality in a way that construes objects in accordance with
the awareness we may have of them, that is, their appearance (Bell 2001). Empirical realism, in
contrast, accepts that objects exist independent of any subjective awareness of them, and
naturalism would be a claim that only empirically real objects exist. Bell argues that appearances are incompatible with naturalism because as phenomena they are not items in the natural
world or subject at all to the laws of nature. Phenomena are instead a primitive. Bell
problematically separates appearances the subject of the transcendental analysis from
empirical reality as if the transcendental analysis did not provide the ground for the scientic
analysis of empirical objects. The problem with his conception is apparent when considering the
transcendental ideality and empirical reality of space and time in which the transcendental
arguments for the necessity of space and time as pure forms of intuition are to guarantee their
empirical reality with regard to all empirical objects.
65
66
nature, such as the fall of a stone, are resolved into their elements and the forces
manifested in them (5:163). This is applied to ethics by a similar reection on
the moral dispositions of our nature encountered in experience:
We have at hand examples of reason judging morally. We can analyze them into their
elementary concepts and, in default of mathematics, adopt a procedure similar to that
of chemistry the separation, by repeated experiments on common human understanding, of the empirical from the rational that may be found in them and come to know
both of them pure and what each can accomplish of itself; and in this way we can
prevent on the one hand the errors of a still crude, unpracticed appraisal and on the other
hand (what is far more necessary) the leaps of genius by which, as happens with the
adepts of the philosophers stone, without any methodical study or knowledge of nature
visionary treasures are promised and true ones are thrown away. In a word, science
(critically sought and methodically directed) is the narrow gate that leads to the doctrine
of wisdom. (5:163)
Kant holds that we have consciousness (Bewutsein) and are aware (bewut)
of the moral law because reason prescribes it to us. This claim is central to
understanding Kants fact of reason, which will be covered in Chapter 3.
Empirical consciousness of the moral law and the general awareness of
moral principles in moral experience do not yet make them empirically real. If
the moral duties are products of the empirical moral agent, they would have to
be considered empirically ideal, that is, dependent upon the human mind. But
if they are grounded in something other than the empirical moral agent, they
could be considered empirically real. Kants division between pure and empirical as two kinds of principles with different grounds points to the possible
67
ground for an empirical realism of moral principles. If the moral principles are
understood as pure because they are grounded in reason, then the status of
reason itself will determine whether these principles are empirically real or
ideal. If reason can be shown to have its independence from empirical moral
agents, then moral principles would have empirical reality. The possible
transcendental ideality of moral principles would similarly depend upon
whether reason can be conceived as independent of the transcendental
moral agent.
The importance of these observations lies in their ontological signicance.
Our awareness of the moral law and moral principles is in empirical consciousness. Empirical consciousness is the subject matter of empirical psychology,
which is part of the theoretical study of nature. Kant believes that causal
determinism is true of empirical psychology, even if we are unable to know
any of the causal laws. Moral principles should, then, be available for use in
causal explanations of human beings in nature. Further, if the moral principles
are available in empirical psychology, the faculty that causes us to be aware of
them by prescribing them to us could be included in empirical psychology. If
so then there is a strong basis for a claim that reason as well as moral principles
are empirically real and subsequently for a robust conception of moral experience that includes the role of practical reason in actual agent deliberation and
causal determination. In addition to these systematic connections, this picture
is supported by some direct textual evidence such as the identication of
a causality of reason in the determination of the will as one of the natural
causes (A803/B831) and the claim that a rational being has the capacity to
act in accordance with the representation of laws which would have to be in
the empirical mind (4:412). If moral principles are available to moral agents
through moral experience, there is reason to call them empirically real. In
addition, if the agents moral actions are caused by reason in a way that can be
understood in empirical psychology, then the property of rational causation of
moral agents would be empirically real as well. In both cases, they could be
understood as transcendentally ideal as well if the transcendental conception of
moral agency also required them.
Kant also provides reasons for casting some doubt on this robust understanding of moral experience as an empirical moral realism. He claims that we
are never aware of the actual causes of our actions (Groundwork, 4:407). Also,
he sharply separates a genuine self that acts on the basis of reason from an
appearance of the self that is determined by desires and inclinations, suggesting that reason can play no part in appearance (Groundwork, 4:457). And he
strongly rejects any attempt to explain freedom using psychological determining grounds in time (Critique of Practical Reason, 5:96). In these cases, Kant
appears to be denying any kind of empirical realism for moral explanations of
action. Kant need not be denying empirical realism for our awareness of moral
68
69
Similarly, Kant uses the language of theoretical and practical reason in making
the division between theoretical and practical philosophy. Since philosophy is
the activity of reason, discussing them together is unproblematic.
Further, theoretical reason and practical reason are, for Kant, not two separate
faculties but one and the same faculty of reason. Kant uses the terms theoretical
reason and practical reason as short for the different uses by one faculty. He
makes this clear in the second Critique when he says it is still only one and the
same reason which, whether from a theoretical or practical perspective, judges
according to a priori principles (5:121; see also 4:391). In the many different
passages referring to theoretical reason, speculative reason, and practical reason,
Kant clearly means to refer to those uses of one faculty of reason. Any mention of
practical, speculative, or theoretical reason ought to be taken to mean these uses
or aims.
Given these clarications regarding the apparent identity of theoretical and
speculative reason, and the real meaning of the different faculties of reason as
different uses of one faculty, the contrast between practical reason and speculative reason becomes approachable. Kant appears to present four different explanations of the contrast.
A. Is/ought: Perhaps most famous contrast between the two is that the speculative aim of reason is to discover what is, while the practical aim of reason
is to tell us what ought to be (A633/B661; A802/B830; A806/B834; A845/
B873).
B. Nature/morality: Reason in its speculative use is said to investigate nature,
while reason in its practical use investigates morality. Kant presents such a
contrast when identifying a metaphysics of nature and a metaphysics of
morals (A841/B869; 4:3878; 5:545; 5:170).
C. Given/creating: The third contrast between speculative and practical reason
lies in their relation to whatever objects they consider, not necessarily in
the objects themselves. Specically, practical reason is said to create its
objects, while theoretical reason concerns objects somehow given to it
(A550/B578, 5:15; 5:5455; 5:89).
D. Knowledge/action: Speculative reason concerns knowledge while practical
reason concerns the determination of the will to action. Not the objects but
the method and purpose of application of reasons powers and principles is
the difference (A327/B384; A365; A475/B503; A803/B831; A841/B869;
5:1920; 5:456, 5:89).
Kants rst way of contrasting speculative and practical (A), using the is/ought
dichotomy, fails to clarify his intended meaning and actual use of the contrast.
The fame of this denition is rivaled only by its obscurity. The meaning of the
is is fairly straightforward. It is important to keep in mind that whatever the
practical turns out to be, it will not include ontological claims. But what does it
70
mean to say that something ought to be? What is the nature of that which
ought to be? In one way, this contrast falls under (C), for Kant says that the
object which ought to be is the moral world which is created by practical
reason through human actions (A808/B836). This moral world which ought to
be is created only to the extent that practical reason is actualized in the moral
agents acts. In another way, the is/ought contrast is like (D) because practical
reason is to determine our actions in accordance with that which ought to be
(A802/B830). The ought in this sense has meaning only as an imperative
commanding the will to action. Because of this ambiguity, I will not treat
Kants references to is and ought to be as comprising a distinct meaning of
the difference between the practical and speculative, although it remains
important as a summary of their fundamental difference.
The contrast between the sets of objects of nature and morality (B) is more
successful as a characterization of Kants intended meaning and use of the
contrast but by itself, insufcient to explain the unique characteristics of
the practical. This contrast can be construed in two ways related to different
characterizations of object, rst as sets of things and second as subject matter.
In the rst, one would assign the theoretical to one set of substantial objects,
perhaps appearances, and the practical to another, perhaps things in themselves.
Besides being limited by relying on a two-world interpretation of transcendental
idealism, this simple kind of contrast also does not t Kants overall theory. The
two uses of reason cannot have completely different sets of objects because they
share concern with at least some of the same objects. On the one hand, God,
immortality of the soul, and freedom of the will, can be approached, Kant says,
both speculatively and practically (A796/B824; 5:45), and, on the other hand,
the human power of choice is treated, as he says in the First Introduction to the
third Critique, as free in practical philosophy and as a cause in theoretical
philosophy as cognition of nature . . . the practical part of a philosophy of
nature (20:197). Further, Kant himself stated that speculative and practical
reason concern the same set of objects of possible experience. As I noted in
the rst section of this chapter, in the Critique of the Power of Judgment, he
distinguished the concepts of domain and territory: the domains of speculative
and practical differ in that they each legislate different sets of concepts, but their
territories are identical because both sets of concepts concern the objects of
possible experience as appearances (5:174). The difference between speculative
and practical cannot lie in distinguishing any corresponding sets of things,
because these sets are coextensive or at least overlap.
Alternatively, the set of objects in (B) can be understood as different domains
construed not as substantial objects but as merely different subject matters. In
the Collins lectures on metaphysics, Kant insists that theoretical and practical
have distinct objects: Practical philosophy is practical not by form, but by the
object, and this object is free acts and free behavior (27:243, Kant uses the
71
latinate Object rather than Gegenstand). The object of practical philosophy is not
any kind of substantial object but is instead said here to be free acts, which as
noted in the previous paragraph have their effect in nature and can be understood
in relation to natural causes. The free acts, however, are a different subject matter
than objects in nature. This interpretation is the better way to construe (B), but it
also turns out by itself not to be specic enough to be helpful since it merely says
that they have different domains, one nature and the other morality, without
characterizing the basis on which to separate them. The quote identies morality
as free acts and free behavior which will be identied as the domain after more
assessment of the nature of the practical.
The third contrast (C) does not itself specify a domain but only a relation
between reason and its objects that can ground a division of subject matter.
Theoretical reason is given its objects by intuition while practical reason
creates its objects. They could both concern the same things in appearance
provided that those objects are subject to both kinds of relations. Practical
reason can be said to create objects in two ways: in the rst way, practical
reason does not actualize its objects; that is, it does not create those objects as
empirical objects, but creates the concept of that object (in some cases, these
are the ideas of reason18); in the second way, practical reason does actualize its
objects in the empirical world. Only in the second way can practical and
theoretical reason relate to the same objects in different ways. In the Dialectic
of the second Critique, Kant identies the objects created by practical reason
as actions in the sensible world and the highest good, which is the nal end
of our right acts (5:134). In particular, the performance of actions by the will
creates objects in the realm of appearances that are then subject to theoretical
reason, which is given the effects of the will in appearances to integrate into
knowledge of nature. Regarding both objects of practical reason our right
acts and the highest good practical reason can be said to create the concepts
of those objects. Through the categorical imperative, moral agents conceive of
the right acts they ought to perform. Likewise, in conceiving of the highest
good, practical reason creates the concept of happiness necessarily connected
with virtue (5:110111). But using these concepts, practical reason seeks to
actualize right acts and the highest good in the world; these activities create
practical objects corresponding to those concepts. Particular right acts of
course occur in the empirical world and are thereby actualized. The highest good,
18
I use the term concepts to discuss that which reason can create because the particular actions
guided by practical reason would be understood in conceptual terms. Certainly, any empirical
ends that would be the goals of actions in accordance with practical reason would be conceptualized as well. When Kant labels the concepts of reason as ideas in the Critique of Pure
Reason (A311/B368), he is not considering practical reason and relations to actions at all. Some
of the concepts of objects to be created by practical reason are ideas, in particular, the
highest good.
72
Kant says, is practical, that is, to be made real through our will (5:113).
Practical reason creates not only the concepts of right acts and the highest good,
it also then uses them as a guide in specic actions in the sensible world which
perhaps work toward but do not fully actualize the highest good.
The claim that practical reason creates its own objects resembles the claim
of constructivists that various principles, properties, or objects are constructed.
Of the specics mentioned by Kant, the actualization of the idea of the highest
good would have to be constructed in an obvious and unimportant sense of
being the result of actions which themselves are of course created as empirical
events in another non-controversial sense. The concept of the highest good is
created by reason when nite rational agents look for a way to reconcile their
moral obligations with their desire for happiness. As such, it stems from the
nature of moral agency rather than any decision procedure and is not constructed. Besides the idea of the highest good, the postulates of practical reason
are ideas created by practical reason (irrespective of the status of their objects).
The fourth contrast (D), between knowledge and action, is not about objects
or relations to objects but about two different ways that reason can comport
itself. This contrast is related to the previous one because reasons creation is
done in part through action and reasons being given its objects through
intuitions provides reason with material to comprehend. The comparison is
not perfect since knowledge can involve creation by reason of speculative
ideas in order to aid the understandings cognitions (A702/B731) and Kant
sometimes talks of practical cognition (5:20). The key difference is that in
both knowing and action, reason has principles or ideas such as the moral law
or the idea of God, but it uses them for different purposes. The two purposes
relate to the is/ought dichotomy already discussed.
This contrast between the theoretical and practical ts best with the domain
of philosophy as a whole and also best justies the internal division within
philosophy that provides its basic structure. Philosophy is identied as the
science of the relation of all cognition to the essential ends of human reason,
of which the highest is the nal end of human beings (A83940/B86768). To
reach those ends, reason must prescribe actions, but in order to act rationally,
reason must know the world in which it acts. Theoretical philosophy provides
knowledge essential to the actions governed by practical philosophy. The basic
division of philosophy into theoretical as knowledge and practical as action
stems from the domain of philosophy as a whole. This result meets Kants
requirement for a rational system of knowledge in which one idea determines
the position of the parts a priori (A832/B860).
I will show here that this fourth contrast also best ts with the identication of
the specic domain of practical philosophy as free acts. Action is the fundamental
characteristic of practical philosophy. Still, the other three contrasts, each also
reect aspects of the practical even if they are not fundamental. In focusing on
73
action, practical reason creates certain concepts as well as rules for free action.
Actions are viewed as created rather than given and so not initially identied as
part of nature to be known; hence, they can be considered as the distinct objects
(as subject matter) of morality as opposed to given objects in nature to be
cognized. At the same time, actions do occur in nature and are secondarily
integrated into the causal series in nature as something given through experience.
Qua actions to be performed, they embody the ought rather than the is of
theoretical philosophy.
Free acts as the domain of ethics
Kant identied free acts as the domain of ethics, that which is to be systematized. He does insist upon the systematization of ends in the kingdom of ends
in Groundwork II, and the systematic relation between virtue and happiness as
part of the highest good in the Critique of Practical Reason, but neither of
these constitutes the complete domain of the practical.
The most explicit argument that Kant gives for the claim that free acts
constitute the domain of ethics is in his course lectures on ethics. He apparently
always began his lectures on practical philosophy by distinguishing the practical from the theoretical (Powalski 27:96; Collins 27:243; Kaehler 3; Mrongovius 29:597; Vigilantius 27:479). For simplicity, I will refer to the Collins
lectures of 178485, which provides the clearest formulation of the distinction
between theoretical and practical philosophy and the domain of the latter, but
the other lectures and even the Groundwork Preface make the distinction and
at least hint at the domain.
The initial denitions Kant provided for the theoretical/practical distinction
on the rst day of his Collins ethics course provides a fair summary of Kants
view of the division:19
All philosophy is either theoretical or practical. Theoretical philosophy is the rule of
knowledge, practical the rule of behavior in regard to free choice. The difference
between theoretical and practical philosophy is the object. The one has theory for its
object, and the other practice. (Collins, 27:243)
19
When reading Kants lectures, one must always be wary that Kant sometimes follows his
textbook closely, so attributing to Kant any structural and denitional passages from the lectures
without checking the text as its possible source is risky. In this case, the distinction clearly stems
from Kant himself because the textbook he was using, Alexander Baumgartens Initia
philosophia practicae primae, begins by discussing practical philosophy as a part of philosophy
without distinguishing theoretical philosophy by name; neither does Baumgarten dene the
domain of practical philosophy as Kant does: for Baumgarten, philosophy in general is the
science of our knowledge of the properties of things, and accordingly practical [philosophy]
is the science of our knowledge of the obligations of human beings (19:9).
74
Kant describes the practical as the rule of behavior in regard to free choice.
This reveals the link between the nature of practical philosophy and freedom.
But the specic conception of freedom is not obvious and must be examined.
The passage mentioned here claims that theoretical and practical philosophies have distinct objects. Kant goes on to clarify that to say that they have
distinct objects contrasts with a distinction one might make in which theoretical and practical philosophy differ only according to form. That is, given a
single object of study, theoretical philosophy can relate to it in one particular
manner, namely, nding concepts of the objects, while practical philosophy
can relate to the same object in another manner, namely, exercising knowledge of the objects (27:243). Kant gives the example of theoretical and
practical medicine, the former presumably consisting in knowledge of health
and bodily conditions, the latter in application of that knowledge to maintain or
alter the state of health. Both share the same object. But this difference
according to form is not what Kant intends by his distinction. Rather, he
insists that theoretical and practical have distinct objects (Object): Practical
philosophy is practical not by form, but by the object, and this object is free
acts and free behavior. The theoretical is knowing, and the practical is behaving
(27:243).
The contrast with theoretical philosophy provides the support for a characterization of freedom at issue. Theoretical philosophy is about knowing what
exists. The practical must have a distinct object, so its object must be other
than knowing what exists. The practical must concern behavior. But since the
behavior at issue cannot be known for then, it would be the object of
theoretical philosophy the behavior in question cannot be understood empirically, as an event whose causes can be known. A study of causes of behaviors
would fall under psychology as the theoretical science or knowledge of the
mind. Practical philosophy must abstract from possible particular causes of
behavior. It must look at an act not as something to be known but as something
to be decided.
But this is the same as saying that all possible choices are open in deliberation as possible courses of action. The individual must view herself as not
constrained by any determinant of her decision-making. Kant puts it this way
in the Collins lecture:
If I abstract from the particular matter in hand, the philosophy of behavior is that which
gives a rule for the proper use of freedom, and this is the object of practical philosophy,
without regard to particulars. So practical philosophy treats the use of free choice, not in
regard to particulars, but independently of all of them. (27:243)
75
philosophy. Here the direction Kants argument takes can be seen; I had earlier
noted that the domain of a science could be an empirical concept (as matter for
physics), but now that option is precluded for the domain of practical philosophy.
The specic meaning of free choice must still be claried. As a starting point
for ethics, one cannot assume that it can be dependent on theoretical philosophy, so the nature of free choice must be internal to practical considerations.
This view of freedom is the nature of the free power of choice facing decisions
for actions from a perspective abstracted from any consideration of determination in nature.
An outline of an argument in a Reection most likely from the mid-1770s that
Kant labels Metaphysical Concept of Morality begins with this same conception of freedom of the power of choice:
We do not abstract the concept of freedom from experience. When we want to act, we
consider the prospective action as completely problematic with regard to the present
moment, and the ought is a condition of the consensus of the prospective action with
reason, which is thus not at all in a pre-determined interconnection with appearances,
i.e., with nature. (R6854; 19:180)
76
77
Part II
82
have priority is to say that reason may insist on these postulates although
speculative knowledge can neither conrm nor deny them. The second, implicit conception that emerges when Kant discusses particular examples of
practical action, I call the priority of action. The priority of action involves
the following claim: when acting, humans are confronted with the categorical
imperative as a command obligating them to act in certain ways. To say that
the practical has priority is to say that speculative metaphysics is irrelevant to
practically oriented action, so that even a speculative denial of the postulates
would not affect practical reason. Kants discussions of the relation between
practical and speculative in fact contain these two conceptions. I will argue that
the former and not the latter conception actually relies on a higher priority of
theoretical reason. I will also argue that the latter conception, Kants implicit
view, is the only one that genuinely asserts the priority of the practical and best
ts with his conception of the nature of practical reason as discussed in my
previous chapter.
The priority of belief
In analyzing the concept of priority, one must rst note that Kant never directly
claims that practical and speculative reason contradict one another. The nature
of priority cannot lie in holding some proposition P (asserted as true by
practical reason) to be true and holding some proposition not-P (asserted as
true by speculative reason) to be false. Kant holds that a condition of the
existence of reason is that it never contradicts itself, and since speculative and
practical uses of reason are one and the same faculty of reason, one can never
be in a position to choose between rival claims to the truths of P and not-P
(5:120).
The basis for a need for priority lies elsewhere in what Kant calls the
interests of practical and speculative reason. Kant denes interest of a
faculty of the mind as a principle that contains the condition under which
alone its exercise is promoted (5:119). He says that the interest of the
speculative use of reason consists in the cognition of the object up to the
highest a priori principles while the interest of the practical use of reason
consists in the determination of the will with respect to the nal and
complete end (5:120). These interests match the contrast between speculative and practical as knowing and acting (the determination of the will).
As principles, these interests are known through concepts1 and would
constitute a priori goals of the uses of reason. These interests as goals mean
that speculative reason is furthered only when working toward the highest
83
Paul Abela argues that the proper understanding of the primacy of the practical is one that avoids
what he calls the integrationist model in which one of speculative and practical reason
whichever is said to have primacy is required to integrate the relevant products of the other
(Abela 2002b). He discusses the paper on which this chapter is partly based, so I will respond
here. On my account, the priority of action is not integrationist in this sense, and I take it to be
just as much Kants view as the priority of belief. Nonetheless, Abela chooses to discuss my
account of the priority of belief including the higher priority of theoretical reason it implies as
representative of my view because in that article, I label it as Kants ofcial position (In this
book, I drop that label and use only explicit and implicit for the two views). The priority of
belief is integrationist, as Abela notes. He fails to stress that the integrationist need in my
interpretation is based on the need of theoretical reason to know. In the process of trying to know
what comes before it, it must try to nd consistency among the principles and concepts and
objects because consistency is a mark of a set of true propositional claims. I do not think that all
of Kants discussions of practical reason include the integrationist claim but only those that
include such a role for theoretical reason, and only because of the nature of theoretical reason.
The priority of action is not integrationist.
84
obvious sense in which reason creates new objects through particular acts of
will that result in actions and alterations in the sensible world, but they are
unproblematically incorporated into the knowledge of nature as events in space
and time. The priority of practical reason concerns its postulation of freedom,
immortality, and God. For practical reason to require priority, it must be the
case that freedom, immortality, and God do not already fall under the interest
of speculative reason; otherwise, practical reason cannot be said to create these
objects. But Kant himself admits in many places that freedom, immortality,
and God are already objects of speculative reason (B395n; A798/B826; 5:34;
5:134). Indeed, the Dialectic of the rst Critique concerns these objects
approached theoretically, as the triad soul, world, and God. It cannot be the
case, therefore, that the priority of the interest of practical reason lies in its
creation of some objects not already subject to speculative reason.
The second possible conict between the interests of the two uses of reason
is that determination of the will to action presupposes some speculative
assertion neither presupposed nor proved by the speculative use of reason,
yet pertaining to objects of speculative reason. In this case, freedom, immortality, and God are to be subject to the interest of speculative reason; practical
reason adds only the assertion that we as moral agents must believe in these
objects. Speculative reason must then integrate this belief into its system of
cognitions.
This latter understanding of the conict between the two interests of reason
is Kants explicit position. Kant concludes that speculative reason must yield
to the demands of practical reason and take up the assertion of certain
postulates freedom, immortality, and God for integration with other beliefs
and cognitions (5:121).
The priority of belief is thus Kants explicit position given in the section of
the second Critique on the primacy of practical reason. This conception of
priority, however, seems at odds with Kants discussion of the nature of the
distinction between speculative and practical. Determination of the will to
action, which is the interest of the practical use of reason, seems tenuously
connected to the notion of belief. Further, Kant does not describe the priority
of the practical in terms of the creation of new objects. Hence, Kant cannot
have in mind the contrast between practical and theoretical as either given/
creating or knowledge/action.
Kants explicit doctrine must presuppose a higher priority of speculative
reason as reasons interest in knowing about given objects.3 Kant is operating
with an unstated assumption that reason must have interest in certain objects,
3
Marcus Willaschek has interpreted the primacy of the practical and the theoretical more narrowly
(Willaschek 2010, pp. 17375). Drawing on the material in Kants second paragraph of the
section on primacy, he understands the primacy of speculative reason only in terms of setting the
85
namely, the objects of the postulates, not for determining right acts but for
understanding the relation between right acts and the rest of our experience.
Hence the postulates appear not as determinants of right acts in the process of
deliberation but only when reecting on the nature of morality, the fact that
right acts do occur, and the fact that agents believe those acts ought to occur.
And for this reection, Kant must be assuming that speculative not practical
reason is at work, and hence that there is a higher priority of speculative
reason. This assumption is a theoretical, not practical, assumption. Moral
agents require knowledge of the non-impossibility of the existence of God,
freedom, and immortality, in order to understand how their right acts can be
related to their experience as a whole. The acts themselves are still determined
by practical reason alone.
One might instead construe this situation as conrming the priority of
practical reason. If practical reason is the use of reason that demands acceptance of its postulates by speculative reason, then one can claim that practical
reason has priority, because speculative reason is constrained to modify its set
of beliefs in order to incorporate the postulates of practical reason. In this way,
practical reason would have priority even concerning beliefs. I believe, however, that such a relation nonetheless reveals the priority of speculative reason
in carrying out the incorporation of all beliefs into one whole. First, these
objects (the postulates) are themselves also objects of speculative reason in the
Dialectic of the rst Critique. Thus, speculative reason is forced to deal with
their possibility as well as their possible integration into a complete set of
beliefs already on purely theoretical grounds, so practical reason does not have
priority in solely insisting that speculative reason consider these objects.
Second, practical reason cannot demand that speculative reason accept beliefs
that the latter has already determined to be false, since then integration of all
beliefs would fail as the postulates turn out to be inconsistent with other results
of speculative reason. Since it is speculative reason that does the required
integrating of beliefs and that could presumably reject beliefs stemming from
limits to what the practical can assume for its purposes. This ts Kants text at that point but does
encompass the larger issue that arises after the speculative is constrained to incorporate practical
reasons beliefs. The larger priority of the speculative lies in its interest in knowing everything,
which includes the relation between the beliefs of practical reason and the knowledge speculative
reason already has at hand. The primacy of practical reason does not entail a surrender by
speculative reason but only that speculative reason must incorporate non-speculative content.
Nathan Rotenstreich also notes as I do that Kants view of the priority of practical reason seems
to rely on a higher priority of theoretical reason (Rotenstreich 1965, p. 130). He argues that the
postulates involve the validity of existence claims and hence must be theoretical. He does not,
however, specically link this claim to the interest of speculative reason in knowledge of objects
but focuses on his claim that validity is a theoretical category. Rotenstreich also uses his analysis
to reject Kants claims about priority of the practical rather than to search for Kants other
conception of priority as I have.
86
practical reason should they have proven false on purely speculative grounds,
speculative reason can be said to have priority.
This priority of speculative reason is exhibited in other passages in Kants
works on ethics, revealing that Kant must have understood that speculative
reason has interests in many issues that arise from practical activity. Kant notes
the interests of the theoretical in some seemingly practical issues such as
freedom in several places, although nowhere does he emphasize it. In the
Groundwork, he quite explicitly contrasts the interest of speculative reason
with the lack (in this case) of interest of practical reason in the question of
freedom: The settlement of that controversy [about freedom] does not belong
to it; instead it only requires of speculative reason that it put an end to the
discord in which it entangles itself in theoretical questions, so that practical
reason may have tranquility and security from the external attacks that could
make the land on which it wants to build a matter of dispute (4:4567). Kant
implies here that the question of freedom needs to be answered for indirect
practical purposes, namely, the defense of morality from outside attacks, rather
than the direct practical interest, namely, any use of the idea of freedom to
determine the will to right acts. And the type of reason that is to answer the
question of freedom is speculative not practical. Speculative reasons interest
in knowledge includes put[ting] an end to the discord in which it entangles
itself regarding the possibility of freedom. But the role of speculative reasons
interest in knowledge about issues such as freedom raised in the investigation
of practical philosophy is not limited to serving indirectly any practical
purpose of defending morality from outside attacks. Rather, Kant says in the
Preface to the Groundwork, speculative reason has its own interest in knowing
what it can about human morality: A metaphysics of morals is therefore
indispensably necessary, not merely because of a motive of speculation for
investigating the source of the practical principles that lie a priori in our
reason but also because morals themselves remain subject to all sorts of
corruption as long as we are without that clue and supreme norm by which to
appraise them correctly (4:38990, modied, emphasis mine). Notice that
speculative reason itself is said to have interest in investigating issues, in this
case, the source of the moral law, which we might otherwise assume belong
exclusively to practical reason.
The second Critique contains some passages making similar points. Kant
notes, for example, that speculative reason concerns itself with certain concepts raised by the practical: it is practical reason which rst poses to
speculative reason, with this concept [of freedom], the most-insoluble problem
so as to put it in the greatest perplexity (5:30). Speculative reason is put in
difculty because it must integrate beliefs required by practical reason with its
other beliefs. This integration is a speculative not a practical issue, a problem
of comprehensive and coherent beliefs not of direct determination to action.
87
See, for example, (Beck 1960, pp. 24244) and (Wood 1970, pp. 3868).
88
postulated, while our reason nds this thinkable only on the presupposition of a
supreme intelligence; to assume the existence of this supreme intelligence is thus
connected with the consciousness of our duty, although this assumption itself belongs
to theoretical reason (5:1256).
To assume the existence of God is not a duty because one cannot have duties
to believe theoretical matters but only to perform actions. The only duties
one has are those that follow directly from the autonomy of reason, namely,
the duties required by the categorical imperative (5:3033). Since these duties
involve working toward the realization of the highest good, one can be said
to have a duty to further the highest good. But any belief in an actual realization of the highest good involves the theoretical claim that God exists. Practical
action implies belief in God but does not conrm the existence of God. One
can be said only implicitly to believe that God exists because ones actions,
when analyzed, are seen to require the existence of God for the realization
of their ultimate aim. But even this conclusion reveals the higher priority of
theoretical reason because even this conclusion involves not determination
of the will to right acts but reection on the determination of the will to right
acts. What Kant calls the moral necessity of belief is a product of theoretical
reection not a prerequisite to action. Since it is not a prerequisite to action,
belief in the postulates is not a motive to action. Kant can thus avoid the
problem some see in his linking the highest good to moral action because the
link is theoretical rather than practical and thus independent of the determination of the will to right acts.
The priority of belief, Kants explicit version of the primacy of practical
reason, thus actually involves a higher priority of speculative reason. Only
when reecting on the circumstances and aims of moral action does one ask
about the prerequisites for realization of the highest good. This reection, itself
speculative, nds reason integrating concepts related to human moral action
with other theoretical matters.
The priority of action
Since the priority of belief actually involves a higher priority of the theoretical, genuine priority of practical reason over speculative reason can occur
only in the other conception of priority, the priority of action. The priority
of action better captures the interest of practical reason in determining the
will to action because only the priority of action involves use of the
categorical imperative. In acting, humans are aware of the categorical
imperative as a command obliging them to act in certain ways by adopting
certain maxims. To say that the practical has priority is to say that the
results of speculative reason are irrelevant to the determinations to action of
practical reason.
89
The priority of action, although never directly stated by Kant, seems to lie
behind many passages he uses to support various arguments. At times, he
separates the practical approach to action from any theoretical issues arising
from an attempt to integrate a conception of action with other beliefs and
cognitions. At other times, when he offers examples of right or wrong acts, he
consistently treats them independently of the theoretical considerations
involved in the priority of belief. When acting, all these passages seem to
say, a moral agent simply does not need to know whether or not speculative
reason precludes the possibility of freedom, God, and immortality. A review of
these passages will show that Kants thought did possess the essential elements
of the priority of action.
Kant hints at the priority of action in several places. In the Antinomy chapter
of the Critique of Pure Reason, he describes the irrelevance of theoretical
considerations for the practical task of doing and forbearing:
If a human being could renounce all [architectonic] interests . . . such a person would be
in a state of ceaseless vacillation. Today it would strike him as convincing that the
human will is free; tomorrow, when he considers the indissoluble chain of nature, he
would side with the view that freedom is nothing but self-deception and that everything
is mere nature. But now if it came to be a matter of doing or acting, then this play of
merely speculative reason would disappear like the phantom images of a dream, and he
would choose his principles merely according to practical interest (A475/B503).
The practical purposes of action trump speculative considerations of the possibility of freedom. When faced with a moral decision, the agent makes his
choice only according to the practical interest of determining the will to action
and not according to any concepts or beliefs of speculative reason. Kant does
not say here that an agent would stop reversing positions by settling on one
position within speculative reason; rather, it appears that the agent stops
speculating at all. A similar passage in the Canon states that the question of
freedom of the will does not concern the practical sphere . . . it is rather a
merely speculative question, which we can set aside as long as our aim is
directed to action or omission (A803/B831; see also 5:172).
His opening paragraph in the Critique of Practical Reason also seems to
assert this priority of the action: If as pure reason it is really practical, it proves
its reality and that of its concepts through the act, and all subtle reasoning against
the possibility of its being practical is futile (5:3, modied). The reality of pure
practical reason is proved not theoretically but in instances of its use. Practical
reason is here given priority because its actual use in action trumps any
speculation regarding its possibility or impossibility. Theoretical rationalizing
is futile when practical reason asserts its priority in action.
Note, however, that in these passages, Kant does not replace a theoretical
basis for inquiry into the possibility of freedom with a practical demand for
belief in freedom as the basis for theoretical inquiry. Had these passages taken
90
that approach, they would accord with Kants explicit doctrine of the priority
of belief. Rather, in both cases, Kant asserts that the theoretical questions
become irrelevant from the standpoint of the moral agent faced with a decision. The agent would choose his principles merely according to practical
interest, and practical reason proves its reality and that of its concepts
through the act independent of futile speculation. The agent would decide
on an action taking into account the dictates of the categorical imperative, not
any beliefs she might hold, because the categorical imperative and not any set
of theoretical concepts is the means by which practical reason determines the
will to action.
The priority of the practical as a priority of action also lies behind Kants
famous claim from the third section of the Groundwork that Every being that
cannot act otherwise than under the idea of freedom is just because of that
really free in a practical respect, that is, all laws that are inseparably bound up
with freedom hold for him as if his will had been validly pronounced free also
in itself and in theoretical philosophy (4:448). In this passage, unlike the two
quoted earlier, Kant attributes to the moral agent, an assumption of freedom:
when a rational being faces moral actions, she must assume that her decision is
free rather than determined by conditions beyond her control. But, Kant says,
her assumption that she is free sufces for the practical reality of her freedom.
As long as she considers herself as free, she is justied in treating herself as if
she were free in her decision. She is not, however, postulating freedom in any
theoretical sense. Kant intends here to distinguish freedom as valid for theoretical philosophy from freedom as sufcient for the practical interest.
Several examples Kant uses to illustrate various aspects of his moral theory
also illustrate the priority of action. In the Critique of Practical Reason, after
arguing that freedom and the moral law reciprocally imply one another, Kant
offers the example of a man commanded by his sovereign to provide a false
deposition who recognizes his freedom to refuse this immoral command. He
judges, therefore, that he can do something because he is aware that he ought
to do it and cognizes freedom within him, which, without the moral law, would
have remained unknown to him (5:30). This person recognizes his freedom
without hesitation when confronted with his sovereigns command to provide a false deposition. And later, after reviewing the possibility of a noumenally free decision underlying the entirety of ones phenomenal character, Kant
turns to conscience to show that our moral experience of placing blame on
ourselves is in agreement with this conclusion; since conscience involves no
theoretical reection but is in perfect agreement with the results of theoretical reection on the possibility of freedom (5:98), conscience must be connected with practical reason rather than theoretical reason. But it appears that
conscience does not require any conrmation of the possibility of freedom
by the theoretical reason, rather, it independently conrms the results of
91
An anonymous reviewer of this book suggests that one might take these examples as evidence
for the theoretical claim of freedom, under the principle that ought implies can, and that thus,
these examples are not limited to action. I discuss the kind of freedom associated with ought
implies can in Chapter 6. Here I will point out that the kind of evidence in these examples
remains at the rst-person practical level. The agent judges that he can do the required deed. The
agents practical belief that he ought to do x does not provide evidence for a theoretical claim that
he can; rather, the principle of ought implies can would only lead the agent to believe that he can.
Any inference from ought to can from the agents perspective must retain that practical status.
92
93
94
explain its possible relation to other objects. Hence the concept of the highest
good, while created by practical reason, could be considered an object for
theoretical reason. In fact, I claim, this is precisely what occurs in Kants
investigation of morality and precisely the nature of the Critique of Practical
Reason. The second Critique is not demanded by the interests of practical
reason but only by the interests of theoretical reason. Practical reasons sole
concern is with the determination of actions.
At least some of the works or parts thereof that we regard as Kants practical
philosophy, therefore, are themselves works of theoretical philosophy about
the practical use of reason. Given that Kant was more a reective philosopher
than an active agent in world or local affairs, we might expect as much. The
philosophical issues surrounding a practical point of view are treated theoretically by Kant. But, ironically, one of those theoretical results may be an
understanding of the priority of the practical as the priority of action, Kants
implicit doctrine, which allows practical reason to function independently of
any results of theoretical reason.
Action and the nature of the practical
Given that Kant had two different senses of the way in which practical reason
has priority, it is legitimate to call both of them Kantian. The priority of
belief preserves the unity of reason at the cost of denying genuine priority to
the practical; the priority of action preserves genuine priority of the practical at
the cost of denying the necessity of a unity of reason. Adopting either will
therefore have both positive and negative consequences for interpreting Kant.
Both capture elements of Kants philosophy. The interpretation I will give
shows that they can be reconciled even if they do not t at all points. This
result has implications for the realism/idealism debate, given that the priority
of action places so much emphasis on the perspective of the moral agent.
The two views are reconcilable in an interpretation of Kant that stresses the
nature of the practical as concerned only with questions of free action, of
determining what moral agents ought to choose, and not with ontological
questions at all. The work in my previous chapter detailing the nature of
practical reason bears fruit here. Practical reasons focus on action rather than
knowledge shields it in practice from any ontological concerns. Practical
reason itself is exhibited only in the moment of decision. As Kant puts it in
the second Critiques Preface, If as pure reason it is really practical, it proves
its reality and that of its concepts through the act, and all subtle reasoning
against the possibility of its being practical is futile (5:3, modied, my
emphasis). The phrase through the acft translates durch die Tat, which
could also be translated as through the deed. Whether the precise act is that
of a moral agent deciding upon an action or the act of reason as a call to follow
95
a certain principle (to be discussed in the next section), the key point is that
practical reason operates in agents actions. Practical reason arises when the
agent is pondering a decision to act and thus within the agent-perspective
understanding of free acts as the domain of ethics. Any and every concept or
idea that practical reason brings to bear on that decision must be understood
from the practical point of view of determination to action and not from the
theoretical view centered on knowledge of what is. The agents perspective, so
to speak, is framed by the action of practical reason. This agent-perspective lies
at the heart of moral experience.
Theoretical reason enters the practical picture only when the agent steps
back from that perspective and reects upon it. This is the step in the priority of
belief in which the demands of practical reason are given to theoretical reason
in order for the latter to incorporate them into a theoretical comprehension of
experience. What is given to theoretical reason, then, is (in addition to the
action as event in nature) the fact that agents do take up a certain perspective
that is determined at least in part by practical reason. Theoretical reason is not
given a set of objects such as God that it must then incorporate into the
ontology of the world but is given a set of principles and ideas that practical
reason uses in determining action. I will discuss this issue in relation to the
nature of the postulates directly in Chapter 5. Theoretical reason does not
obtain any ontological claims from practical reason but only the principles and
ideas that constitute the agent-perspective. There is, then, no possible clash
between the ontology provided by theoretical reason and that demanded by
practical reason because the latter demands no ontology.
At the heart of Kants moral theory, then, is the claim that practical
philosophy must be separated from theoretical philosophy. The basis of his
claim that moral theory must seek an a priori moral law surely a touchstone
of Kantian morality lies in the nature of morality as practical philosophy
concerned with a rule for free action in contrast to and without reference to
theoretical philosophy concerned with knowledge of what exists. This basis
must always be kept in mind when assessing particular claims in Kants ethical
theory. In general, one must always be skeptical about anything that looks like
an existence claim in Kants practical philosophy.
Another way to see the importance of this interpretation is that Kant
simultaneously allows one type of naturalism while rejecting another type.
He rejects, quite clearly, a methodological naturalism that would base
morality on something empirical, such as an empirically derived claim
about human nature. The agent-perspective view on decision-making precludes anything other than a priori practical principles. No particular empirical fact about human beings, or about the human environment, or about
nature at all, can serve as a basis for investigation into the ultimate principle of morality. This ultimate principle of morality, then, must be a priori
96
rather than a posteriori, and must be a necessary moral law rather than just a
contingent practical principle.
While insisting on this methodological anti-naturalism, however, Kant also
shows that any resulting morality must not make existence claims; Kant
thereby at least allows for a metaphysical naturalism. The domain of the
practical must be sharply distinguished from that of the theoretical, and only
the theoretical is allowed to be about knowing what exists. Kantian morality,
then, must be prima facie compatible with a naturalistic metaphysics.
These rather broad conclusions are quite general and could be misinterpreted. The separation does not mean that practical philosophy never concerns
things that exist. Such a conclusion would be absurd, since practical laws
govern human beings, who exist, in their relations with each other and with
things, which exist. On the one hand, Kant himself insists that morality must
be applied to human beings as they exist in nature (this is the study of
practical anthropology, 4:388). Here, however, knowledge from theoretical
philosophy as a study of human beings in nature is brought into the practical
from outside, so to speak, in order to provide the material needed to apply the
pure a priori ethics to actual human beings in nature. On the other hand,
however, the parts of his pure moral theory that are not borrowed from an
empirical study of nature yet appear to make existential claims I am thinking
in particular of the postulates of freedom, immortality, and God must be
capable of being interpreted in a way that avoids any existence claims.
The interpretation I have just given will be eshed out in more detail. In the
remainder of this chapter, I want to show how the best interpretation of Kants
fact of reason ts with the role of the priority of action.
The fact of reason and agent-perspective
The interpretation of the fact of reason will deepen the discussion of the
difference between the empirical and transcendental agent introduced in
Chapter 1. Empirical agents exist in nature as individuals who undergo a moral
experience. As in Kants examples, they are faced with particular decisions
regarding moral (and of course non-moral) situations. From their agentperspective, they must consider themselves able to make that decision freely,
that is, they act under the idea of freedom. The priority of action holds that this
perspective is self-validating and need not concern itself with the metaphysical
question of whether the agent really is free. The transcendental moral agent is
the conception of agency that results from a transcendental deduction of
morality, the description of the a priori conditions for there being empirical
moral agents. This agent need not be a transcendent being who in some way
exists and acts independently of nature in space and time unless such a nonnatural ontology of agency is required by the specic a priori conditions of
97
Pauline Kleingeld offers a very similar interpretation that stresses the agent-perspective, calling
it radically agent-based (Kleingeld 2010, p. 72). With some minor differences, we share this
general interpretation. The largest difference is the way that I relate the interpretation to the
empirical/transcendental division.
98
elements of Kants ethics. Finally, I will examine the kind of freedom that the
fact of reason is said to reveal to us.
Categorical imperative and moral law
Although the fact of reason plays an extremely important role for Kant, he
never clearly dened it, forcing commentators to attempt to elucidate it in light
of different, even conicting, references.7 I will argue here that the fact of
reason consists of the categorical imperative as our consciousness of the moral
law. This denition involves Kants distinction between the moral law as such
and the categorical imperative.
Kant discusses this distinction in the Critique of Practical Reason and the
Groundwork. The moral law holds for purely rational beings who would use it
as the sole principle of action because they lack any other possible determinants for action. Purely rational beings have no sensuous impulses, so would
always act rationally. They would conceive of the moral law not as a command
but as a description of their principle of action. We, humans, however, can
guide our actions not only using laws of reason but also by incorporating
sensuous needs and desires. Unlike purely rational beings, we do not automatically guide our action according to the moral law. We humans encounter the
7
The initial assessment in English is by Lewis White Beck in his Commentary on Kants Critique
of Practical Reason (Beck 1960, p. 166f), where he quotes four passages from the text and
concludes that the fact is either consciousness of the moral law, the moral law itself, or
autonomy. Beck cites additional passages from the second Critique and further identies four
additional references to the fact of reason after the second Critique that, he says, fall in line with
the earlier ones. See also (Beck 2002). Darlei DallAgnol quotes eleven passages in which Kant
mentions the fact in the Critique of Practical Reason and concludes that it has a family of
meanings (DallAgnol 2012a, p. 115). Other commentators identify a single meaning for the fact
of reason, generally taking it to be the consciousness of the moral law. Dieter Henrich takes it to
be simply the demand of the good (Henrich 1994, p. 83); Henry Allison, the consciousness of
standing under the moral law (Allison 1990, pp. 23139); John Rawls, that in our common moral
consciousness we acknowledge the authority and direction of the moral law (Rawls 1999c,
p. 102); Guido de Almeida, the consciousness that nite rational beings have of the moral law as
a categorical imperative (De Almeida 2012, p. 147). Two quite different understandings of the
fact are given by Marcus Willaschek, to be discussed later on, and Ian Proops, who argues that
the fact is both the moral law itself and our consciousness of it, although he stresses the former
(Proops 2003). What makes Proops interpretation different is that he interprets Kants German
term Faktum in light of a legal proceeding in which the factum (the Latin term) is the thing
identied as the matter for a legal judgment, as opposed to the juris which is the legal basis for
the judgment. Kant uses this terminology in the Critique of Pure Reason, in his discussion of the
difference between the possession of a concept such as causality and the justication for that
concept, the latter being the subject of the transcendental deduction (A8487/B11619). Proops
claim is weakened when one notes that in the second Critique, Kant uses the German term rather
than the Latin term, a difference from the rst Critique and at odds with his normal practice in
discussing legal terminology. Pauline Kleingeld also rejects Proops reading on the basis that it
does not t with Kants own direct explanation of the fact and that it cannot explain the passages
in which Kant describes the fact in active terms (Kleingeld 2010, pp. 6465).
99
8
9
100
imperative. How this happens and what precise status the fact has requires
more explanation of the meaning and function of the fact of reason.
Fact and act
The precise way in which this imperative strikes us in our moral experience is
as a command. We take commands to issue from some authority, and this case
is no exception. The authority is assumed to be reason; hence, the fact of
reason is a command of reason of which we are conscious. We also understand
commands as acts, and since reason provides the command, we can ascribe this
act to reason. The fact of reason, then, can be understood as having a twofold
signicance as our actual conscious awareness of an act of command by
reason. The term that Kant uses, Faktum, allows for this meaning.
Every translation of the Critique of Practical Reason into English simply
translates the term Faktum as fact without even mentioning the availability
of an alternative translation of the term as act or deed. The German term
Faktum at the time was a germanization of the Latin term factum, and was
allowed to have two meanings (both related to the Latin term): fact and
deed.10 Kant uses the Latin term factum in the Metaphysics of Morals and
throughout his Reections and lectures on ethics and natural right as equivalent to the German Tat, which is best translated as deed (6:224), although at
6:371, he equates Factum with Tatsache (fact). In the second Critique, had
Kant meant fact in the sense of a piece of knowledge, he likely would have
used the German Tatsache instead of Faktum.
Marcus Willaschek has argued convincingly that the meaning deed or
act makes more sense than fact in many of the passages in which Kant
uses the term in the Critique of Practical Reason.11 These passages tend to
discuss reason as announcing itself (5:31), giving (5:31), forcing itself
(5:32), and determining the will in relation to dynamic laws that can
determine its causality (5:42). At least in these passages, it is better to
understand the Factum of reason as an act of reason in providing the moral
law. Particularly strong evidence comes from a quotation from the Critique of
Practical Reason in which Kant uses the terms Tat in a way equivalent to the
use of Faktum: For, if as pure reason it is really practical, it proves its reality
and that of its concepts through the act [die Tat], and all subtle reasoning
10
11
Neither the word Faktum nor Factum appears in the Grimm Brothers dictionary. Duden
indicates that the term Factum (or Faktum) was rst used in Germany in the seventeenth
century, and that it had meanings corresponding both to fact and to deed. Kleingeld cites
Zedlers Universallexikon from the mid-century in noting that deed was the preferred
meaning (Kleingeld 2010, pp. 6263).
See (Willaschek 1992, pp. 174193). The discussion of the Factum of reason appears
separately as well in (Willaschek 1991).
101
against the possibility of its being practical is futile (5:3, modied). These
considerations at least ought to put the translation of Factum as fact in
doubt, and provide compelling evidence that Kant intended the term to refer in
some way to the activity of reason.
To see the entirety of Kants use of Faktum as deed, however, would leave
out the important element of our awareness in the discussion of the Faktum.
Pauline Kleingeld suggests interpreting the Faktum as consciousness of the
fundamental law (of pure practical reason), a consciousness that reason produces in agents rather than a fact as something alien confronting the moral
agent; the Faktum is a fact generated by a deed. She gives an analogy
comparing the fact of reason to a decision by a king or a painting by
Rembrandt: the act results in a thing that individuals can become aware of as
a product of that act.12 I agree that consciousness is necessary and believe that
Kleingeld has provided the best interpretation of the Faktum of reason yet, but
I think that she still separates the conscious awareness from the act of reason
too much. My distinction between the moral law and the categorical imperative
allows for a deeper connection between the senses of fact and deed.
A moral agents awareness of the categorical imperative is not the result of
an act of reason, it is an act of reason in commanding us to act a certain way.
There is no lag between the generation of the command to act so that ones
maxims can serve as practical laws and the awareness of the categorical
imperative. They are the same thing within the agent-perspective. In moral
experience, when the agent is faced with a free decision with moral implications, reason acts in pressing the agent with the categorical imperative as a
command.13 Unlike the decision of a king or painting by Rembrandt that can
be known well after the actual deciding or act of painting, the awareness is
simultaneous with the deed. A better analogy, and one that illustrates what
I take the status of the fact of reason to be as a kind of deed, would be to
watching a dance by a ballet dancer because the result that one is aware of is
inseparable from the current action of the dancer. One can still remember the
dance later, but the phenomenal experience of the memory is different from the
experience of the performance. Analogously, the experience of the obligatory
force of the categorical imperative during moral deliberation is different from
the later memory of that experience. The Faktum of reason is the direct
experience of the categorical imperative.
12
13
102
The best interpretation of the Faktum of reason, then, is that of the consciousness of the categorical imperative as a command being issued by reason
to an agent caught up in the process of deliberation. Seeing the Factum of
reason as awareness of this deed of reason instead of a static fact ts the
philosophical function of the Faktum. Kant introduces the Faktum after presenting and resolving the two problems of nding the property of the will
that makes it determinable by mere lawgiving form, the solution being freedom, and of nding the law to determine a free will, the solution being
lawgiving form. Kant concludes that freedom and unconditional practical
law reciprocally imply each other (5:29), a claim Henry Allison has deemed
the reciprocity thesis.14 Without an independent basis to ground either side
of the reciprocity, neither freedom nor law can be justied. Kant is faced with a
circle15 (although he does not use that term). When the Faktum is seen as the
awareness of the categorical imperative as a command in moral experience, it
is clear how Kant is able to nd a way to justify both moral law and freedom
despite the limits of their reciprocal derivation. The circle is broken by invoking the practical experience of reasons commanding us in the categorical
imperative.
This practical perspective is within the agents perspective. Given the
assessment earlier in this chapter of the nature of the practical/theoretical
divide and my claim that the priority of action emphasizes the practical point
14
15
103
of view from the agent-perspective, the fact of reason can be understood within
that perspective as an unquestioned contribution to moral deliberation. The
agent views herself as subject to the categorical imperative and her will as
having the capacity to be determined by the categorical imperative to action.
The moral agent, when faced with a situation that demands action, is simply
aware of the call of the categorical imperative. The agent at that time and qua
agent cannot doubt the validity of the imperative since it is experienced as an
undeniable command. The possibility for doubting the validity of the categorical imperative comes only when the moral agent is not faced with a situation
that demands action, and thus when the moral agent in not directly involved in
moral deliberation and is not caught up in the agent-perspective. At this time,
the agent engages in theoretical reection about moral deliberation itself, and
the experience of the categorical imperative is subject to philosophical examination. The priority of belief is applicable at this point since this reection
involves a higher priority of theoretical reason.
Our experience of the categorical imperative can cause us to ask in reection
about its ultimate source. One possible source is pure practical reason, that is,
the categorical imperative might be derived from the moral law as a particular
instantiation for nite sensuous and rational beings. But it is also possible
before completion of the philosophical reection that the experience of the
categorical imperative does not have its source in pure practical reason; it
might, for example, merely stem from empirically conditioned reason (5:15).
In the rst case, the explanation of the categorical imperative will invoke
transcendental considerations about rational agency as such. In the second
case, our awareness of the categorical imperative in our experience stands on
its own as a practical fact about empirical agency. Because the categorical
imperative might not stem from pure practical reason, Kant is impelled to
investigate its source. The status of the categorical imperative will depend
upon the status of its source, if any, discovered during the investigation
conducted in the Critique of Practical Reason and elsewhere. My next chapter
will look in some detail at the ways in which Kant attempts to explain the
nature of reason, both empirical and transcendental, in order to provide a
justication for the categorical imperative upon reection. It needs no justication in practice.
The fact of reason is ideal rather than real at the empirical level because it
has no ontological status outside of the agent-perspective. It is the agents
conscious awareness of the categorical imperative. At the transcendental level,
the fact of reason must be understood in relation to the broader understanding
of the nature of reason itself. If Kant is able to defend the claim that the
possibility of moral experience depends upon a certain kind of moral agent,
and that this agent must experience the moral law as a categorical imperative,
then the fact of reason has transcendental justication. But since the fact of
104
Karl Ameriks provides this analysis in (Ameriks 1982, pp. 203220). He couches this interpretation in terms of a relation to Groundwork III in which Kants attempt to provide a
deduction of the moral law by rst proving that we are free falls short. Kant is then said, in
the second Critique, to reverse the order of proof and begin by asserting the moral law as an
undeniable premise, from which Kant is then able to assert that we really are free as a
metaphysical claim, as though we have practical insight into what theory cannot prove.
105
of freedom; the concept is also linked to the concepts of God and immortality
so that those concepts, not the objects corresponding to them, obtain objective
reality (5:34). The moral agent who cognizes freedom within himself does
so at the moment of decision when faced with the demand by his prince to
provide false testimony; the agent is caught in the agent-perspective from
which the agent must believe his decision is freely made (5:30). And it is
not freedom but the concept of freedom (my emphasis) that is said to have
objective and, though only practical, undoubted reality (5:49). This latter
sentence contains several caveats: not freedom itself but only its concept is
what is at issue; its reality is undoubted (unbezweifelte), implying that the
belief in the concept of freedom is inescapable rather than that freedom is itself
real; and most important, that the concept has only practical reality.
This quick look at three passages is by itself insufcient to show that
freedom is not an objectively real property of human beings. There is, however, also evidence that Kant intends freedom to be seen merely as an
assumption taken up for practical purposes. Of all the intelligibles, Kant says,
only freedom has reality for us only insofar as it is a presupposition inseparable from that law, which he glosses to on behalf of that law and of the use of
pure practical reason, implying that this reality of freedom is only from the
agent-perspective when confronted with the moral law when using practical
reason (5:70). Freedom is identied as one of the postulates of practical reason
(5:132, see also 5:45) and is subordinate to the moral law that is itself also
declared a practical postulate (5:46), clearly in a less technical sense than
Kant uses the term in reference to immortality, freedom, and God, but nonetheless something to be taken up only in a practical point of view.
When one takes into account the larger context of Kants practical philosophy, the freedom that is revealed by the fact of reason should be understood in
relation to the two priorities of practical reason. Under the priority of action,
the fact of reason is experienced by a moral agent from the agent-perspective
when faced with a decision. If freedom is understood in this way, it too would
be valid only from the agents perspective and not be an ontological claim.
Under the priority of belief, freedom would be tied to the fact of reason when
the agent is later reecting upon her experience. In this case, the immediacy of
the disclosure of freedom through the experience of the fact of reason is lost.
Reection upon the awareness of the categorical imperative would raise the
question of how such an imperative is possible, and freedom would be the
answer derived from analysis performed from speculative grounds. This kind
of analysis is provided by Kant in the second Critique, to be sure, but freedom
in this sense must be understood as one of the presuppositions for a
practical claim.
The fact of reason discloses freedom to us as a practical postulate. Kant
often uses the term freedom by itself, although he also identies freedom of
106
the will (Wille) directly (e.g., 5:2829). The Wille, however, is identied as a
practical reason in the Groundwork (4:412), and practical reasons determining
the will through the moral law is what constitutes its freedom. In order to
understand freedom of the will, it is necessary to determine the status of pure
practical reason such that it is able to determine a free will.
Concluding remarks
While the previous chapter viewed practical philosophy from the perspective
of philosophy in general, this chapter has examined practical philosophy
directly. I showed that Kants famous claim that the practical has primacy
over the theoretical can be understood in two different ways. The priority of
belief understands the practical as raising certain possible beliefs, the postulates in particular, that stem from reection on our moral lives and presenting
them to theoretical reason, which in turn is tasked with integrating those beliefs
into its own claims about the world. This task reects a higher priority for
theoretical reason because part of its goal is to understand human experience;
reection on practical experience presents theoretical reason with content it
then tries to incorporate into overall comprehension of experience. I argued
that even the Critique of Practical Reason can be seen as a work of theoretical
reason examining the nature and use of practical reason. The priority of action,
in contrast, gives the practical independence from any possible theoretical
claim that might cast doubt on the practical. When faced with deliberation
about a choice for action, agents decide as if they are free and using the
categorical imperative of which they are conscious without requiring any
theoretical conrmation indeed, even if there were theoretical disconrmation, Kant sometimes allows. This sense of priority is, I argue, operative in the
discussion of the fact of reason. I take the fact of reason to be the awareness of
the categorical imperative as a command issued by practical reason during
deliberation. Agents do not question the categorical imperative when deliberating but may later, upon reection, ask themselves for explanation and
justication. At that point, the experience of the practical becomes the subject
matter for theoretical investigation. The need for freedom, for the moral law
as a principle (and the conception of pure rational agency itself), for the
highest good, and for the postulates arise through that theoretical reection
but are still practical claims because they are based on practical experience.
The priority of action also gives a general indication of how Kants moral
theory is idealist both at the transcendental and empirical levels, since practical
experience is understood from the perspective of the moral agent. The following chapters will provide more details about specic practical claims, starting
with Chapter 4 about the status of the moral law and practical reason itself.
The previous two chapters discussed the place of the practical in Kants
philosophy as a whole and the specic nature of the practical. Practical philosophy is about free acts understood from the agent-perspective (Chapter 2).
The only method for systematizing those free acts is through an a priori moral
law. Moral duties must not be derived empirically. Yet the nature of Kants
transcendental method allowed that even pure duties could be found in empirical moral experience; their justication would be priori. A closer look at the
nature of practical reason (Chapter 3) showed that there are two main ways
of understanding it: in relation to action and in relation to belief. In the rst,
practical concerns do not need any conrmation by theoretical reason; in the
second, the task of understanding the place of the practical in experience falls
ultimately to theoretical reason. I identied the fact of reason with human
empirical consciousness of the categorical imperative in moral experience
complete with its normative commanding nature and argued that it is compatible with moral idealism at both the empirical and transcendental levels.
This picture is overly simplistic, however, since the transcendental method
of justication itself might show that morality requires some non-natural
metaphysics, or even transcendental realism. By that, I mean that some moral
principle, property, or object is independent of the transcendental moral agent.
It is possible that a conception of reason itself or God would have to exist
independent of moral agents themselves. The identication of the practical
with the agent-perspective does not by itself preclude this non-natural metaphysics. This chapter will examine this issue more closely with a focus on
practical reason.
I believe that the Wille as reason, not will (as Willkr or faculty of choice), is
the core faculty in Kants ethics, and that understanding its relation to the
transcendental and empirical levels is crucial for determining whether Kants
ethics should be considered ideal or real. Philosophy is a product of the activity
of reason in determining and pursuing its fundamental ends. The categorical
imperative is grounded in reason and duties stem from reason. The value of
humanity itself is based upon reason, either as a product of the categorical
107
108
Kant had many different kinds of freedom, and for many different faculties. Lewis White Beck
identied ve concepts of freedom with their own internal variations, and he did not even
include political freedom. The ve identied by Beck are (1) an empirical, psychological,
comparative conception that provides a basic distinction between voluntary and coerced action,
(2) moral freedom, or the ability to act on the moral law, which Kant says requires transcendental
freedom, (3) freedom as spontaneity, used for imputation and implying that the decision had no
external cause, (4) transcendental freedom, or a causal force independent of causality in nature,
and (5) postulated freedom, a two-perspective view that Beck nds the most plausible Kantian
conception of freedom (Beck 1987). Beck also emphasizes the dual character of freedom in the
Wille/Willkr distinction. He correlates Wille with the lawgiving in the conception of autonomy
and Willkr with the initiation of a new causal series in the conception of spontaneity (Beck
1960, p. 177). I disagree with Becks insistence that we must never suppose that there are two
faculties related to each other in some external, coercive way. There is only one, but it has prima
facie two kinds of freedom (Beck 1960, p. 180). Even while a famous and widely cited work
on freedom in Kant Henry Allisons Kants Theory of Freedom uses the singular, its author
acknowledges the bewildering number of ways in which Kant characterizes freedom and the
variety of distinctions he draws between various kinds or senses of freedom before announcing
his focus on transcendental freedom (Allison 1990, p. 1).
109
Translations into English of Kants works now generally render them as will
and choice or faculty of choice respectively. Unfortunately, this device
leaves the unsuspecting reader to associate what in English is generally called
the problem of freedom of the will with Kants term Wille. Since the
problem of freedom of the will is generally understood in terms of a free
choice, it is preferable to associate that problem with Kants term Willkr;
Chapter 6 will look specically at the problem of freedom of the will in
relation to the Willkr. The faculty of Wille is identied by Kant as nothing
other than practical reason (Groundwork, 4:412) and as the lawgiving capacity of rational beings (Metaphysics of Morals, 6:226). My interpretation
stresses this legislative role. Since, as I will show, Kant identies this legislative role specically with reason (Vernunft) in the Third Antinomy and the
Groundwork, I prefer to identify the legislative faculty as reason. In what
follows, I will stress reason but will continue to use the translation will when
Kant uses the term Wille and use reason only for Kants Vernunft.
Reason in nature in the Critique of Pure Reason
Despite the fact that Kants most famous work is entitled Critique of Pure
Reason, Kant does not provide any coherent denition of pure reason. Reason
is identied as the faculty of principles (A299/B356). He sometimes discusses
the pure as opposed to empirical use of reason, implying that there is one
reason applied to different subject matters (A563/B591). Other times, he contrasts empirical reason as appearance with reason in itself (A549/B577). I want
to contrast empirical and pure reason by showing that empirical reason is the
embodied faculty of reason in appearance and pure reason is the transcendentally free structure of reason considered independently of its empirical
manifestation.
The Critique of Pure Reason in general and the Third Antinomy in particular include claims that reason itself is a cause that must be conceived as part of
nature but also as transcendentally free. In this section, I will argue that the
Third Antinomys discussion of freedom centers on reason rather than on the
faculty of choice. I will show the evidence that Kant has a place for reason in
nature and offer an interpretation that best shows how reason can be both a
natural cause and have transcendental freedom.
In the solution to the Third Antinomy, Kant highlights the causal power of
reason as a faculty necessary for morality. Now that this reason has causality,
or that we can at least represent something of the sort in it, is clear from the
imperatives that we propose as rules to our powers of execution in everything
practical(A547/B575). There are in general two different possibilities for
understanding the causal relation of reason to these particular actions in nature.
First, reason could have some status that makes it independent of natural
110
Kant sometimes discusses animals actions as determined by instinct. See, for example, R7199
(19:272).
111
112
Many of the issues I raise in my discussion of Kants critical works are treated by Henry Allison
in (Allison 1990). Unlike many commentators on Kants ethics, Allison recognizes the importance of some conception of freedom for reason itself in addition to the problem of the freedom of
the will. I will note my specic disagreements with Allison particularly regarding the nature of
transcendental freedom in some of the following notes. In general, he equates Kants use of the
term reason in the Third Antinomy with a more expansive rational behavior, so that behavior
based on maxims is seen as the causality of reason (Allison 1990, p. 33). He also takes the
freedom of reason to consist in acts of judgment made on rational grounds rather than, as I do,
the principles of reason themselves. Practical spontaneity is then assigned to judgments, not the
faculty of reason as the source of a priori principles (Allison 1990, p. 36f).
Here, I largely concur with Keith Ward when he notes that the Third Antinomy is solved by
conceiving of reason as the faculty which, as changeless and eternal, is the transcendentally
free source of free actions. Ward uses the term noumenal freedom rather than transcendental
freedom. See (Ward 1972, pp. 757). We differ in explaining the status of that changeless and
eternal reason.
113
There are several places in the solution to the Third Antinomy that make it
clear that reason is the key faculty under discussion. When asking how human
beings are able to attribute to themselves more than just an empirical character,
he answers in regard to certain faculties, he is a merely intelligible object . . .
we call these faculties understanding and reason (A54647/B57475). Neither
will nor a faculty of choice is mentioned. Reason provides greater support for
this claim than the understanding because the understanding is linked to empirically conditioned powers while reason relates to objects only through its own
ideas. In the very next paragraph, Kant singled out imperatives as evidence that
reason must have causality.
Now that this reason has causality, or that we can at least represent something of the sort
in it, is clear from the imperatives that we propose as rules to our powers of execution in
everything practical. The ought expresses a species of necessity and a connection with
grounds which does not occur anywhere else in the whole of nature. (A547/B575)
This discussion of reason as a cause will be taken up in my next section, but for
now, the key point is that reason is specically cited for its role in practical
determination to action. Kant continues in his discussion by saying that
reason does not give in to those grounds which are empirically given
(A548/B576); it is at least possible for us to assume that reason actually does
have causality in regard to appearances (A54849/B57677) and there is
a certain causality of his reason (A549/B577); that reason is the cause of
producing actions and that the ideas of reason have actually proved their
causality in regard to the actions of human beings as appearances (A550/
B578); asks could reasons action then be called free and answers in part by
invoking the fact that pure reason, as a merely intelligible faculty, is not
subject to the form of time (A551/B579); and among many other occurrences
in the following pages, Reason is thus the persisting condition of all voluntary
[willkrlich] actions under which the human being appears . . . every action,
irrespective of the temporal relation in which it stands to other appearances, is
the immediate effect of the intelligible character of pure reason; reason therefore acts freely (A553/B581).5
Even the discussion of freedom in the solution to the Third Antinomy that
seems most focused on the decision of the agent actually reveals this causal
5
Although this last passage clearly identies reason as the freely acting intelligible, it is sometimes used in support of very different claims. Eric Watkins, for example, quotes this passage up
through the semi-colon but omits the last four words (reason therefore acts freely) and instead
concludes that Kant . . . refrains from attempting to explain which particular intelligible
character we might be able to impute to things in themselves on the basis of the empirical
characters that cause the particular actions we perceive in the sensible world (Watkins 2005,
pp. 32728). On the same page, he similarly overlooks Kants claim in R5611 that the activities
of reason are not appearances by taking the reference to more generally apply to things in
themselves.
114
power of reason. Near the end of his discussion, Kant offers the example of the
malicious liar to show the way that the freedom he has discussed has an
empirical use. The malicious lie, we are told, can be traced back causally to
bad upbringing, a wicked natural temper, and so on. These determining causes
seem to fully explain the lie. Yet, Kant says we blame the agent despite all of
this empirical explanation. This blame attributes the cause of the lie to a lack
of causation by reason: This blame is grounded on the law of reason, which
regards reason as a cause that, regardless of all the empirical conditions just
named, could have and ought to have determined the conduct of the person to
be other than it is (A555/B583). Further, Kant says, we transfer blame from
the agent to reason: now, in the moment when he lies, it is entirely his fault;
hence reason, regardless of all empirical conditions of the deed, is fully free,
and this deed is to be attributed entirely to its failure to act. This attribution of
blame to reason for its failure to act is at odds with blaming an agent for
making the wrong decision. And the active way that reason could have and
ought to have determined the conduct of the person rather than a claim that
the agent could have chosen to act out of reason also reveals that the key
faculty under discussion in the solution to the Third Antinomy is reason rather
than free choice.
Reason as a natural cause
I will now discuss in more detail the nature of empirical reason.6 In particular,
I will show the plausibility of a robust conception of empirical reason that
allows reason to be an entirely natural cause within appearances.
A non-naturalistic interpretation of Kants discussions of reason, as well as
the faculty of the understanding and other active faculties, is that they exist
only in things in themselves and not in appearances, but that reason and the
other active faculties have effects in appearance. But there is also a place in
Kants metaphysics to count reason as among the appearances in space and
time. Things in themselves can have counterparts in the world of sense as they
appear. Certainly, if reason is to be considered to be a thing-in-itself or faculty
in something-in-itself, then it can have a counterpart as appearance or as part of
6
Empirical reason is not directly mentioned by Kant very often. This term empirical reason can
be and certainly was at times used by Kant as shorthand for empirical use of reason, as opposed
to a pure use. The earliest use of the term I have found is in R4582, dating from 177275, in
which it is listed alongside pure reason and practical reason in theological matters: The need of
reason to cognize a highest being is that of a necessary hypothesis of the employment of reason:
(1) of pure reason; (2) of empirical reason (both are speculative); of practical reason (17:601).
A slightly earlier note from 1771 asks what are the appearances of reason? (Subjectively
determined general cognitions) (R4375; 17:525). In R4582, Kant appears to be discussing the
empirical use of reason, but in R4375, he is certainly asking about the existence of reason in
appearances.
115
some beings appearance. This is the sense of empirical reason at stake here.
Kant even describes the task of the Critique of Practical Reason as taking for
granted that there is empirical practical reason and asking whether a corresponding pure practical reason can exist and be related to empirical actions. One
of Kants notes made in the 1780s on a loose piece of paper rather than in his
textbooks, and thus presumably made in preparation for the Critique of
Practical Reason rather than in preparation for his lectures, stresses this point:
The critique of practical reason has as its basis the differentiation of empirically conditioned practical reason from the pure and yet practical reason and
asks whether there is such a thing as the latter (R7201; 19:27576).
In the Solution to the Third Antinomy, he holds that an empirical manifestation of reason is necessary in order for reason to be understood as a cause at
all. Kant identies the causal power of reason, at least in part, with the
application of a rule:
Thus every human being has an empirical character for his power of choice, which is
nothing other than a certain causality of his reason, insofar as in its effects in appearance
this reason exhibits a rule, in accordance with which one could derive the rational
grounds and the actions themselves according to their kind and degree, and estimate the
subjective principles of his power of choice. Because this empirical character itself must
be drawn from appearances as effect, and from the rule which experience provides, all
the actions of the human being in appearance are determined in accord with the order of
nature by his empirical character and the other cooperating causes. (A549/B577)
116
117
118
particular thoughts that would require the existence of this faculty for consistent and comprehensive explanation of these appearances in nature.
Two other possible objections deserve discussion. In the Solution to the
Third Antinomy, just after his discussion of empirical reason, Kant appears to
hold that when reason is held to be the cause of actions, it cannot in any way be
considered part of nature but must be considered as an alternative cause to
natural causes. He contrasts the causality of ideas of reason with empirical
causes and insists that reason must be independent of the conditions of time,
thus not in nature. There are two variations of this reasoning, one focused on
practical reason, the other on reason in general. It is important to look at these
arguments to determine whether they preclude the strong sense of empirical
reason I am advocating that would include a representation of the categorical
imperative in empirical consciousness as the natural cause of some actions.
First, Kant argues that reason as practical cannot itself have an empirical
character in nature as a natural cause.
But if we consider the very same actions in relation to reason, not, to be sure, in relation
to speculative reason, in order to explain them as regards their origin, but insofar as
reason is the cause of producing them by themselves in a word, if we compare them
with reason in a practical respect then we nd a rule and order that is entirely other
than the natural order. For perhaps everything that has happened in the course of nature,
and on empirical grounds inevitably had to happen, nevertheless ought not to have
happened. At times, however, we nd, or at least believe we have found, that the ideas
of reason have actually proved their causality in regard to the actions of human beings
as appearances, and that therefore these actions have occurred not through empirical
causes, no, but because they were determined by grounds of reason. (A550/B578)
Kant can be taken to present the following argument here: Reason as cause
provides ideas that themselves reect a normative rule and order regarding
what ought to happen. This normative rule and order is different from the rule
and order of nature, in other words, nature is not what it ought to be. Because
the normative rule and order is other than the natural order, any idea that
expresses a normative rule and order cannot be a natural cause in the natural
order. The argument is ambiguous between an incompatibility of the content
(the order) with the order of nature and an incompatibility of the normativity
(the ought) with the order of nature.
Either way, this argument does not show that reason cannot be in nature.
Suppose, on the one hand, the argument relies on a non-natural content in the
idea. The content of a representation in inner sense may be incompatible with
the natural order, but the idea itself may be part of that natural order. It is
entirely conceivable to have a representation in inner sense whose content
is incompatible with the natural order (for example, a dream in which I am
speaking with the living Immanuel Kant) yet which as representation is
entirely within the natural order. Suppose, on the other hand, that the argument
119
relies on a claim that the normativity of the idea is non-natural. But mere
normativity is not non-natural; the existence of desires shows that representations in nature can contain both preferences for certain outcomes and a moving
force beyond themselves. Thus, it appears that an idea of reason can exist in
nature even though it may represent normative preference for a non-natural
order. Empirical reason, then, can include a representation that provides a
moral order, namely, the categorical imperative. Note that this conclusion is
different from saying that empirical reason can be the only source of a
representation of the categorical imperative. To allow that the categorical
imperative can exist in empirical reason is not the same as allowing that it
can be justied by empirical reason. Justication of the categorical imperative
can still require pure practical reason.
A second objection that might be given to show that Kant has no room for
empirical reason is that he repeatedly attributes spontaneity to reason and other
faculties. Henry Allison makes this point central to his interpretation of Kants
theory of freedom7. Spontaneity of the faculties of understanding and reason
involves judgments (which Allison calls takings) and creation of ideas
(which Allison calls framings). These together imply that the faculties of
reason and understanding must be understood as involving deliberation and
justication, which in turn cannot be simply understood in natural causal
terms. This epistemic spontaneity is mirrored by a practical spontaneity, in
which decisions of the power of choice for particular actions cannot be
understood in natural causal terms but must be understood in terms of an
incorporation of ones incentive into ones maxim. The incorporation
requires rational deliberation and spontaneity of the agent. Hence, rationality
and agency can be understood only in terms of an intelligible character of the
agent as spontaneous. The requirement of spontaneity is so strong that, he says,
we cannot both deny such a character and afrm our status as rational agents,
that is, rational agency can be understood only in terms of non-natural spontaneity of reason and will. For present purposes, the important point is that
Allisons argument would preclude the possibility of an empirical reason, for
rationality could be understood only as spontaneous and different from the
causal order of nature.
Allisons position has generated much debate in the literature.8 Here, I will
focus only on one claim: namely, that Allisons interpretation reads too much
into the requirement for epistemic spontaneity. Kant does hold that reason
must be spontaneous, but this spontaneity need not involve elaborate transcendentally free deliberation independent of natural causal relations. Empirical
deliberation may proceed according to causal laws in nature precisely because
7
8
(Allison 1990, p. 35ff). The quotation later in my paragraph is from (Allison 1990, p. 41).
See some of Allisons replies in (Allison 1996, pp. 10942).
120
some of those causal laws are psychological rules of reason. If the resulting
conclusion is correct, it is in part because the deliberation utilized rules of
reason that are themselves justied, even though the deliberation might have
been entirely determined as a series of events in nature in the mind. What is
required for the justication of those rules of reason themselves is a separate
issue. Here, transcendental freedom for reason would be required. But this
justication need not involve any extended deliberation. It may result from the
timeless nature of reason itself. I will say more about this in the next section.
Reason as a transcendentally free cause
Reason as transcendentally free in the Third Antinomy
In understanding the solution to the Third Antinomy, one must keep in mind
Kants claim that the Thesis has proved not that a single cause (Ursache) is
necessary but that a non-natural type of causality (Kausalitt) must exist
(A444/B472). Indeed, a review of the argument for the Thesis shows that
Kant does proceed this way. At least one instance of an Ursache of the proper
Kausalitt must exist to avoid the contradiction that would result were all
causality merely natural. Kant then announces that We are now permitted also
to allow that in the course of the world different series may begin on their own
as far as their causality is concerned (A450/B478). Whoever gave Kant that
permission was in error, since the argument (if successful), strictly speaking,
has shown only that one cause of that type would be necessary and one
normally cannot make an argument from a single instance to other similar
particulars. But for the sake of argument, let us allow that if there must be at
least one cause of a type of causality, then further causes of that type could be
invoked if other reasons call for them.9
Freedom as a type of causality should be understood as the causality of
reason in determining the will as power of choice to action, conceived
alongside the natural causality of empirical reason but as a different type
of causality that also exists in nature. I will show that there is no metaphysical conict with holding that reason is a cause in both types of
causality, each understood for different purposes. The requirement for
reason to be a transcendentally that is, justiable under a transcendental
argument free cause does not entail that reason is a transcendent that is,
existing outside nature cause. A transcendentally free reason can exist in
Watkins also stresses that Kant is working with a different type of causality but still gives this an
ontological reading rather than seeing both types of causality as attributed to beings in nature
(Watkins 2005).
121
nature just as other transcendental conditions for the possibility of experience such as causal relations can exist in nature.
That the transcendentally free causality Kant discusses in the Third Antinomy is supposed to be an alternative kind of causality among objects in
nature is, in addition to tting the philosophical function of practical freedom,
supported by some direct evidence in the text. An overlooked passage in the
solution to the Fourth Antinomy describing the solution to the Third Antinomy
is very clear. After clarifying that the necessary being required by the Fourth
Antinomy would not be a member of the series of empirically conditioned
beings in the world of sense, Kant contrasts that metaphysical claim with a
claim that the transcendentally free cause in the Third Antinomy is part of the
series of conditions in nature:
Hence this way of grounding an unconditioned existence [in the Fourth Antinomy]
would be distinguished from the empirically unconditioned causality (of freedom) in
the previous article in that in the case of freedom, the thing itself as cause (substantia
phaenomenon) would nevertheless belong to the series of conditions, and only its
causality would be thought as intelligible, but here the necessary being would have to
be thought of as entirely outside the series of the world of sense (as an ens extramundanum) and merely intelligible; this is the only way of preventing it from being
subjected to the law of the contingency and dependency of all appearances. (A561/
B589, emphasis in original)
Only the causality of freedom and not the thing that possesses the causality of
freedom would be intelligible; the object possessing transcendentally free
causal power is an object in nature.10 Kant makes a similar point in a passage
in the solution to the Third Antinomy when he denes intelligible and relates
it to a different kind of causality: I call intelligible that in an object of sense
which is not itself appearance and states that one can consider the causality
of this being in two aspects, as intelligible in its action as a thing in itself, and
as sensible in the effects of that action as an appearance in the world of sense.
Of the faculty of such a subject we would accordingly form an empirical and at
the same time an intellectual concept of its causality (A538/B566, partly my
emphasis). The causality, not the being, is understood as intelligible yet as
attributable to the object of sense. Admittedly, Kant uses the term thing in
itself here, and there is some justication for the metaphysical claim that Kant
intends to discuss a non-natural being. But the evidence I presented earlier that
ties freedom to reason suggests that it is worth looking at an interpretation that
identies the transcendentally free causality as some kind of causality of
reason.
10
Henry Allison notes this passage but interprets it to mean that the intelligible character is
assigned to a being with a phenomenal side instead of recognizing that Kant intends the entire
being to be phenomenal and only the causal relation to be intelligible (Allison 2012, p. 29).
122
In the Canon, Kant revealed this need for the possible transcendental freedom
of the faculty of reason when he admitted that practical freedom can exist within
sensible nature while transcendental freedom deals with the non-sensible status
of the faculty of reason as well as with the origin of actions:
We thus cognize practical freedom through experience, as one of the natural causes,
namely a causality of reason in determination of the will, whereas transcendental
freedom requires an independence of this reason itself (with regard to its causality
for initiating a series of appearances) from all determining causes in the world of the
senses, and to this extent seems to be contrary to the law of nature, thus to all possible
experience, and so remains a problem. (A803/B831, emphasis mine)
The question of transcendental freedom arises when the status of the faculty of
reason itself is questioned about its ability to begin a series of appearances
through its causality. And with regard to its causality not the entire faculty
of reason itself is to be independent of determining sensible causes.
Transcendentally, free causality is tied to the intelligible character of reason,
which ought not to be equated with a transcendent reason. Kant introduces the
idea of an intelligible character as opposed to the sensible character. The
character of a cause is dened as a law of its causality without which it
would not be a cause at all (A539/B567). An agents actions in appearance
are said to have an intelligible character, through which it is indeed the cause
of those actions as appearances, but which does not stand under any conditions
of sensibility and is not itself appearance. This causal relation does not stand
under conditions of time, so no action would arise or perish in it as intelligible.
As intelligible, then, it would not change, yet it would have causal effects in
the world of sense in time. Thus freedom and nature, each in its full signicance, would both be found in the same actions, simultaneously and without
any contradiction, according to whether one compares them with their intelligible or their sensible cause (A541/B569). Kant looks for a faculty that is
only intelligible, in that its determination to action never rests on empirical
conditions but on mere grounds of the understanding (A545/B573).
As I have noted, human beings count themselves as intelligences because they
experience the effects of the understanding and, above all, reason (A54647/
B57475). The understanding with its a priori concepts and reason with its ideas
appear to exhibit spontaneity because these concepts and ideas cannot be
understood as arising from sensible causation. The ideas and concepts cannot be
represented as sensible themselves. Hence, individuals must view themselves as
intelligences in order to believe that their reason is valid.11 To view oneself as an
intelligence is simply to take oneself to possess valid cognitive faculties.
11
In contrast, Ameriks takes this passage to be evidence that Kant retained pre-critical
rationalist views about freedom into the early 1780s. Freedom is taken to be a property of the
123
The causal power of reason on the sensible will is, importantly, a timeless
cause since reason itself is timeless. Many passages in the solution to the Third
Antinomy point this out: Pure reason, as a merely intelligible faculty, is not
subject to the form of time . . . The causality of reason in the intelligible
character does not arise or start working at a certain time in producing an
effect (A551/B579). For since reason itself is not an appearance and is not
subject at all to any conditions of sensibility, no temporal sequence takes place
in it even as to its causality (A553/B581). Reason does not alter (even if its
appearances, namely the way in which it exhibits its effects, do alter)
(A55556/B58384). And reason is present to all the actions of human beings
in all conditions of time, and is one and the same, but it is not itself in time, and
never enters into any new state in which it previously was not (A556/B584).
The causal condition that lies in pure reason is a timeless, persistent condition that is able to inuence the sensible will to action. Kant does not describe
it as different for different individuals, as individual free choices certainly
would be. This general, time-independent, determinant of human sensible
choices must be the moral law itself understood as the pure structure of reason
itself applied to the circumstances of possible human free choice. Pure reason
can be a cause of actions in the sensible world by dictating that they ought to
occur. Reason as a faculty can be understood as providing a structure for
experience, specically, a structure for systematic and consistent actions.
Pure reason as timeless structure
The previous sub-section showed that in the Critique, Kant is concerned with
the transcendental freedom of reason with regard to its causality. This subsection will show the kind of transcendental structure that pure practical reason
exhibits, in particular in relation to the moral law. The following section will
discuss whether Kant is justied in attributing transcendental freedom to pure
practical reason by means of that structure.
It is rst useful to review Kants few descriptions of reason in general in the
Critique of Pure Reason. Reason combines its logical use, where reason
abstracts from content, and a real use as a faculty in which it contains the
origin of certain concepts and principles, which it derives neither from the
senses nor from the understanding and itself generates concepts (A299/
B355). The real (or transcendental) and logical uses of reason are combined
in his denition of reason as the faculty of principles (A299/B356). He
implies that this denition entails that reasons principles rest on thought alone
soul to think and act independently of determination in nature, and the faculties of reason and
understanding are evidence that human beings have immaterial souls that can think and act
freely (Ameriks 1982, p. 190).
124
125
beings apart from any of the particular conditions that distinguish these
particular rational beings, is the structure of that faculty.
This structure of pure practical reason that generates the moral law is the
autonomy of reason in morality. In the second Critique, Kant claims that the
moral law expresses the autonomy of pure practical reason because it is a
lawgiving of its own on the part of pure and, as such, practical reason (5:33).
Autonomy is the independence of practical reason, which is equivalent to the
will (4:412) from determination by anything outside itself. The moral law is
determined by the structure of reason, so in one sense, reason is not free
because it simply subsists with the structure and its corresponding moral
law; I believe that this is the sense that Kant has in mind when in the
Metaphysics of Morals, he says that Wille is neither free nor unfree. Reason
is considered independent of determination by laws of nature because, in this
sense, there is no determining ground for reason at all (6:213). It is fair to say
that for Kant, autonomy of pure practical reason in generating the moral law is
the transcendental condition for their being any morality for freely acting,
rational beings. This transcendental structure is essential for Kants ethics. It is
familiar to everyone who has studied Kants ethics in its guise as the autonomy
of the will as opposed to the heteronomy of all other foundations for ethics,
and the arguments that the moral law is based on form alone as opposed to any
content. It is expressed through Kants insistence that the moral law be a priori
and thus necessary rather than a posteriori and thus contingent. (I will say more
about autonomy in relation to moral value in Chapter 7.)
This reason must be transcendentally free in its causality, that is, an
independence of this reason itself (with regard to its causality for initiating a
series of appearances) from all determining causes of the world of the senses,
to which Kant adds that this independence seems to be contrary to the law of
nature (A803/B831). The transcendental freedom of reason itself is here
understood to be the ability of the faculty of reason to begin a new causal
series in nature independently of the prior causal series in nature. There are
many places where Kant seems to take transcendental freedom of the faculty of
reason to require an ontologically distinct realm of things in themselves
independent of appearances in nature in space and time. But a closer consideration of this issue shows that Kant does not in fact require transcendental
freedom to be understood as ontological independence of the faculty of pure
reason from nature, and that transcendental freedom can be met within a
metaphysically naturalistic ontology.
Independence from causation in nature does not mean independence from
existence in nature. There are two ways in which something (A) can be
independent from determination by something else (B). First, A may exist
apart from B in such a way that B never has any contact with A. I will call
this transcendent independence or freedom. An example of this sort of
126
127
128
Corinna Mieth and Jacob Rosenthal raise this issue but do not go far enough in recognizing its
import. Since, they say, an electronic calculator operates according to both the laws of nature
and the laws of mathematics, there is a clear way in which one and the same result (in this case,
a mathematical conclusion) is the result of two kinds of laws without contradiction (Mieth and
Rosenthal 2006, pp. 27273).They take Kant to reject this kind of double understanding of
processes that are ontologically in nature. They quote the following passage from the Groundwork in which Kant raises a possible contradiction between freedom and nature: It would,
however, be impossible to escape this contradiction if the subject who seems to himself free
thought of himself in the same sense or in the very same relation when he calls himself free as
when he takes himself to be subject to the law of nature with regard to the same action (4:456).
Yet precisely in that passage, Kant allows that different relations could be at work in a being
understanding itself as free and as naturally caused in the same action, and the different causal
relations of an empirically determined causality and a transcendental free causality. This is
exactly the kind of view I am attributing to Kant. Mieth and Rosenthal recognize that the view is
plausible but do not attribute it to Kant.
129
the Third Antinomy (A50/B478). The solution to the Third Antinomy requires
that both thesis and antithesis be true, so some events are determined in the
series of natural causes and effects in time and are also understood as the
initial spontaneous beginning of a new series. In the resolution of the Third
Antinomy, Kant identies those events as decisions of the empirical power of
choice determined by reason. Because reason is timeless as a certain kind
of structural relation among concepts, it has no preceding state that would
have been different and from which it, qua reason itself, would have arisen.
(This cannot be said of individual manifestations of reason in particular
individuals which, as an empirical faculty of an organism, would have come
into existence and developed with the emergence and growth of that organism.) When a particular moral action is attributed to reason as a cause, this
timeless structure of reason that has generated the moral law as a categorical
imperative is viewed as the spontaneous cause, even as simultaneously, the
categorical imperative itself is understood in terms of the fact of reason as
existing in the empirical agents conscious mind. A different type of causality, rational causality, is at work in addition to the deterministic causality of
natural psychological causal forces. That empirical reason exists in the
empirical mind as described by empirical psychology does not prevent that
faculty of reason from being understood in terms of its structure. As Kant had
claried in his discussion of the Fourth Antinomy, in the case of freedom,
the thing itself as cause (substantia phaenomenon) would nevertheless belong
to the series of conditions, and only its causality would be thought as
intelligible (A561/B589). The transcendentally free causality of pure practical reason in providing the moral law is a timeless cause that can be
understood to begin a series of effects in nature from a practical point
of view.
Moral agency is delineated by the relation an agent has to the structure of
reason. To be a moral agent is to be capable of having the decisions of ones
empirical power of choice in nature determined by pure practical reason
through its empirical manifestation as empirical reason. The particular structure that reason imposes on decisions is the systematic determination of the
conditions for the free actions of all moral agents. The moral law as a description of the activities of purely rational agents provides the idea of such a
systematic union; the categorical imperative, as the fact of reason, is the
manner in which the human empirical power of choice experiences the obligation to conform ones actual decisions as much as possible with that moral law
by opting to resolve deliberation in favor of the principles of reason. This last
decision is understood from the agents perspective; from the theoretical
account given in empirical psychology, it would be seen as the motive of
reason rather than the impulses of sensibility determining ones power of
choice.
130
This section has shown that pure practical reason as a faculty possesses a
timeless structure that generates the moral law. Reason as a faculty in general
is a set of formal relationships and laws to provide systematic connection
among concepts. Pure practical reason provides the formal systematic basis for
decisions, namely, that they be based on law-like principles that systematize
free actions. There is thus a transcendental basis for the moral law and a reason
to attribute transcendental freedom to pure practical reason. The transcendental
freedom of reason had been dened in the Canon as an independence of this
reason itself (with regard to its causality for initiating a series of appearances)
from all determining causes of the world of the senses (A803/B831). The
structural account of the freedom of reason that I have provided here meets all
the criteria that Kant offers for transcendentally free causation in moral actions.
The transcendental freedom of reason cannot be proved
While Kants conception of morality includes this transcendental structure, he
also denies that it can be proved to be real. This denial sets up a complicated
situation for determining the extent to which Kant is a moral realist. A later
section will assess realism and idealism about reason in detail. This section
will show why Kant insists that the picture of morality involving a transcendentally free faculty of reason providing an a priori moral law cannot be
proved.
The end of the solution to the Third Antinomy contains Kants own limitation on the conclusions to be reached. Freedom, he says, has not only not been
shown to be real; it has also not even been shown to be possible. The reasons
he gives anticipate the reasons he provides in both Groundwork III and the
second Critique to limit the status of claims about freedom. First, from
experience we can never infer something that does not have to be thought in
accord with the laws of experience and second, from mere concepts a priori
we cannot cognize anything about the possibility of any real ground or any
causality (A558/B586). The rst reason stresses that the kind of transcendental freedom shown in the Third Antinomy is part of the minimal explanation of
experience in accordance with the laws of nature. This reasoning does not
preclude transcendentally free causality from existing in nature, but as a
distinct kind of causality, it cannot be proved from any empirical data. The
second reason limits the status of transcendental freedom to a purely conceptual level, a limitation that likens it to a transcendental idea, the status of which
will be part of my next chapter on the postulates.
In the second Critique, Kant denies that the moral law is subject to a
transcendental deduction. Unlike the categories of the understanding which
are susceptible to the transcendental proof of their necessity for experience, the
moral law has no proven transcendental role in experience. He makes this
131
132
deduction of that faculty on the other hand, mirrors the distinction in the
Critique of Pure Reason between the metaphysical and transcendental deductions of the categories. The metaphysical deduction provides and defends the
list of pure concepts of the understanding but does not prove that they are
necessary for experience; the transcendental deduction proves that these pure
concepts as a set are necessary for experience but does not explain their details
(B159). The categories themselves are in essence the structure of the pure
understanding, and that structure is known prior to its transcendental deduction. What the structure of the pure understanding is, is not in question in the
transcendental deduction. In the same way, the structure of reason necessary to
legislate the moral law is not questioned in the passage I quoted.
Reason in Groundwork III
The interpretation of the importance of reason and the nature of its transcendental freedom in the rst Critique holds for Groundwork III as well. In
Groundwork III, Kant is not concerned with individual choice but with the
faculty of reason. He asks whether we as empirical moral agents are also
justied in attributing to ourselves a transcendental conception of moral
agency centering on a valid faculty of reason. In this section, I will defend
this reading of Groundwork III.13
The argument Kant provides in Groundwork III is primarily a justication of
reasons ability to provide valid law and not an argument for free human moral
action in any sense other than its determination by that valid law. Once Kant
can justify this legislative power of reason, he can defend the human obligation
to will maxims that conform to the categorical imperative, and will have
secondarily defended freedom of the will as the empirical power of choice
able to follow reason. He will have defended the obligation human beings have
to try to make their actions conform to the rational dictates of morality, and to
that extent, will have shown that it is the responsibility of human beings to
attempt to make their actual world conform as closely as possible to a perfectly
rational world.
13
The interpretation of Groundwork III presented here builds on the work of several other
interpreters while not claiming the full endorsement or agreement of any of them. I am indebted
in particular to the work of Marcus Willaschek, which has been discussed in my previous
chapter in relation to the fact of reason (Willaschek 1992), Julio Estevess argument that Kant
does not invoke theoretical arguments in Groundwork III and that his claims are fully compatible with the Critique of Practical Reason (Esteves 2012), Corinna Mieth and Jacob Rosenthals
interpretation of the nature of the freedom of reason itself in relation to justication (Mieth and
Rosenthal 2006), and Jens Timmermanns assessment of freedom of the will that shows it as
merely a capacity to be determined by principles of reason rather than a free choice among open
possibilities (Timmermann 2003).
133
Others see this as the claim that Kant is discussing freedom of judgment here,
that is, a transcendentally free particular act of the power of judgment independent of any causal determination according to the laws of nature.14 It is
thought that if my judgment is determined in accordance with the laws of
nature, then it is determined only by an impulse or some previous efcient
cause in accordance with the laws of nature. Were my judgments or choices
determined by an impulse, I would have no reason to ascribe validity to them.
I must assume, then, that the particular judgment I make is not determined by
impulses. This approach concludes that every rational being must assume she
is free in making particular judgments or moral decisions.
But in focusing on particular judgments as acts, this approach makes too
broad a jump.15 The passage does not require that Kants focus is on particular
14
15
Henrich stresses the importance of the freedom of judgment in reasoning (Henrich 1975, p. 65).
Allison discusses epistemic spontaneity as part of Kants argument for the freedom of
decision for the will in relation to a judgment (practical spontaneity) and more recently in
terms of a distinction between the space of reasons and the space of causes (Allison 1990,
pp. 22223) and (Allison 2011, pp. 30708).
Mieth and Rosenthal provide an excellent critique of the standard approach to this question.
They note that there are ve different ways that one can discuss freedom, the rst of which is
freedom as rationality. Freedom as rationality entails that ones deliberations and judgments are
in accord with reason. Mieth and Rosenthal note that that this allows these deliberations and
134
16
135
way of stating the question that Kant posed at the end of Groundwork II,
namely, whether the moral law is itself possible.
At this point in Groundwork III, Kant could simply present a further
argument for the possibility of a free reason in order to resolve the issue. But
before attempting to do that, he stresses the importance of the upcoming
argument. The rst four paragraphs of the section entitled Of the Interest
Attaching to the Ideas of Morality (4:448450) focus on the question
of obligation. Having shown that rational beings unavoidably attribute an
as-yet-unproved freedom to themselves (and hence attribute the moral law to
themselves), he recognizes that mere self-attribution is not enough to provide
obligation. But why, then, ought I to subject myself to this principle, and do
so simply as a rational being? he asks (4:449). Perfectly rational beings act
rationally without any interference from impulses. But human beings are not
perfectly rational beings, and instead nd our actions determined by sensibility. Why should human beings not simply allow their actions to be determined
by sensibility? Why take any interest in the moral law? If the moral law is
merely something human beings assume for themselves, there appears to be no
reason to take it as binding. Put another way, if human beings merely assume
that they are free, that is, possess a reason that can produce independently valid
principles, then what guarantee is there that these alleged principles of reason
really are valid? Kant concludes It seems, then, that in the idea of freedom we
have actually only presupposed the moral law, namely the principle of autonomy of the will itself, and could not prove by itself its reality and objective
necessity (4:449). On what grounds, he asks, is the moral law binding?
This consideration leads Kant to admit to a kind of circle in his argument.
The way that he describes this circle is telling: he notes that he has shifted from
talk of efcient causes to talk of the order of ends:
It must be freely admitted that a kind of circle comes to light here from which, as it
seems, there is no way to escape. We take ourselves as free in the order of efcient
causes in order to think ourselves under moral laws in the order of ends, and we
afterwards think ourselves as subject to these laws because we have ascribed to
ourselves freedom of the will: for, freedom and the wills own lawgiving are both
autonomy and hence reciprocal concepts, and for this very reason one cannot be used to
explain the other or to furnish a ground for it. (4:450)
When he says take ourselves as free in the order of efcient causes, Kant is
referring to the claim we make that our reason is independent of impulses.
When he then says that we think ourselves under moral laws in the order of
ends, he implies that from the lack of efcient impulsive cause, we assume
that we are free in the use of our reason to legislate the moral law for ourselves.
But why is this categorical imperative really valid? Merely our assumption that
we are governed by reason is insufcient. Hence, the problem that we have
136
only so far ascribed freedom to ourselves. Our mere assumption that we are
subject to the moral law is not enough to show that we are actually subject to
the moral law. There would have to be some additional basis for the claim that
we are rational beings and can then assume that reason freely provides its own
principles.
Reason and the intelligible
The circle can be escaped only if there is an independent way to afrm that we
can legitimately take ourselves to be rational. This will afrm the validity of
the moral law, and thus afrm the reality of autonomy or freedom of the will,
its ability to act on the basis of the moral law. It will conrm that the order of
efcient causes in nature is not the source of the categorical imperative, but
that a different source of the categorical imperative is available. This is the path
Kant takes to escape the circle by invoking the intelligible world as the view
that human beings take themselves as rational and therefore as subject to the
moral law provided by reason. The circle problem is declared resolved a few
pages later when Kant says:
for we now see that when we think of ourselves as free we transfer ourselves into the
world of understanding as members of it and cognize autonomy of the will along with
its consequence, morality; but if we think of ourselves as put under obligation we regard
ourselves as belonging to the world of sense and yet at the same time to the world of
understanding. (4:453)
Henrich, for example, emphasizes that Kants argument for the consciousness of our freedom
depends on his crucial two-world doctrine so that the consciousness of our freedom depends
upon the claim that our will belongs to a world other than the sensible world (Henrich 1975,
p. 98).
137
138
139
19
Strictly speaking, the causal relation is between reason as a cause and the feeling of respect as
an effect. The feeling of respect can then be understood as a purely natural cause of a decision of
the will (4:460). The relation between a timeless idea and a sensible feeling operating at a
particular time in the human mind remains.
Patrick Kain is a clear example. In (Kain 2004), he lays out Kants position regarding the
authorship and legislation of the moral law. Human beings can be seen to legislate the moral
law, as can God, because to legislate means only to place someone under an obligation to obey
the law. But the law is not itself created by either individual moral agents or by God; rather, the
law itself has no author. It is given by the nature of reason itself (Kain 2004, p. 288). Allen
Wood also makes Kant a moral realist in part on the basis of an objective conception of rational
will that is independent of the existence of any particular rational agent and that creates a moral
law valid for all rational beings whether they are aware of it or accept it (Wood 1999, p. 157).
But I will show in the following text that Wood also hesitates in a way that allows for a moral
idealist interpretation.
140
Three options
Moral realism or idealism regarding reason and the moral law could be either
transcendental or empirical. As I dened these terms in Chapter 1 on realism,
the transcendental distinction between moral realism and moral idealism
concerns whether the transcendental conception of the moral agent requires
any principles, properties, or objects independent of the conception of moral
agency itself. This transcendental conception is the result of philosophical
work examining the necessary conditions for there to be any moral agents at
all. If all the principles, properties, and objects required for morality can be
said to be dependent upon the conception of the moral agent, then morality
would be transcendentally ideal. If, however, one or more principles, properties, or objects were required for morality and had to be conceived independent
of the moral agent, then morality would be transcendentally real.
The empirical reality of morality could follow from either transcendental
reality or transcendental ideality of morality. Empirical reality considers
whether the actual principles, properties, or objects required for morality as
part of the transcendental assessment of morality would exist independent of
actually existing moral agents. An empirical moral idealism would hold that all
principles, properties, and objects required for morality are entirely dependent
upon the moral agent qua moral subject. Empirical moral idealism would
actually preclude a transcendental realism (this is one way in which these
terms are not used in the same sense that Kant uses them with regard to objects
in space, for Kant takes transcendental realism to imply empirical idealism
[B6971, A36970]).
Here the aspects of morality under discussion are the categorical imperative,
the moral law, and the faculty of reason that generates them. The issue is
whether practical reason itself must be considered transcendentally and empirically real, transcendentally ideal but empirically real, or only empirically
ideal. Since reason is the source of the moral law it, will have the same status
as reason. The categorical imperative is tied both to the fact of reason qua
experience, which is empirically ideal, and to reason if reection upon the fact
of reason can provide justication for its validity.
The strongest interpretation one can give to the claim that reason has a transcendental structure is transcendentally realist. A necessary condition for moral agency
would be that the moral agent is subject to a law that holds independent of the
structure of moral agency, and thus is independent of the transcendental moral
agent. This interpretation would entail a conception of reason that somehow exists
in the nature of things and not merely as the wills or reasons of moral agents.20
20
I take this phrase from Patrick Kain, who holds to this view (Kain 2004, p. 303).
141
21
Kain admits that the conception of realism that stresses reason as the source of the moral law
independent of any particular moral agents does not rule out what he calls global constructivist
anti-realism that relies on the nature of Kants transcendental idealisms restrictions on the
validity of a priori synthetic judgments (Kain 2004, p. 261). The position of the transcendental
ideality of reason can be understood in terms of a global view. He offers only a very brief
objection to global anti-realism and takes a global anti-realism based on transcendental idealism
to require that morality would then, like the categories, have to be valid in nature in space and
time. He takes this to imply an empirical realism. To this extent, his reasoning is correct.
However, he then limits the understanding of nature in space and time to restrict all comprehension of objects in nature in moral ways. He objects that causal determinism in spatiotemporal
nature would preclude the empirical reality of autonomy. My explanation of the structural
account of transcendental freedom of reason shows that the autonomy of reason can describe an
empirically real faculty of reason. He also objects that morality cannot be limited to beings in
nature because morality is supposed to be valid for all rational beings, not simply those who
share our forms of intuition. It is true that the moral law would be valid for all rational beings,
but in the same vein, the pure concepts of the understanding would be valid for any possible
experience. In both cases, they are in practice limited to nature in space and time if, as
naturalists would claim, there are no other forms of intuition and our ontology is restricted to
nature in space and time. Validity for hypothetical beings and for hypothetical nonspatiotemporal existences does not translate to actual existence restrictions, and the global
idealist position would still be valid for all possible rational beings.
142
I discuss the immanent rather than transcendent validity of the ideas of reason in (Rauscher
2010).
143
agents would not work since God would also be a moral agent. I have always
been puzzled by a claim that there can be a faculty like reason independent
of the existence of beings with that faculty. There are two ways that reason
could be independent of rational agents: rst, by being embodied in the things
that exist as their rational structure and second, by existing abstractly as a kind
of relation of thoughts. The rst option, I have already rejected in the previous
paragraph because it would violate Kants critical restrictions on reason. The
second option is essentially identical to the transcendental ideality of reason as
structure and so does not work as an ontological alternative to the conception
of reason as restricted to moral agency. The transcendental conception of
moral agency with regard to reason thus points to a transcendentally ideal
morality.
Transcendental idealism and empirical moral realism
I believe that the interpretation that makes Kant a moral transcendental idealist
and empirical realist would t some of Kants and most of contemporary
Kantian intentions better, and that this version of realism regarding the faculty
of reason and its principles is quite plausible as a philosophical position and
interpretation of Kant.
The reasons that point to a transcendentally idealist and empirically realist
view begin with a look back at Kants claim that no deduction of the moral law
is possible. I believe that Kant is mistaken in denying that a transcendental
deduction can be given for the moral law. Kant had argued that there was no
empirical object in morality and so no possible empirical conrmation for any
possible deduction. It is true that the moral law has no objects of experience for
which it can serve as the transcendental basis, and thus, a transcendental
deduction could not proceed in the same manner as the theoretical transcendental deduction of the categories. But Kant does not have to deny that there is
a kind of object for which the moral law is a basis, namely the experience of a
free rational agent deliberating courses of action. No empirical object is
involved in this deliberation, only a decision about free action. The moral
law could be seen as the necessary basis for any rational deliberation at all.23
I believe that some sort of transcendental argument about the conditions for
agency is what Kantian realists must strive for. If such an argument is
successful, it still ts within a naturalistic framework. The transcendental
justication of the moral law would mean that for a particular kind of experience to be possible in this case, any kind of deliberative decision-making
23
I take Christine Korsgaard to have attempted something like a transcendental deduction in this
sense in (Korsgaard 1996b). Robert Stern provides a positive assessment for a Korsgaardian
transcendental argument for morality in (Stern 2011a).
144
145
perspective that, as part of practical philosophy, has no ontological commitments but is geared only toward determining free actions. The difculty in
interpreting Kant as a transcendental moral idealist and an empirical moral
realist, then, is in his refusal to afrm that human beings can be cognized
objectively as beings with pure practical reason. We might be required to
believe that morality is transcendentally justied, but that is different from a
claim that it really is transcendentally justied.24
Empirical moral idealism
The third possible interpretation would make Kant simply an empirical moral
idealist. Practical reason and the moral law would have no transcendental
status at all and no meaning outside the empirical moral agent. We can believe
but we cannot prove that we are rational beings. Although it is clear that Kant
believes that human beings possess reason, he does not make this belief an
objective theoretical cognition. This point concerns the distinction between the
practical ought and the theoretical is in Kant. Practically, human beings
conceive themselves as possessing the transcendental structure of reason that
demands consistency, uniformity, universality, and the like. Theoretically,
however, human beings cannot prove that they are obligated to the moral
law (the failure of the deduction). We can believe that there is a moral law, that
is, that there is a transcendental structure of pure practical reason. But we
cannot prove that simply because human beings possess pure practical reason,
they are obligated to follow its dictates. Obligation is a practical rather than a
theoretical matter, a matter of ought rather than is.
This empirical moral idealism is not simply based upon the lack of success
of a deduction. The positive aspect of this view is that human beings moral
lives are simply constituted by the perspective they take when faced with
action. As an inescapable perspective, it would not be relative to individuals or
cultures but would be species wide (and extend to any other species with
similar mental structures).
As empirically ideal, the moral perspective would not be objective. With no
deduction to prove that only this point of view is legitimate for active deliberative beings, the possibility would remain that other kinds of deliberative
24
Allen Wood brushes over this important distinction. He claims that his interpretation of Kants
deduction of the moral law draws no distinction between our having to take ourselves, from a
practical standpoint, to be capable of judging according to objective reasons, and there actually
being such reasons for us to judge according to . . . [and] between taking ourselves (from a
practical standpoint) to be capable of setting ends with objective worth and there really being
objective worth for those ends to have . . . [and] our taking ourselves (from a practical
standpoint) to be responding to moral requirements that are unconditionally obligatory and
the actual existence of such categorical requirements (Wood 1999, p. 381, note 30).
146
beings would have other possible determinants of behavior besides just desires
and reason. This possible relativism would not be itself part of the practical
view but would only become apparent when one steps aside from practice and
reects theoretically on the nature of this moral perspective. From within the
moral perspective, no deduction is necessary. I will say more about this
perspective in Chapter 7 in a discussion of the nature of the objective order
provided by reason.
I conclude these observations about the relation of reason and the moral law
to moral realism, then, with the following alternatives. If, on the one hand,
there is a way to conrm that the transcendental structure of moral agency has
more than simply the status of a set of subjectively held but unconrmable
beliefs as part of the practical point of view, then Kant must be a moral
transcendental idealist and empirical realist. The structure of reason would
be known through a transcendental argument but the existence of reason
depends upon the contingent existence of particular deliberative beings. If,
on the other hand, morality is nothing more than the practical point of view
that human (and similar) beings must adopt, and there is no guarantee that it
has any further validity, then Kant must be considered a moral empirical
idealist.
Concluding remarks
In this chapter, I have presented the core of my interpretation of Kants ethics.
I argued that reason as the source of the moral law is the most important faculty
for Kants ethics, more important than the power of decision in the power of
choice. I showed that reason has a place in nature as empirical reason. The
Third Antinomys emphasis on the free timeless causality of reason was
understood in terms of the structure of intelligible reason itself. Transcendental
freedom does not require any non-natural entities but is simply the manner in
which the structure of reason can be understood, and can function as a rational
cause, independently of determination by alien causes. Reasons role in morality, and transcendental freedom itself, can be understood within a metaphysically naturalistic framework. I concluded by arguing that the role of reason in
Kants ethics would t with a transcendental moral idealism and an empirical
moral realism best, although elements of Kants thought also point toward only
an empirical moral idealism. The next few chapters will examine other parts of
Kants moral theory to show how they t within the metaphysically naturalist
framework and moral idealism.
Part III
The two previous chapters looked at the lack of ontology in the primacy of
practical reason and the status of reason within nature, showing that they are
compatible with both a metaphysical naturalism and moral idealism. This
chapter will argue that the postulates of practical reason particularly belief
in God can readily be understood in a metaphysically naturalistic framework.
The argument will also thereby reject any realism about God. I will show that
Kant had two different tendencies in his discussions of the postulates, one of
which focuses on God as an object and the other on the concept of God. (In this
chapter, I will generally use the term concept in discussing the idea of God,
any transcendental hypothesis about God, and the postulate of God; I intend
this term in its broad sense in which concept is usually contrasted with object.)
The latter aspect of his discussion of the postulates, I claim, provides a basis
for the naturalistic interpretation. I also argue that Kants nal conception of
the postulates in the Opus Postumum reects this concern with concept over
object.
Along with the idea of the highest good and as a precondition of its
realization, Kant offers the postulates of immortality of the soul and the
existence of God. His arguments have often been rejected by commentators,
mainly on the basis that the kind of rational hope for happiness Kant stresses is
not really required for moral obligation, thus undermining the practical nature
of these postulates; although recently more positive attention has been paid to
them and to Kants philosophy of religion in general.1 I will not assess the
1
In Chapter 3, I showed how the priority of belief claries the way in which moral obligation is
independent of the postulates. For an early example of traditional criticism of Kants argument
for the postulates, see (Beck 1960, p. 274f.). A defense of Kants argument can be found in
(Wood 1970, p. 129f.). In the past dozen years, a plethora of work on Kants philosophy of
religion has appeared, generally stressing an interpretation that takes Kant to be religious and his
philosophy to require adherence to Christian doctrine. In general, I think that this interpretation
overemphasizes the extent to which some of Kants positions require religion. My chapter will
provide an example of the kind of language that Kant uses in his philosophy that weakens their
claim. Further, Manfred Kuehn has shown convincingly that Kant had no strong personal
religious beliefs (Kuehn 2001). For a philosophical defense of a Kantian atheism, see (Denis
2003).
149
150
151
My interpretation bears some resemblances to Vaihingers; the differences will become apparent
as my argument proceeds. I would like to mention a few key differences here. First, unlike
Vaihinger, I stress that this theory of ctional concepts is for Kant mostly practical; I think that
Kants theoretical philosophy of nature uses ctions only in the regulative use of reason, but not
in physics or mathematics themselves. Kant is an empirical realist about both pure and empirical
concepts of the understanding. Second, I try to be clearer about the possible referential relations
between ctitious concepts and objects and the sense of immanence Kant employs. Third, I take
the ctions to be a priori and not merely part of the empirical psyche. Finally, I do not use the
term ctions but stick with Kants terminology.
Peter Byrne advocates a similar interpretation of the postulate but offers a more comprehensive
view of Kants use of the idea of God overall. He convincingly argues against the realist view of
Kants religious claims and in favor of a subjective, anti-realist view. In particular, he stresses
that Kants critical philosophy precludes any traditional metaphysical view of God and of
religious language (Byrne 2007).
152
At this time, Kant had not yet distinguished practical postulates as such from
other kinds of postulates. He uses the terms postulate and hypothesis
interchangeably, showing that he had not nalized any special or distinct role
for what will become the postulates of practical reason. For example, in
R4113, written sometime from about 1765 to 1775, Kant notes that the idea
of God is a necessary logical ideal, the necessary hypothesis of the natural
order and the necessary hypothesis of the moral order (17:421). Hypotheses are seen as serving both theoretical and practical purposes (see also
R4582, 17:601, and R4928, 18:30).
The postulate is also construed as a concept that must be held by the mind in
order to play some sort of functional role while the existence of any object
corresponding to that concept is not claimed. However, this suggestion is made
not with regard to a practical but with regard to a theoretical proof of God;
theoretically, only a concept of God is attainable, practically, proof of the
object of that concept is also attainable. In a Reection from the early 1770s,
Kant distinguishes between the need for a concept and the need for its proof;
however, the proof in question is a moral proof necessitated by the experience
4
5
For example, see the Inaugural Dissertation, 2:402, for the postulate of pure time, and 2:418, for
the postulate that nothing material at all comes into being or passes away.
The text of Meiers Logic is reprinted in Kants gesammelten Schriften, here 16:668.
153
of moral order rather than any purely a priori reection in pure reason as part of
an explanation of the theoretical pure concept of possibility:
The transcendental concept is necessary, not the transcendental proof; in that one sees
that one cannot make heterodox claims without any ground. The necessity of the divine
being as a necessary hypothesis either of pure concepts of possibility or of experience in
this world, and the latter as hypothesis of morality. Absolute necessity cannot be
proved.
The proof of the existence of God is not apodictic but hypothetical sub hypothesi
logica und practica [under logical and practical hypotheses].
A hypothesis that is necessary in relation to the laws of pure reason is originaria
[original], a hypothesis that is necessary in relation to experience is conditionalis sive
relativa [conditional or relative]. The rst is necessary, the second necessitated; the
former originaria, the latter subsidiaria [subsidiary], e.g. suppletoria[supplementary]
for explanation of order in the world. (R4580, 177275, 17:600)
This moral proof is one that is said to be necessitated rather than necessary,
that is, required for some particular empirical circumstances rather than purely
a priori. The moral order in the world must be explained using a transcendental
hypothesis, yet that hypothesis remains conditional upon or relative to the
alleged moral order in the world. For theoretical purposes, only a transcendental concept of God is allowed, but no proof. When a proof is available, then, it
is only contingent upon a certain empirical fact.
In the Critique of Pure Reason
Even in the Critique of Pure Reason itself, the term postulate is used
theoretically, occurring most often in the name of the three modal principles
of the understanding, the postulates of empirical thought.6 He also calls
reasons search for the unconditioned in the Dialectic, a postulate of reason
(A498/B526). There is no hint of Kants later special status of practical
postulate (although there is a relation to the nature of belief in the Canon
of Pure Reason, which will be discussed later on in this chapter)
The term hypothesis, however, is restricted to theoretical claims. The
section The Discipline of Pure Reason with Regard to Hypotheses is worth
a detailed look. Hypotheses, as Kant denes them, are not merely any kind of
Paul Guyer has argued that the use of the same term postulate here and in the second Critique
is not accidental, and that the practical postulates Kant gives in the second Critique must be
understood in terms of the postulates of empirical thought, in particular that in practical
postulates, the concepts must be understood as possessing real possibility, not merely logical
possibility (Guyer 2000, pp. 33371). This claim reads too much into the use of the same term.
The different status of the postulates of practical reason and the postulates of empirical thought
are clear in their being assigned to practical reason, the practical point of view of the ought,
and empirical thought, the theoretical point of view of the is.
154
155
world of the senses is a mere image, which hovers before our present kind of cognition
and, like a dream, has no objective reality in itself; that if we could intuit the things and
ourselves as they are we would see ourselves in a world of spiritual natures with which
our only true community had not begun with birth nor would not cease with bodily
death (as mere appearances), etc. (A77980/B80708)
The hypothesis, then, is not asserted as true. It is not part of the world-view of
the person who makes this claim. It is asserted as not-impossible. The person
makes this claim only for a particular purpose without actually asserting the
truth of the claim. The particular purpose is the main point for the claim. As
Kant sums it up, Thus one sees that in the speculative use of reason hypotheses have no validity as opinions in themselves, but only relative to opposed
transcendent pretensions (A781/B809).
These two differences between ideas of reason and transcendental hypotheses
should not obscure the similarities. An important similarity to keep in mind is
that both concern concepts that we cannot have insight into in concreto
(A771/B799, modied). Kant notes this with regard to the example he gives
of a transcendental hypothesis. He says strongly We do not know or seriously
assert the least thing about all of this. He also notes that it remains wrong to
extend the principles of possible experience to the possibility of things in
general; that is, it remains wrong to apply the categories and principles of the
understanding beyond the boundaries of possible experience in nature (A781/
B809). Hence transcendental hypotheses cannot be afrmed in any detail.
A transcendental hypotheses, then, is not a reference to an object at all. If
anything, it is a placeholder, that is, it serves to highlight a lack of concrete
theoretical knowledge of objects in nature. This is to say, the transcendental
hypothesis has the function of pointing out to an opponent a lack of complete
156
support for the opponents claims. That is how the transcendental hypothesis is
able to be used as a defense. Since the hypothesis is never afrmed as such,
and in fact cannot ever be afrmed because not known in concreto, its meaning
is equivalent to its function. Its work is merely negative, to remind the
dogmatic opponent that he is making a claim that transgresses over the
boundaries of experience and thus over the boundaries of knowledge. No
ones ontology is extended by this claim. It is meant to remind others of the
limits of their ontology.
I think this is the best way to understand transcendental hypotheses in Kant.
I also think that this is how Kant ought to have, and eventually does, understand the practical correlate of transcendental hypotheses, namely the postulates of practical reason. But Kant does not reach this sort of conclusion right
away. Rather, he explains practical ideas in a way that makes them stronger
than theoretical ideas. Within the discussion of transcendental hypotheses,
Kant says that in regard to its practical use reason still has the right to assume
something which it would in no way be warranted in presupposing in the eld
of mere speculation without sufcient grounds of proof (A776/B 804). Thus,
the negative element of transcendental hypotheses is not at issue for practical
hypotheses. There will be further, positive reasons to afrm something about
the practical hypotheses. The positive reasons, of course, will turn out to be
related to the need for the realization of the highest good.
The canon of pure reason
The rst Critique never uses the term postulate to describe a moral claim and
seems to restrict the term hypothesis to theoretical claims. Still, moral claims
that will later be called postulates of practical reason are raised in the Canon
of Pure Reason chapter. The Canon is a rich resource for assessing the state of
Kants concerns in both ethics and systematic issues; its importance comes in
part from the fact that when Kant wrote the Critique of Pure Reason, he
planned no further critical works before turning to his metaphysics of
morals. Thus, the Canon contains what Kant considered before 1785 as the
important critical discussion of ethics in its broader relation to epistemology,
metaphysics, and his critical system.
The focus for the present topic is rather limited: the status of the concept of
God as a belief rather than a hypothesis. To what extent does Kant claim that
belief in God is a claim that the concept God refers to an actual object? And
to what extent does Kant provide good reasons for claiming any difference
between the practical and the speculative uses of the concept God.
Only reason in its practical use, Kant claims, can have a canon, or the sum
total of the a priori principles of [its] correct use (A7967/B8245). The
whole of the Critique up to this point, Kant claims, shows that speculative
157
reason has no correct use and thus cannot have a canon. (This claim goes too
far, even if one accepts that the negative use of transcendent concepts just
described does not count as a correct use; for the ideas of reason do have a
correct and positive regulative use.7) The only legitimate use of reason then
must be practical. Kant denes the practical use of reason as everything that is
possible through freedom (A800/B828). In Chapter 2, I showed that the
practical is dened as the rules for free action and free behavior, and as
determining what ought to occur as opposed to what is in existence. The
practical, then, concerns moral laws, and thus these alone belong to the
practical use of reason and permit a canon.
The concept of God, then, is permitted because of its relationship to the
moral law. Presumably this means the following: If there is a moral law
stemming from reason, then some other concepts or ideas of reason are so
tied to it that they must be afrmed. That is, one would have reason to use the
concept God if reason already provides moral laws that somehow demand
that the concept God be invoked; this relationship is exhibited in Kants
arguments in the second Critique. Kant initially does not present the problem
this way in the Canon but instead reverses their relationship: These [three
problems of freedom, God, and immortality], however, have in turn their more
remote aim, namely, what is to be done if the will is free, if there is a God, and
if there is a future world(A 800/B828). Such a statement implies that the
existence of God and immortality are to be used as determinants of the moral
law itself or our motivation to follow it.8 Kant states the proper relationship
when asking his three famous questions that summarize philosophy. What
may I hope or alternatively if I do what I should, what may I then hope? is
the link between morality and God (A805/B833). Roughly speaking, this
reverses the previous conditional. Here is the proper Kantian formulation, for
only on this basis can we start with the moral law and proceed to God.
Kant admits that the question he focuses on, what may I hope, is simultaneously practical and theoretical, so that the practical leads like a clue to a
reply to the theoretical question and, in its highest form, the speculative
question (A805/B833). One might object that the canon for practical reason
is misplaced, since it is now a canon that mixes together some elements from
the practical (namely, determination of our duties) with some elements of the
theoretical (determining causal relations and effects, in this case, happiness).
7
8
See my explanation of the positive role of reason in the Critique of Pure Reason, covering both
the Appendix to the Dialectic and the Canon of Pure Reason (Rauscher 2010).
Henry Allison points out that this is a view that Kant ultimately abandoned. Allison identies it
with the claim that the existence of God and immortality are necessary as motivations to follow
the moral law (Allison 1990, p. 67). Strictly speaking, the sentence is ambiguous between
determining ones duties (what is to be done) and looking for motivation toward those duties,
but as Allison points out, the general context favors the claim about motivation.
158
The relation between these two elements lies at the core of the interpretive
controversy over the postulates: determining whether the practical element
stressing action and abstracting from ontology or the theoretical element
stressing ontology should predominate.
The theoretical question, of course, is whether God exists. But even here in
the Canon, there are indications that Kant does not require there to be an
independent being God as substance. And since theoretical reason is at
work, it seems as if Kant ought to invoke some of the same concerns that
arose in the discussion of transcendental hypotheses, in particular, the way that
the concept or idea of reason would not itself be afrmed but only entertained
for the function of making some other claim, or defeating a claim made by
another.
Indeed, Kant hints at this status, although he does not clearly state it. This
comes in the following section of the Canon, On having an opinion, knowing,
and believing.9 For the most part, this section presents belief in God as belief
that there is an object matching the concept God. While denying that human
beings have any knowledge that there is a God, Kant argues that human beings
inescapably believe that there is a God. He contrasts knowing as taking
something to be true on the grounds of both objective and subjective sufciency on the one hand, with believing something to be true on the grounds of
subjective sufciency but objective insufciency on the other hand. Belief in
God, then, appears to be a taking to be true that there is a being God
matching the concept God.
But a closer look reveals some concerns on Kants part to weaken this
interpretation. To see this, it will help to focus on the denition of conviction
as subjective sufciency, which leads to the issue the nature of the practical
relation that lies at the basis of belief.
Kant explains subjective sufciency as conviction, and denes conviction this way: If it is valid for everyone merely as long as he has reason, then
its ground is objectively sufcient (A820/B848) and contrasts this with
persuasion, which is not based on reason but on the particular constitution
of the subject. Insofar as a human beings reason is the basis of the taking to
be true, then, the human being believes something on a subjectively sufcient
basis. But it does not have objective sufciency, for then it would be knowledge. Kant denes objective sufciency as certainty (for everyone) as
opposed to conviction (for myself) (A822/B850). Why would conviction,
This section of the Canon can be interpreted as either a general explanation of levels of assurance
in theoretical claims, as in (Chignell 2007), or as mainly concerned with the practical (Stevenson
2003). Given its place in the Canon, I take Kant to be using this section to explain the peculiar
status of practical beliefs.
159
My moral disposition, that is the relation between my will (or choice) and the
moral law originating in my own reason, is the sole basis of belief in God. As
such, this relation is individual, depending as it does on each individuals own
choices. Kant notes in the following paragraph that an individual who was
indifferent to morality, that is, whose will did not take up a relation of
obligation to follow the law of reason, would not have this conviction, and
for such an individual, the existence of God would be a merely speculative
rather than a practical question.
Kants minimization of belief is perhaps clearest in his claim that the
effect the concept of God would have on a persons moral disposition is
160
what constitutes the belief. He claims that a belief has even less not
more, as one might expect weight than a transcendental hypothesis. Kant
says,
If here too I would call merely theoretically taking something to be true only an
hypothesis that I would be justied in assuming, I would thereby make myself liable
for more of a concept of the constitution of a world-cause and of another world than
I can really boast of; for of that which I even only assume as a hypothesis I must know
at least enough of its properties so that I need invent not its concept but only its
existence. The word belief, however, concerns only the direction that an idea gives
me and the subjective inuence on the advancement of my actions of reason that holds
me fast to it, even though I am not in a position to give an account of it from a
speculative point of view. (A827/B855, long emphasis mine)
Here Kant claims that a transcendental hypothesis will take a previously given
concept and invent (erdichten) only its existence. With a belief I am not in
possession of the appropriate concept to any degree of distinction, let alone am
I in a position to claim that an object exists matching that concept. Instead, the
belief is a functional relation between the idea (or concept) and the advancement of my actions. The belief has the following peculiar relation between idea
and, as it were, its referent: the idea refers not to an object but to somehow to
the individuals moral disposition. The belief functions to strengthen the moral
disposition.
That Kant offers these limitations on belief at least suggests that it is not
appropriate to interpret his conception of belief, the equivalent of a postulate in
the rst Critique, as a straightforward assumption on the practical side of what
could not be proved on the theoretical side. The nature of belief is rather
something different than simple assertion that an object exists to correspond to
a concept.
We see that there are two noteworthy elements to Kants theory of the
practical postulate of God in the period before and in the Critique of Pure
Reason. First, Kant at least sometimes assumed that theoretical and practical
hypotheses were essentially alike. He thought that both of them operated
along similar lines, falling short of proof but instead satisfying reason by
the introduction of concepts. Second, Kant, perhaps simultaneously, argued
that the practical offered a different, or stronger, kind of hypothesis that he
will later call a postulate. When he developed this difference in the rst
Critique in the Canon as a theory of belief, he hinted that beliefs were not
directed toward objects but toward the subject holding the belief. These
elements, when put together, indicate that Kant had a way to develop his
theory of postulates as concepts that do not refer to objects but that
nonetheless play a necessary functional role in our practical lives. He still
insisted, however, that the postulates did more than just that, as we will see
in the next section.
161
10
11
I skip over some Reections from the period between the rst and second Critiques. In R6099
(18:45153), Kant says that a postulate is more than a hypothesis, and that this is because
morality is more than just pragmatic. R6109 (18:45657) and R6111 (18:45859) both stress
that God is a postulate not mere hypothesis, and that this is due to the inescapable moral interest
rather than just the contingent pragmatic interest. These can be taken to mean that postulates are
like hypotheses, except that instead of merely offering them temporarily for defense, they are
held in the mind as a kind of constant support for morality. See also R6283 (18:54950).
My conclusion parallels that of Markus Willaschek, whose detailed look at the postulates of
practical reason also stresses the importance of the concept over the object: as he phrases the
point, the practical extension of our cognition by the postulates does not concern the existence
of God, freedom, and immortality, but only the objective reality of our concepts of God,
freedom, and immortality (Willaschek 2010, p. 191). My chapter was originally published as
an article (Rauscher 2007) before his article was published and I am largely leaving it intact
because my analysis puts the postulates in the context of the course of Kants entire career, even
though Willaschek provides more detail about the postulates in the second Critique.
162
possibility is not at all theoretical, hence also not apodictic, i.e., it is not a necessity
cognized with respect to the object but is, instead, an assumption necessary with respect
to the subjects observation of its objective but practical laws, hence merely a necessary
hypothesis. I could nd no better expression for this subjective but nevertheless
unconditional rational necessity. (5:11n)
Kants attempt to clear things up, as is often the case, only makes them more
obscure. The postulates are said not to postulate an object itself but to postulate the possibility of an object. Yet possibility of objects is a theoretical
matter, not a practical determination. Further, Kant holds that the postulate is
a necessary hypothesis, thereby using a term, hypothesis, that as we have
seen, he had specically contrasted with moral belief in the rst Critique
(A827/B855), although one that he had also used to refer to both theoretical
and moral claims about God in his Reections. Kant also tries in this passage
to explain the nature of the necessity involved as not a necessity with respect to
the object but a necessity with respect to the subject, in fact with the subjects
observation of its objective but practical laws, which he calls a subjective
rational necessity. Kants attempt to clear up what he must have thought was
a confusing explanation of the postulates does not itself shed much light
on them.
The more detailed argument about the postulates comes in the Dialectic of
Pure Practical Reason, beginning with his introduction of the term in a
discussion of the immortality of the soul (5:122) and continuing through the
end of the Dialectic (5:148). I noted earlier that there are two tendencies in this
discussion. The rst is the tendency, easily recognized, to describe the postulates as claiming on behalf of practical reason the existence of objects that
theoretical reason itself could neither prove nor disprove. Kants denition of
postulate shows that the theoretical concepts appear to be at issue: a postulate is a theoretical proposition, though one not demonstrable as such, insofar
as it is attached inseparably to an a priori unconditionally valid practical law
(5:122). The bulk of section VI, On the Postulates of Pure Practical Reason in
General is devoted to showing how the postulates of immortality, freedom,
and the existence of God give objective reality to the related or identical
concepts from the Dialectic of the rst Critique (5:132133). This tendency is
also certainly quite strong in section III Primacy of Pure Practical Reason,
where Kant claims that the primacy of practical reason entails that it can insist
that speculative reason must accept these propositions and, although they are
transcendent for it, try to unite them, as a foreign possession handed over to it,
with its own concepts, implying that speculative reason must accept the
postulates as being of like kind with theoretical concepts (5:120). And of
course there is the abundant use of the term exists, both in reference to the
immortal soul and God (see, for example, 5:122 and 5:124). In fact, on face
value, it appears that this tendency to treat the postulates as the practical
163
12
13
14
Patrick Kain highlights many such passages in his argument that the postulates of practical
reason require a realist interpretation of Kants ethics. He acknowledges that Kant does not
prove the existence of God to any satisfactory extent but only proves the rational indispensability of believing that they are really possible and real (Kain 2006, p. 459). Kain notes that
Kants failure to prove that such supersensible objects are real does not entail that he is
committed to anti-realism about them, nor does the possibility of skepticism. In this, he is
correct. The idealist interpretation of the postulates that I am presenting does not rest upon any
fallacy of that kind, on any move from lack of proof of God to the claim that the postulates do
not require God to exist. Instead what is important is that the actual status of the nature of belief
is given immanent rather than transcendent reference, that belief is related to the content of the
concept rather than the object, and that it is only from the practical point of view that abstracts
from ontological claims that such beliefs are considered. The arguments I present in this chapter
attest to those positive considerations that buttress a naturalist, idealist interpretation of the
postulates.
Adina Davidovich relatedly notes that the function Kant assigns to the postulates would stand
even if God were not to exist. Not Gods existence but only belief in Gods existence is required
for practical purposes. She makes this claim: [E]ven if God does not actually exist, and we
only believe in the existence of God, the practical consequences for obedience to the moral law
are the same. Therefore, a postulate of practical reason does not have to be known to be true,
and probably does not even have to be true, in order to serve its practical purpose (Davidovich
1993, pp. 201).
Guyer and Willaschek note the importance of this claim (Guyer 2000, pp. 36667) and
(Willaschek 2010, p. 192).
164
Kant himself here weakens a crucial step in his argument for the postulate of
God. The existence of God had been postulated in order to serve as the ground
of the correspondence of happiness and virtue precisely as the means to ensure
that this correspondence is possible (5:12425). Kant now concedes that it is
possible that nature itself in accordance merely with its own natural laws
would provide the correspondence (recall that these natural laws will include
laws regarding human behavior). Objective grounds for this possibility
cannot be ruled out, although human beings reason cannot fathom how this
possibility might be conceived in any detail, hence, Kant claims, there are
subjective grounds for human reason to reject the possibility.
Still, the possibility is a live option objectively even if not subjectively for
our reason. Kant calls the existence of two possible objective explanations of
the possibility of correspondence between virtue and happiness this irresolution of speculative reason. Both options are genuine. Both allow for the
possibility of the highest good. Notice how Kant describes these two options as
equal before he invokes a deciding factor in tipping the balance toward the side
of God rather than nature:
The command to promote the highest good is based objectively (in practical reason); its
possibility in general is likewise objectively based (in theoretical reason, which has
nothing against it). But as for the way we are to represent this possibility, whether in
accordance with universal laws of nature without a wise author presiding over nature or
only on the supposition of such an author, reason cannot decide this objectively. Now a
subjective condition of reason enters into this, the only way in which it is theoretically
possible for it to think the exact harmony of the realm of nature with the realm of morals
as the condition of the possibility of the highest good, and at the same time the only way
that is conducive to morality (which is subject to an objective law of reason). (5:145)
Both options are way[s] we are to represent this possibility. Thus there are
two ways that human beings can conceive in the abstract the single possibility
of the highest good: one purely in accordance with natural laws (although
without the details that are apparently beyond our reason) and one invoking
God. Either of these two possibilities will fulll the function of explaining the
possibility of the highest good. Kant claims that human beings are constrained
to represent this function by using the postulate of God because this postulate
is the only one of the two conducive to morality and because this is the only
way theoretically possible for human beings to represent this function.
Although he does not spell this reason out, presumably Kant means to invoke
something about the way that the concept of God is subjectively
165
15
For a discussion and detailed rejection of the arguments Kant gives that an atheist would have
insurmountable difculty with moral incentives related to the highest good, see (Denis 2003).
She suggests at one point that an atheist might in fact be able to picture a non-theistic manner
for the supreme virtue to correspond to happiness, although she does not invoke Kants
distinction between subjective and objective comprehensions of possibility that Kant mentions
at 5:145 (Denis 2003, pp. 21011). For an analysis that defends Kants claim, see (Kielkopf
1997).
166
objects, though it cannot determine them more closely and so cannot itself extend this
cognition of the objects. (5:13435)
Allen Wood argues that Kant cannot possibly mean to deny the synthetic claim of existence
(Wood 1970, p. 148f). But Woods insistence is based on an assumption that Kant must
understand the postulates to be asserting the existence of a transcendent substance. I am arguing
that this assumption is unwarranted.
167
most remarkable something that was quite lacking in the progress of reason on the
path of nature, a precisely determined concept of this original being. (5:139)
Earlier in the second Critique, he held that freedom is immanent practically although
transcendent theoretically (5:105).
168
represented theoretically and positively is not thereby seen; that there is such a causality
is only postulated by the moral law and for the sake of it. (5:133)
Theoretical reason was forced to grant that there are such objects, though it cannot
determine them more closely and so cannot itself extend this cognition of the objects
(which have now been given to it on practical grounds and, moreover, only for practical
use); for this increment, then, pure theoretical reason, for which all those ideas are
transcendent and without objects, has to thank its practical capacity only. In this they
become immanent and constitutive inasmuch as they are grounds of the possibility of
making real the necessary object of pure practical reason (the highest good), whereas
apart from this they are transcendent and merely regulative principles of speculative
reason, which do not require it to assume a new object beyond experience but only to
bring its use in experience nearer to completeness. (5:135)
In these two passages, Kant contrasts immanent with transcendent. Further, in the rst passage, Kant also holds that the postulate is only in reference
to the moral law.
Kants use of the term immanent here can be explained in part by
reference to some of the Reections from earlier periods. In several Reections
from the 1770s and into the mid-1780s, Kant uses the term immanent to
refer to nature in space and time and thus to what can be known, and
transcendent to refer to what goes beyond it and thus to what cannot be
known (R5639, 18:276, 17781788; R6154, 18:470, 178384). At one point
in a draft of the Antinomy of Pure Reason, he contrasts immanent principles,
which restrict themselves to appearances, to transcendent principles, which
stem from the spontaneity of pure reason to give a unity beyond appearances
(R4757, 17:70304, 177577). The First Critique also ts this use of immanent when, in the Dialectic, Kant says We will call the principles whose
application stays wholly and completely within the limits of possible experience immanent (A29596/B352). The meaning of immanent that can be
drawn from these Reections and applied to the second Critique is roughly as
follows: something immanent is something that refers to appearances, that is,
to nature.
Given this meaning of immanent, Kants use of the term to describe the
postulate seems puzzling. How could the concept of God ever be something
that refers to appearances? In both passages, Kant notes that the immanence
attributed to the postulate God is in its relation to highest good as the object
of practical reason that duty commands us to try to create. Our moral duties, of
course, apply to human beings as living in nature in space and time, that is, in
appearance. The postulate of God is immanent, then, in its relation to the
actions of human beings in nature as they strive to create the highest good,
keeping in mind that the possibility of this highest good is for them subjectively represented by the concept God created specically for this purpose
by practical reason itself. I think this is what Kant means by this key sentence
from one of the passages referred to earlier: For we thereby cognize neither
169
the nature of our souls, nor the intelligible world, nor the supreme being as to
what they are in themselves, but have merely unied the concepts of them in
the practical concept of the highest good as the object of our will, and have
done so altogether a priori through pure reason but only by means of the moral
law, and, moreover, only in reference to it, with respect to the object it
commands (5:133, my emphasis). We unify the concepts of God and of
the other postulates together with the concept of the highest good in order to
complete in our minds a picture of the possibility of the latter. But this is done
not by claiming that the concept of God corresponds directly to any object, but
rather that the concept is merely functioning to sustain our hopes in the highest
good as a result of our action in accordance with the moral law, or as Kant says
in reference to the moral law. The concept of God merely functions as a
support for the moral law, and thus in reference to it, and to the immanent
moral life human beings live in nature, rather than to any object that might be
God as a transcendent being.18
One might still wonder how a mere concept can function to satisfy reasons
need to conceive of the possibility of the highest good since, of course, only an
actual God and not the mere concept would be capable of bringing about the
highest good.19 But this is the same kind of question as asking in the theoretical sphere how a concept of God could create the teleological order in the
world, the parallel situation in the Appendix to the Dialectic in the Critique of
Pure Reason. Kant there denies that we suppose the existence of God in itself
but use the concept only in relation to an explanation of the world, albeit a
necessary one. And that shows clearly that the idea of that being, like all
speculative ideas, means nothing more than that reason bids us consider every
connection in the world according to principles of a systematic unity, hence
as if they had all arisen from one single all-encompassing being (A686/B714).
I am claiming that Kants postulate of God can be understood in the same way.
Reason uses the concept as a means for approaching the world in a particular
way. In the theoretical case, the approach is in regard to systematizing our
experience; in the practical case, it concerns moral action. But in both cases,
the actual existence of a being corresponding to the concept is not posited.20
Marcus Willaschek phrases this well when he says with regard to the postulates: the practical extension of our cognition by the postulates does not
concern the existence of God, freedom, and immortality, but only the
18
19
20
Here, I will note that there is yet another strain to Kants discussion in the second Critique about
God. He says, at one point, that the postulate is not a theoretical cognition of the existence of
God but a personal desire that there be a God (5:143). This strain is discussed in detail by Drew
Pierce in (Pierce 2004) where he argues that the postulate of God should be interpreted as a
claim that we ought to believe in God.
I owe this objection to Burt Louden.
I discuss this connection in detail in Rauscher 2010.
170
21
22
23
171
In this passage, Kant invokes the contrast between the mere idea of God and
the existence of God corresponding to that idea. He holds that only the idea is
required for moral purposes, with one small addition, namely that the minimum of cognition that it is possible for God to exist is required. Kant is still
holding on to the claim that the idea must at least possibly correspond to
reality, although he does not require the idea to be thought actually to
correspond to reality.
Even given this minimal theoretical aspect, in the Religion within the
Boundaries of Mere Reason, the practical import of the concept of God, as
well as other related religious concepts and doctrines, is stressed over any
theoretical implications. In fact, Kant even argues that the practical effect of
some doctrines requires that their theoretical details be ignored. For
example, in a long footnote regarding theoretical consideration of an afterlife, Kant comes very close to mocking the idea of hell as punishment
(6:6971). He argues that on the one hand, eternal punishment in hell would
seem by many to be too harsh a punishment for a nite amount of evil, so
evil doers might expect easy escape through last-minute conversion and
divine mercy. But on the other hand, a nite duration of punishment in hell
would seem to some as survivable and tolerable as a tradeoff for a certain
amount of pleasure obtained by evil means in this life. After discussing this
and related matters for two pages, Kant concludes that it is better to simply
ignore these issues: if, instead of [extending it to] the constitutive principles of the cognition of supersensible objects, into which we cannot in
fact have any insight, we restricted our judgment to the regulative principles, which content themselves with only their practical use, human
wisdom would be better off in a great many respects, and there would be
no breeding of would-be knowledge of something of which we fundamentally know nothing (6:71). The point Kant is making is that the detailed
tenets of religion might not survive theoretical scrutiny, and that if their
moral effect depends upon their theoretical truth, then morality will be
harmed. Instead, Kant holds, we should use religious tenets as, in essence,
172
regulative principles for our moral duties, principles we hold in our minds
only to the extent necessary for them to aid our moral duties.
In the Metaphysics of Morals, Kant notes that any duties we might think we
have toward God are really duties to ourselves:
This idea [of God] proceeds entirely from our own reason and we ourselves make it,
whether for the theoretical purpose of explaining to ourselves the purposiveness in the
universe as a whole or also for the purpose of serving as the incentive in our conduct.
Hence we do not have before us, in this idea, a given being to whom we would be under
obligation; for in that case its reality would rst have to be shown (disclosed) through
experience. Rather, it is a duty of the human being to himself to apply this idea, which
presents itself unavoidably to reason, to the moral law in him, where it is of the greatest
moral fruitfulness. In this (practical) sense it can therefore be said that to have religion
is a duty of the human being to himself. (6:4434)
The idea of God, rather than a given being here is said to be of importance,
and only because the idea can be applied to the moral law (Note also that Kant
puts theoretical and practical purposes on the same footing as both holding to
the idea of God without holding to Gods reality.). Later in the book, Kant
stresses that we abstract from the existence of God when utilizing the idea of
God in philosophical morals: The formal aspect of all religion, if religion is
dened as the sum of all duties as divine commands, belongs to philosophic
morals, since this denition expresses only the relation of reason to the idea of
God which reason makes for itself; and this does not yet make a duty of
religion into a duty to God, as a being existing outside our idea, since we still
abstract from his existence (6:487). In morality, we abstract from the existence of God while nonetheless utilizing the idea of God. I think this kind of
relationship can be understood only as a functional relationship in which the
concept of God is, qua concept, playing a role in human moral life.
The functional role of the postulate of God and the consequent limitation of
this postulate to the concept qua concept, is also evident in some of the
unpublished notes and drafts Kant wrote in the 1790s prior to the Opus
Postumum. There is a reintroduction of the idea of as if. This reection from
his notes for lectures on Logic, most likely from the 1790s, can be taken as
paradigmatic of this line of thought:
One cannot obtain or prove objective reality for any theoretical idea except that of
freedom alone, and only because it is the condition of the moral law whose reality is an
axiom. The reality of the idea of God can be proved only through this and thus only
with a practical purpose, i.e. act as if there is a God, thus it can be proved for this
purpose. (R2842, 16:541)
The dating of this Reection itself shows that Kant could have held these
views as early as 1776, but certainly, he held them in the 1790s because this
point is included in Kants Logic book edited by Jsche (9:93). Certainly in the
1790s, Kant has established to his own satisfaction that the idea of God is valid
173
only for the practical purpose of defending the possibility of the highest good,
and here he expresses this position with the hypothetical act as if there is a
God, not the assertoric act because there is a God.24
This position accords with the position Kant laid out in his unpublished draft
essay What real progress has metaphysics made in Germany since the time of
Leibniz and Wolff?, written sometime in between 1793 and 1795 but not
published until after his death. He claims there that the moral arguments for
God, immortality, and freedom are not proof of their truth, nor of any objective teaching as to the reality of their objects, for in regard to the supersensible
this is impossible; instead, it is merely an injunction, subjectively and indeed
practically valid, and in this respect sufcient, so to act as though we knew that
these objects were real (20:298). Regarding whether this also involves existence claims, Kant is less clear. He adds to the as if discussion a claim that we
voluntarily grant them objective reality, but only a few paragraphs later
claims that they have reality in a subjective context but are not knowledge
of the existence of the object corresponding to this form (20:299300). He
also claims that the practical proof is thus really not a proof of Gods
existence absolutely (simpliciter) but only in a certain respect (secundum
quid), namely in relation to the nal end which the moral man has and should
have, and thus with reference merely to the rationality of assuming such a
being; whereby man is then enabled to accord inuence upon his decisions to
an idea which he frames for himself, on moral principles, exactly as if he had
drawn this idea from a given object (20:305). A proof in relation to
something, only in a certain respect but not absolutely does not sound
like a proof of anything but only a subjective expression or need, and Kants
gloss on this claim that we take the idea of God that we have framed for
ourselves and allow that idea to inuence our behavior. The idea that human
beings construct the concept of God in practical reason is further discussed in
the Progress essay. Kant notes that sometimes we forget that these ideas
have been arbitrarily framed by ourselves, and are not derived from the
objects (20:300).
This review of Kants writings in the 1790s shows that before the Opus
Postumum, Kant was making claims that were major elements of the view that
we have only an idea of God and do not also afrm a being, God. Let us now
briey assess his nal position in the Opus Postumum.
24
There is a similar comment, Practical idea as regulative principle, to act as if there were a God
and another world, in a draft headed the whole of critical philosophy that Kant wrote in
November or December 1797 (R6358, 18:683), although later in that same Reection, Kant
seems to indicate that God is to be understood in terms of a metaphysics of the supersensible
(18:685).
174
All ideas of reason are treated the way that I have interpreted the postulate of
God. This ought not to be surprising, given the earlier discussion of theoretical
hypotheses as well as the regulative use of the ideas of reason. The idea God
is but one idea that reason has created to use merely as an idea without any
assertion about Gods existence. It is true that Kant also uses terms to describe
God that might seem to have existential import. He calls God the highest cause
(21:19), a person (21:30), and a highest being (21:33). Yet, at the same time,
he says that God is a thought object and that in philosophy, it is necessary
to derive the predicates contained in that concept analytically whether or
not there may [be] such a substance (21:32). In the context of these claims,
it is clear that Kant does not intend for these predicates to have to apply to
175
any really existing God but that they have the transcendental function as part
of an idea of reason.
Kant also identies this idea of God with pure practical reason itself. Here,
Kant is thinking about the concept of God as a lawgiver. He equates God not as
a substance but as the highest moral principle in me . . . God is moral/practical
reason legislating for itself (21:1445). By equating God with the selflegislation of reason, Kant clearly abandons any conception of God as substance and instead identies God with the functioning of pure practical reason.
Such a conception conforms to the earlier claims that only the idea of God is
necessary, and that it functions within practical reason as a conrmation of the
possibility of the highest good. To say that God is pure practical reason itself
in its moving forces is to admit that the idea of God is all that is needed, and all
that is meant, when postulating God.
The second aspect I want to show in the Opus Postumum is the moral or
practical nature of the idea of God. Obviously, the identication of God with
pure practical reasons self-legislation is strong evidence in this direction. But
Kant also stresses the moral function of the idea of God:
Difference between unconditioned and conditioned duty of practical reason. The former
has God as originator God is thus not a substance to be found outside me but merely a
moral relationship in me. (21:149)
In these notes, Kant discusses God in terms of the motivating force of the
categorical imperative. We human beings create the concept of God in order to
provide ourselves with a concrete idea of the power and authority that we
accord to the moral law.
This reason for postulating God is different from the reasons that Kant gave
in the arguments in the Critique of Practical Reason.25 God in the second
Critique is needed to serve as a representation for the possibility of the highest
good. Here in the Opus Postumum (as in some of Kants other writings and his
course lectures), God is seen as the legislator posited as commanding us to
obey the moral law.26
25
26
Eckart Frster argues that Kants focus on God not as a being but a concept is made possible by
an important change in Kants conception of the highest good. In the Opus Postumum, Kant
views the highest good not as an objective relation among the totality of beings in the world but
a subjective state of a particular being. The highest good no longer is a result to be achieved
beyond human efforts, so no independent God as substance is required. I think Frster is
reading too much into Kants lack of discussion of the highest good (Frster 2000).
There is one passage in this series that does discuss the highest good: That philosophy (doctrine
of wisdom) is called worldly wisdom in German is relevant, for wisdom, the science of
wisdom, aims at the nal end (the highest good). Now this wisdom, in the strict sense, can
only be attributed to God, and such a being must be given at the same time all power; for
without this the nal end (the highest good) would be an idea without reality; so the proposition:
there is a God is an existential proposition (21:149). I admit that this passage is evidence
against my interpretation, in that Kant claims that there is existential claim regarding God
176
The third aspect that I wanted to show in the Opus Postumum passages is the
way that Kant discusses the concept of God as a creation of reason. This is
perhaps the most important aspect for understanding the real intent of Kants
doctrine of the postulates. For here, Kant not only shows that the postulate of
God is a concept without reference to an object, he not only shows that this
concept functions for human beings for moral purposes, but he also shows that
it is essential to the core of Kants ethics, the autonomy of reason, that the
concept of God is one created by reason for its own purposes. Here is one of
Kants clearest statements of this claim:
Transcendental philosophy is the consciousness of the capacity of being the originator
of the system of ones ideas, in theoretical as well as in practical respect [to the right of
this sentence:]. Ideas are not mere concepts but laws of thought which the subject
prescribes to itself. Autonomy. (21:93)
Reason prescribes certain laws of thought to itself. These laws are the ideas of
reason. Kant quite clearly here identies ideas of reason not with concepts
purporting to refer to objects but instead with laws that prescribe how reason
must think. Here in the Opus Postumum, we see that autonomy is understood to
extend to every aspect of the moral, including the ideas that reason requires
to support morality, namely, the postulates. The idea of God is given by reason
to itself for its own purposes. These moral purposes are intrinsic to reason, and
thus the ideas reason creates are not arbitrary but necessary. Kant stresses this
point in a passage that comments on a denition of God:
The concept (thought) of such a being is not an ideal (ctive) but a necessary [one]
emerging from reason in the highest standpoint of transcendental philosophy.
It is not a ction (arbitrarily constructed concept, conceptus factitius) but one
necessarily given by reason (datus).
There is a God. For there is a categorical imperative of duty before which every
knee shall bend in heaven as on earth etc. and whose name is holy, without the
assumption of a substance being permitted which would represent this being for the
senses. (21:64)
177
Patrick Kain is a notable exception. Kain shows how Kant regarded God as a legislator, not an
author (creator) of the moral law in (Kain 2004) and how Kant took human beings to be
required to believe in God in order to represent God as the author of the divine commands (Kain
2005). John Hare, who presents a modied divine command theory based on the weakness of
human nature, is another exception (Hare 1996) and (Hare 2001). Both Kain and Hare offer
realist accounts.
178
moral agent. This aspect of morality is, so to speak, within the moral agent.
Chapter 6 will continue that topic with a look at the freedom of choice of a
moral agent. Moral agency also entails a relation to things that appear to some
to be outside the moral agent herself: in particular, God in the present case and
the value of humanity in other moral agents. It is natural to assume a realist
stance with regard to these things, since God as an object is independent of the
moral agent, and the value of humanity would be a property of other moral
agents and so independent of the moral agent as subject.
When I dened moral realism and moral idealism, I distinguished between
the principles, properties, and objects required for morality and moral principles, properties, and objects. Since God could play a role in non-moral areas
such as explaining the creation of the universe or the ground of truth, God
would be an object required for morality. That is to say that the mere existence
of God does not entail moral realism because even were there no such thing as
morality at all, God would still exist. However, this assessment is incomplete
because the existence of God necessarily includes the existence of justice and
good, and other moral principles and properties as part of the nature of God.
These would be, in my classication, moral properties of the object God and
their existence independent of the moral agent as subject would entail moral
realism.
God as an existing being would be transcendentally real as independent of
the transcendental moral agent. That is, if the postulate about God entails that
God is a really existing being, then God would have to be independent of the
moral agent even if the existence of God is considered a transcendental
condition for morality. Even given my previous chapters argument that the
moral law stems from reason which in turn can be understood only as the
existing reason of particular moral agents considered to have a certain structure, the reality of God would still be needed. Since God is also considered to
be rational, Gods reason would also contain the structure that produces the
moral law and thus, the moral laws existence would not be dependent simply
upon the (nite) moral agent.
The argument I have given in this chapter provides an idealist understanding
of the postulate in which the postulates do not require anything independent of
the moral agent. This interpretation understands the postulate as about the
concept God rather than the being, God. In order to make the moral demand
of the fact of reason coherent with other aspects of the agents self-conception
as a happiness-seeking being, a moral agent is required to believe that the
highest good is possible. In order to picture how the highest good is possible, a
moral agent creates the concept God as, as it were, the placeholder for that
which makes the highest good possible, recognizing however that the highest
good might be possible in accordance with nature alone. This concept is not
intended to refer to any actually existing being. Rather, the moral agent uses
179
This chapter will assess Kants metaphysics of choice in order to see the extent
to which what is usually referred to as freedom of the will is compatible with
a metaphysical naturalism. The previous two chapters already showed that the
faculty of reason and the postulates can be interpreted within a metaphysically
naturalistic ontology; the following chapter will do the same for moral value.
My approach to freedom of the will as choice mirrors my approach to
reason. Regarding reason, I showed that there is an empirical faculty of reason
in human beings in nature, one studied by the science of empirical psychology.
I then considered Kants transcendental arguments that pure practical reason
must be transcendentally free in order to legislate the pure moral law without
the inuence or determination from any causal sources in nature. I argued that
for this requirement, Kant does not need to postulate that pure practical reason
exists as a transcendent thing in itself (or a property of a thing in itself)
ontologically distinct from nature; rather I showed that Kants transcendental
method allows him to identify pure practical reason with the structure of reason
determined philosophically but existing within nature. The transcendental
freedom of reason is not a nonnatural freedom. This current chapter asks the
questions: is there a counterpart to the distinction between pure and empirical
reason in Kants conception of the power of choice? To what extent would a
nonnatural power of choice be needed for ethics? Is there a way to understand
Kants views on freedom of the will without recourse to nonnatural objects or
abilities?
This chapter has two goals. First, abstracting from my naturalist interpretation, I want to show that Kants conception of free choice is not as metaphysically demanding as many assume. I call this part of my argument many to
one because it shows that the metaphysics of free choice requires only one
instance of a nonnatural free act that explains each and every empirical
instance of an agent making a moral decision, instead of many particular acts,
each corresponding to a particular empirical decision. I will show that even
when Kant is insistent that the freedom of choice of the moral agent be free in a
nonnatural way, he requires only one nonnatural free act rather than a multiplicity. Individual choices of the empirical power of choice in nature are
180
181
182
In Kants notation in his own copy of the A edition of the rst Critique he does mention the Wille
twice in the margins of the solution to the Third Antinomy, rst saying that the will is the faculty
of desire in a being with reason and that it is a pure will when pure reason has causality (23:41, at
A538/B566) and second in reference to the specic phrase reason has causality remarking i.e.
is the causality of its objects. This causality is called the Wille. But in transcendental philosophy
one abstracts from the Wille (23:50, at A547/B575).
And of these two, one of them (4:451) does not even appear to mean Willkr as faculty of
choice but is more closely related to willkrlich, or voluntary: Vorstellungen, die uns ohne
unsere Willkr kommen (representations which come to us involuntarily)
183
While Kants own use of the terms in the 1780s does not consistently allow
for a distinction between the two terms, commentators generally agree that
Kant appears to have a rm distinction between Wille and Willkr in mind in
1797s Metaphysics of Morals.3 There he denes and applies his terminology in
the preface to dene Willkr as a faculty to do or refrain from doing as one
pleases . . . [i]nsofar as it is joined with ones consciousness of the ability to
bring about its object by ones action. Wille is dened as the faculty of desire
whose inner determining ground, hence even what pleases it, lies within the
subjects reason (6:213).
In the end, then, Kant distinguished Wille and Willkr in terms which equate
Wille with practical reason as the source of the moral law one possible
ground determining choice to action and Willkr with latitude of decisionmaking in doing or refraining as one pleases the actual determination of
action. I will not take up the details here of exactly how this distinction is
foreshadowed in Kants earlier writings, nor will I attempt to further disentangle Kants terminology. Here I assume that this distinction captures the basic
structure of Kants ethics in which Wille as reason provides a moral law for
Willkr as the power of decision. For the remainder of this paper, I will use the
English terms will, power of choice or simply choice, and reason in
my discussion to translate Wille, Willkr, and Vernunft, respectively,
and note when Kants use of Wille appears to refer to reason instead of
choice when it is otherwise not clear.
To the extent that Wille is free, it is so as reason, the freedom of which has
been shown to concern its structure independent of natural causes. In the
Metaphysics of Morals Kant says that Wille is neither free nor unfree. By this
he does not retract the picture of reason as independent of determination by
laws of nature but rather emphasizes that there is no determining ground for
reason at all (6:213) since it is merely concerned with providing law as
practical reason (6:226). This denial of freedom does have consequences for
understanding autonomy, as I discussed in Chapter 4 on reason, because the
Allison reviews the distinction in (Allison 1990, pp. 12936) but does not emphasize the dual
ontology for Willkr that gives it an empirical role in nature and a nonnatural transcendentally
free existence independent of nature. This approach forces him to gloss many passages in which
Kant talks about the empirical Willkr. For example, forgoing his normal practice of leaving the
term Willkr untranslated, he even draws attention away from Kants admission that there is a
distinct phenomenal Willkr when he quotes from Kants Metaphysics of Morals: Some have
tried to dene freedom of Willkr as the power to choose between the alternatives of acting with
or against the law (libertas indiffereneiae). But freedom of Willkr cannot be dened in this way,
although choice [Willkr!-FR] as phenomenon gives frequent examples of this in experience.
(Allison 1990, p. 133, quoting 6:226). Allison also rejects Kants empirical psychological
account of action determined in accordance with natural causality as studied by empirical
psychology, claiming that Kant rejects freedom of the empirical character. Practical freedom is
instead, for Allison, not to be understood in terms of natural causality at all.
184
law is something that simply stems from reason and is not freely chosen;
autonomy cannot be understood in terms of any kind of choice.
Freedom of the power of choice, the kind of freedom related to decisionmaking, is what must be analyzed here. Kant places human freedom of choice
at least in part in nature, referring to the human ability to act as beings in space
and time in a way similar to animals but not necessitated in the manner that
animals are. In order to nd out what Kant requires of freedom outside of
nature, we must rst see what he allows as (one form of) practical freedom
within nature.
One of the best known descriptions of this practical freedom is in the
solution to the Third Antinomy:
It is especially noteworthy that it is this transcendental idea of freedom on which the
practical concept of freedom is grounded, and the former constitutes the real moment of
the difculties in the latter, which have long surrounded the question of its possibility.
Freedom in the practical sense is the independence of the power of choice from
necessitation by impulses of sensibility. For a power of choice is sensible insofar as it
is pathologically affected (through moving causes of sensibility); it is called an animal
power of choice (arbitrium brutum) if it can be pathologically necessitated. The human
power of choice is indeed an arbitrium sensitivum, yet not brutum but liberum, because
sensibility does not render its action necessary, but in the human being there is a faculty
of determining oneself from oneself, independently of necessitation by sensible
impulses. (A5334/B5612)
185
in a practical sense. Take this passage from the 1792 Dohna metaphysics
lectures. Kant denes, as is typical, stimuli with sensitive impulses and
motives with intellectual.
We are affected by stimuli but not determined. Whoever is determined by motives is
free, for he acts according to the laws of his own reason according to spontaneity and
not according to receptivity. Will is the faculty of desire insofar as it is affected by
representation of a rule . . . The free power of choice is called pure if it is determined
merely by the representation of the law, by the form of conformity to the law. . . The
capacity for determining oneself through pure representations of reason is freedom.
(28:677, modied)
Here representations of a rule are taken to determine the free power of choice,
which is free when determined by motives. Representations, of course, can
occur only in the mind in inner sense. Thus, the empirical power of choice is
affected by stimuli and motives. To the extent it is determined by the motives,
it is free. For similar claims, see 29:89899 (Mrongovius, 1782), R1021
(15:45758, 177379) and R1054 (15:470, 178588).
The power of choice is, on this picture, a natural faculty. It exists in
human beings as appearance. The determination of the power of choice thus
occurs within the causal nexus of nature. But all of this still means that
human beings are determined in accordance with laws of nature as
described in empirical psychology. Kant calls this freedom because of the
role of principles and ideas. Kant uses the term freedom to refer to the
ability human beings have to act in accordance with principles and ideas as
opposed to instinct and sensible impulses. In a Reection from the mid1770s Kant even uses the term blind (blinde) to refer to the determination of the power of choice by sensible impulses: Power of choice is
either blind or free power of choice (brutum oder liberum). . . The blind
power of choice is that which is necessitated and determined through
impulses (stimulos) (R1028, 15:460). This word blinde is the same word
Kant uses in the rst Critique at A51/B75 when he says Thoughts without
content are empty, intuitions without concepts are blind. And just as Kant
distinguished animal from human power of choice, he also distinguished
animal from human cognitive faculties. Animals do not have reason or
understanding but do have the ability to connect representations of outer
sense they have no inner sense according to laws of sensibility as
Kant says in his lectures on metaphysics (28:276, see also 28:690 and
29:879). The link is this: animals do not have concepts or principles,
humans do, and these concepts or principles make knowledge and rationally
determined action possible. If human beings but not animals can act
according to concepts or principles as representations of laws, then the
human power of choice is the one that decides upon actions based upon
principles rather than directly based upon mere sensible impulses of nature.
186
The mere sensible impulses of nature are outer causes (R6931, 19:209) or
instincts (R7220, 19:289, 1780s) or stimulus (Powalski, 27:123) or
inclinations (R6931 19:209, 177678). But human beings are capable of
guiding their behavior according to principles rather than these mere animal
impulses.4 Decisions based upon principles are deemed free because they
stem from empirical reason, even though empirical reason is within the
causal nexus of nature. That same reason considered as a structure independent of natural causal determination is transcendentally free. Even if
Kant ultimately believes that practical freedom is insufcient for morality,
he has still at least provided a picture of human action, which completely
incorporates it into nature. This view is sufcient for the practical perspective of the moral agent (A803/B831). Moral agents are able to view their
decisions in general as determinable by reason rather than only by sensible
impulses. For mere practical purposes of determining oneself to action in
reference to a moral law, this compatibilist conception of freedom is sufcient by itself.
Kant, of course, did not believe that this sort of compatibilist freedom is
sufcient for a complete explanation of morality. He famously dismissed
Leibnizs related notion as the freedom of a turnspit (5:97). The rejection
of this compatibilist, naturalist freedom of choice must have grounds other
than that of identifying and justifying right acts. The following section will
provide those grounds and ascertain whether these grounds require there to be
many or only one nonnaturally free act. I will also examine the metaphysics of
his deeper conception of freedom for the power of decision of the faculty of
choice.
First, a word about the enumeration of the free decisions of the power of
choice. When I claim that I want to reduce the number of nonnatural free
decisions from many to one, I mean to refer only the transcendentally free
acts of the power of choice. The freedom of reason has already been
accounted for. The practical freedom of empirical choice has also already
been accounted for in a compatibilist, naturalist way. The remaining question for my naturalist interpretation is simply what the nonnatural free
decisions of the power of choice would be and how many there are. The
transcendentally free decision would be, like reason, timeless. Exactly how
that timeless choice relates to particular choices in time is the issue. Some
commentators claim that every individual decision of the power of choice is
4
My account of the empirical power of choice is relatively short and only a sketch. A more
extensive explanation of Kants empirical psychology in its relation to the human power of
choice in nature is Patrick Frierson, Kants Empirical Account of Human Action (Frierson
2005). Frierson emphasizes Kants determinism through the role of causal laws that can be
attributed to individual empirical character. See also his book (Frierson 2003) for discussion of
the empirical elements of Kants discussion of motivation.
187
Henry Allison is the most famous proponent of this view. I have discussed his view elsewhere in
this book and will do so later in this chapter. Regarding timelessness, he says that strictly
speaking the transcendentally free decision is not timeless but merely independent of the
conditions of time, meaning independent of causal determinism in nature (Allison 1990,
pp. 513).
Hud Hudson offers such a view that interprets Kant as offering a token-token identity between
transcendentally free acts and the corresponding natural occurrences of these acts (Hudson
1994). Ralf Meerbotes invocation of anomalous monism functions in a similar way (Meerbote
1984).
Allen Wood accepts the single, timeless decision as the basis of an intelligible character, which is
the basis of the empirical character exhibited in each particular action (Wood 1984). My account
in this chapter adopts his general approach.
188
have the ability to do it despite any natural causal determination that would
appear to make the action impossible.8 Kant states the basic claim that ought
implies can in several places.9 In the Critique of Practical Reason Kant
offers the example of a man threatened with execution unless he falsely
testies against an innocent person. Kant asks whether he would resist this
injustice. He judges, therefore, that he can do something because he is aware
that he ought to do it and cognizes freedom within him, which, without the
moral law, would have remained unknown to him (5:30). It appears that the
range of free actions must be quite wide if every action we ought to do is an
action we must be free to do. Consider as an extreme case a drug addict who
ought to stop ruining humanity in her own person by quitting the habit:
ought implies can would indicate that despite the irresistible urges to
continue the drug, she is free at all times to refuse to take it because she
has a power to overrule the causal determinism in nature that otherwise
would cause her to continue the drug. One could apply this to the agent
independently on each occasion on which she succumbs to her addiction,
because on each occasion her empirical Willkr is determined by the sensible
impulses caused by the addiction rather than by reason as it should be. In
order for the agent to be obligated to act against the addictive impulses, she
must be able to act against those impulses despite their necessary determination of her actual decision.
The consequences for Kants theory of freedom, however, are not as severe
as they rst appear. There are two different ways to understand the freedom
associated with can in ought implies can. First, one can hold that for every
particular action that one ought to perform, one actually has the separate and
individual freedom to perform that action in spite of all other natural causation
opposing it, including other acts and the character of the same agent. Even
Robert Stern has recently argued convincingly against a claim that ought implies can in
general and particularly that Kant has a different meaning (Stern 2004). What Stern calls the
strong interpretation of ought implies can gives priority to the can in claiming that the
ought should be restricted on the basis of the actual abilities of agents (whether particular or
considered in general). This strong interpretation would thus make the can controlling
over the ought. For example, if a study of human nature were to show that it is impossible
for an individual to be able to calculate or act upon a utilitarian standard, then no such
standard could be obligatory for human beings. Instead of this interpretation of ought
implies can, Kant is shown to hold to the interpretation that prioritizes the ought such
that the moral law is determined a priori and then applied to human beings. Whatever the
moral law dictates is understood to be within the capacities of human moral agents. This is
the proper interpretation of Kants claims, and it shows that ought implies can implies a
nonnatural ability to act in accordance with moral law regardless of the limitations of natural
causality.
Stern collects eleven passages from Kants published works in (Stern 2004, pp. 535) but does
not include the passage I cite here (he makes no claim to comprehensiveness).
189
10
This should not be overblown to make the agent responsible for all the past causal antecedents
of her actions, which given Kants strict necessity in natural causal law would be vast. Wood
offers the correct response to objections of that sort (Wood 1984, p. 92).
190
Beck is asking Kant whether in fact all actions that ought to be performed, that
are commanded by the moral law, can be performed, that is, are compatible
with the natural order.
Kant replied to this question about a month later:
As for the question, Can there not be actions incompatible with the existence of a
natural order but which are yet prescribed by the moral law? I answer, Certainly! If you
mean, a denite order of nature, for example, that of the present world. A courtier, for
instance, must recognize it as a duty always to be truthful, though he would not remain a
courtier for long if he did. But there is in that typus only the form of a natural order in
general, that is, the compatibility of actions as events in accord with moral laws, and as
in accord too with natural laws, but only as regards their generality, for this in no way
concerns the special laws of any particular nature. (11:348, July 3, 1792)
Kants reply is to suggest that the can involved in ought implies can is a
general ability, that is, that such actions are not incompatible with the concept
of natural laws in general. The implication is that specic circumstances in
which those natural laws apply, and even the specic natures of particular
substances within a natural order, may actually contradict the requirements of
the moral law, even if that moral law is compatible with natural laws in
general. The example he gives of the honest courtier illustrates the point.
A courtier, which we might say is the natural concept of a permanent atterer
to royalty, is subject to the moral command not to lie. But the natural order in
which the courtier exists is one which requires lying as part of the job of
attering royalty. Without lying, there could be no courtiers. The particular
natural order, which includes courtiers, then, is inconsistent with the moral
ought. Ought in this case does not imply an empirical can, for by denition a
courtier cannot always tell the truth. There can be no honest courtiers. Nonetheless the courtier ought to be honest.
One might object that the courtier, qua courtier, may be unable consistently
to tell the truth, but qua human being is certainly capable of telling the truth.
But that would be to misunderstand Kants point. His point is that certain
particular natural laws may be inconsistent, and thus make impossible, certain
moral commands within that particular natural order. If the natural order is
changed, in this case, if the profession of the courtier is disallowed or abandoned, then of course the command may be followed. Given that allowance, it
might be incumbent upon the individual to resign his courtiership, and even to
attempt to change the particular natural order that includes courtiers, but these
presuppose that the person is being understood not qua courtier but qua human
being. There is a difference between the claim Courtiers cannot always tell
the truth, although they ought to and Human beings employed as courtiers
cannot always tell the truth, although they ought to. The rst is true because
the individual in question would cease being a courtier by denition, the
second false false, at least, on the assumption that no other special law of
191
the particular natural order prevents it. However, in both cases the person as
courtier is still subject to the moral law and is still obligated to tell the truth,
because that moral command, although inconsistent with the particular natural
order, is consistent with natural order in general.
The case of the honest courtier can be generalized to include the special laws
of any particular natural order, and this is what makes Kants discussion here
so important. Moral obligation must be in accord with the general laws of
nature but need not be in accord with the particular natural order, or particular
natural laws, or particular natures of particular substances. The categorical
imperative determines what ought to be the case, which actions ought to be
performed, not by basing its assessment on the particular natural order of
things but on the general idea of consistency with laws of natural order in
general. Only then does this picture of how things ought to be get applied to
the particular natural order experienced by human beings. The can of ought
implies can here is the general can of an entire order of nature rather than a
particular can of individual actions.
This more general sense of order is used in the Critique of Pure Reason
when Kant discusses the malicious liar in the solution to the Third Antinomy.
Here he discusses the alternative in which the liar refrains from lying in terms
of an entirely different empirical character caused by a different intelligible
character (although Kant refers to this in terms of reasons intelligible
character):
[One can ask] only: Why has it [reason] not determined appearances otherwise
through its causality? But no answer to this is possible. For another intelligible
character would have given another empirical one; and if we say that regardless of
the entire course of life he has led up to that point, the agent could still have refrained
from the lie, then this signies only that it stands immediately under the power of
reason, and in its causality reason is not subject to any conditions of appearance or of
the temporal series; the difference in time might be a chief difference in appearances
respecting their relations to one another, since these are not things in themselves and
hence not causes in themselves, but it makes no difference to action in its relation to
reason. (A556/B584)
To say that the liar could have refrained from lying is to say that if reason
had had a different intelligible character, that would have determined the
particular action in nature differently through a different empirical character.
The empirical character as a whole is seen as the result of a timeless
intelligible character, and any change in actions from the natural order would
have required a different natural order constituted in part by a different
empirical character, which could result only from a different timeless intelligible character.
To be sure there are passages in which Kant suggests that ought implies
can holds for each particular action. In the Metaphysics of Morals, for
192
193
Eric Watkins has suggested something akin to this picture. In his reading, Kants concept of
causality involves causal powers of objects, or as he puts it, the natures of individual things, that
then are codiable into laws of nature. Individual agents choose their own natures: to say that
personal agents freely choose their own natures is simply another way of saying that personal
agents are responsible for their noumenal [intelligible] and empirical characters, which is
consistent with what common sense dictates, namely that we be at least partially responsible
for our characters (Watkins 2005, p. 336). The nature that an individual chooses determines
her empirical character, the law of her own causality in nature. Because individuals are thus
responsible for creating some of the laws of nature, they are capable of doing what they ought to
do to the extent that such capability is dependent upon their own individual natures (they cannot
violate other laws of nature).
194
195
The act of the agent is described as an act of reason, which could have
determined the agents act differently. Reason is a cause, which is fully
free and this deed is to be attributed entirely to its failure to act. The agent is
responsible for the lie because reason could have instead seen to it that the
agent tells the truth; ultimately, then, reason is responsible.
These passages stressing the action of reason and reason as cause make
sense only if they are seen as taking reason not as an independent being in
itself but as an empirical faculty that does actually cause the empirical decisions of an agents faculty of choice.12 Reason could cause some of the actions
of rational beings, but does not always determine the actions of moral agents.
This conception would, however, not solve the problem of blaming particular
agents. For if blame is placed on reason rather than the agent, the agent ought
not be punished because the transcendental power of reason is beyond her
12
Allison considers but rejects such a view of reason as a cause. First, he correctly recognizes that
Kant does not advocate the choice of an intelligible character in the rst Critique (the intelligible character explains the empirical character, but nothing explains the intelligible character),
so rejects interpretations of the Third Antinomy that attribute such a view to Kant (Allison 1990,
p. 51). However, second, he goes on to interpret the Third Antinomy in terms of his own
Incorporation Thesis, which is based primarily upon the position Allison draws from the
Religion. Allison uses this later reference to incorporation to overrule the plain statements of
Kants Third Antinomy that reason itself is the cause. Allison states that he wants to construe
all of this in such a way that it does not commit Kant to the view that reason is literally a causal
power and to the conception of a timeless noumenal agency, which seems to be inseparable
from this commitment, so he tries to avoid anything that would make reason an efcient cause
(despite such language from Kant even outside the Third Antinomy, for example, A317/B374).
In order to do that Allison interprets Kants discussion of reason simply as the legislating of the
laws to a power of choice that itself either incorporates them into its maxim or does not in a
transcendentally free manner. The freedom of choice is spontaneous and the genuine causal
factor of an action. This interpretation itself cannot be squared with the text of the Third
Antinomy in which Kant discusses intelligible causality only of reason.
196
control. Evil actions are freely decided upon only because reason could have
instead caused the right action to have occurred. There is no corresponding
claim that good actions are freely decided upon because some transcendental
cause could have caused the wrong action to be taken.
As I have interpreted the Third Antinomys emphasis on reason in Chapter 4,
this causal power of reason is understood in a nonontological way. Reason is
an efcient cause in nature as part of the set of psychological determinants of
an individuals empirical power of choice. The empirical faculty of reason is
the empirical character of reason. Transcendentally reason functions as a
spontaneous cause because the very structure of reason itself, which is the
intelligible character of reason, is the ground of the moral law, the representation of which in the empirical mind is reason as a natural cause. The topic in
this chapter provides a different aspect of that relation. If reason as intelligible
cause is the ground of reason as empirical cause, Kant appears to be arguing,
then any transcendental responsibility for actions determined by empirical
reason lies in the structure of reason itself (empirical imputation, as already
mentioned, is a complex of empirical factors). This conclusion raises the
question of the sense in which reason can be responsible for anything.
Indeed, it is inexplicable why reason acts as it does and not in some lesser or
greater way. Kant offers this admission of a fatalism about the course of nature
and agnosticism about reason in the paragraph immediately following the
malicious lie discussion. The key is that we cannot ask why reason is what it
is or why it causes only what it causes: one cannot ask: Why has reason not
determined itself otherwise? But only: Why has it not determined appearances
otherwise through its causality? But no answer to this is possible (A55657/
B58485). With this admission, Kant appears to have abandoned any attempt
to explain transcendental moral responsibility for actions. Both right and
wrong actions occur in the empirical world not because the agent chose to
do those actions but because reason has determined the empirical course of
nature in such a way that they occur. But then moral responsibility seems
entirely divorced from the agent. Kant might have wanted to take this route in
order to avoid difculties with assigning an atemporal nature to moral agents
for purposes of assigning moral responsibility. The result, however, is that he
has avoided moral agents as well.
This is an aspect of the well-known Sidgwick problem, which dates back to
Kants time as an objection raised against Kant by Karl Leonard Reinhold.13
I will not add anything to this discussion beyond noting that this problem for
moral responsibility arises because Kant appears to hold that the only transcendentally free cause which can be postulated is reason, not simply that the
13
The details of Reinholds and Sidgwicks objection and a critique of the standard manner in
which Kant is defended are discussed by (Fugate 2012).
197
will is free only when it is acting rationally. In order properly to attribute moral
responsibility, one wants instead there to be a transcendentally free choice by
the agent, not a transcendentally free causal power of reason.
Perhaps because Kant realized this problem, or perhaps because of other
reasons unknown to us, Kant does offer such a conception of a transcendentally free power of choice. I will discuss this transcendent freedom of the agent
as presented in the Critique of Practical Reason, which I take to exemplify this
conception of freedom for moral responsibility. He introduces this passage
with the claim one must recall what was said in the Critique of Pure Reason
or follows from it as a solution to the clash between freedom and natural
necessity (5:97). Indeed, Kant invokes the Third Antinomy to the extent that it
explains how it is possible for there to be such a thing as transcendental
freedom independent of natural causation. This insistence by Kant appears to
require an ontological interpretation of the Third Antinomy in which the moral
agents power of choice, and not simply the faculty of reason, in some way has
an existence independent of her existence in nature.
Kant goes so far as to call empirical, compatibilist freedom a wretched
subterfuge (5:96) and at bottom . . . nothing better than the freedom of a
turnspit, which, when once it is wound up, also accomplishes its movements of
itself (5:97). After arguing for the insufciency of empirical practical
freedom, Kant determines that genuine freedom requires independence from
the mechanism of nature. Here instead of claiming that reason is free from the
mechanism of nature and thus that reason is the free cause of moral actions,
Kant stresses the existence of the moral agent as a thing in herself independent
of nature. The free cause which is independent of nature is not reason but is the
moral agent as a thing in itself who views his existence insofar as it does
not stand under conditions of time. He goes on:
in this existence of his nothing is, for him, antecedent to the determination of his will,
but every action and in general every determination of his existence changing
conformably with inner sense, even the whole sequence of his existence as a sensible
being is to be regarded in the consciousness of his intelligible existence as nothing but
the consequence and never as the determining ground of his causality as a noumenon.
So considered, a rational being can now rightly say of every unlawful action he
performed that he could have omitted it even though as appearance it is sufciently
determined in the past and, so far, is inevitably necessary, for this action, with all the
past which determines it, belongs to a single phenomenon of his character, which he
gives to himself and in accordance with which he imputes to himself, as a cause
independent of all sensibility, the causality of those appearances. (5:9798)
Here Kant limits this freedom to rational beings, but not to their reason (or to
reason itself independent of them). The moral agent is said to give himself his
character, which appears as his phenomenal existence in nature. His creation of
his character is the basis of the legitimacy of imputation of actions to him in
198
nature. Kant stresses this point in the following paragraphs of this section of
the second Critique, even admitting that this noumenal character might be evil:
[Some people] show from childhood such early wickedness and progress in it so
continuously into their adulthood that they are taken to be born villains and quite
incapable of improvement as far as their cast of mind is concerned; and nevertheless
they are so judged for what they do or leave undone that they are censured as guilty of
their crimes . . . This could not happen if we did not suppose that whatever arises from
ones choice (as every action intentionally performed undoubtedly does) has as its basis
a free causality, which from early youth expresses its character in its appearances
(actions); these actions, on account of the uniformity of conduct, make knowable a
natural connection that does not, however, make the vicious constitution of the will
necessary but is instead the consequence of the evil and unchangeable principles freely
adopted, which make it only more culpable and deserving of punishment. (5:99100)
Here Kant argues that moral responsibility is attributed to the agent as a thing
independent of nature in space and time. Reason is not seen as constituting the
nonnatural free cause of moral actions; in fact, the free nonnatural cause mentioned is the evil and unchangeable principles freely adopted that result in the
single phenomenon of the persons empirical character in nature. The transcendent free adoption of these principles is the ultimate basis of moral attribution.
Despite his invoking empirical considerations or the causality of reason,
moral attribution after all does require for Kant the postulation of a free
nonnatural ability to decide upon principles for action. But as with the claim
about the need for a transcendent causal power to serve as the can in the
principle ought implies can, this freedom need not be distinct for each moral
action. The free transcendent choice that Kant thinks is involved in moral
attribution is the choice not of actions but of principles, presumably maxims.
One principle or maxim can of course serve as the basis for a plurality of
particular actions. So here again it appears that the transcendentally free power
of choice need not be understood as a plurality of transcendent free decisions,
but just one. Transcendental freedom of choice in Kant is not a series of
independent free transcendent decisions but a single free transcendent decision
of the power of choice that has as its effect the entirety of the empirical
character of the individual. This conception reduces the nonnaturalistic
requirements of Kants theory considerably. The next section will explain
how Kant ultimately describes this single free transcendent choice and how
it can be understood in a metaphysically naturalistic way.
The one transcendentally free decision
The previous sections looked at the reasons Kant gives for requiring a free
choice. This section will argue that, as a metaphysical claim on its own, it is
best not to interpret transcendent freedom of choice as allowing that every
199
200
intelligible deed, cognizable through reason alone apart from any temporal
condition; the latter is sensible, empirical, given in time (factum phenomenon)
(6:31). The single supreme maxim is freely chosen in a transcendentally free
way, and each particular empirical deed in time that reects that maxim can be
understood as free ultimately only as a reection of the supreme maxim.
Kant argues, in essence, that the power of choice must be understood as
choosing actions in nature and as having chosen a general course of action
independent of nature. Kant states this problem in terms of the source of human
propensity to evil. The propensity to evil, that is, the consistent placing of selfinterest above the moral law, must have a free origin. Physical propensity, that
is, the consistent decisions of the sensible power of choice in nature, is not free.
Thus the propensity to evil must have some other free source. The free source
would have to be a decision independent of nature. There are two corresponding
understandings of the term deed. There is only one deed independent of
nature, and that is the fundamental choice to follow the moral law or to instead
prioritize self-interest. This deed is the decision which formulates ones highest
maxim. Then there is the understanding of the word deed that applies it to the
particular application of that highest maxim to actions. These deeds in appearance are referred to as free, although presumably determined according to laws
of nature, because they can be traced to a nonempirical origin in the single
transcendent deed which formulates ones highest maxim.14
Combining the views presented in the Critique of Practical Reason and the
Religion results in a plausible account of how Kant can understand the freedom
of the power of choice at two levels in such a way that it requires only one
nonnaturally free choice. The single act of choosing ones character reverberates, as it were, throughout all of the appearances of ones self in nature. The
single phenomenon of ones empirical character can be traced to a single
nonnatural free decision. This nonnatural freedom is not needed at every
particular instance as an independent explanation of each particular action.
Rather, one such free act sufces to serve as the ground for all particular actions
in nature. The transcendent self need not be conceived as involved in a multitude
of free decisions that correspond with a multitude of empirical decisions.
14
Later in the Religion Kant speaks of a revolution in the disposition of a moral agent, referring
to an alteration of the fundamental stance one takes toward morality, the choice to change ones
fundamental disposition from self-love to duty (6:47). If understood as a genuinely independent
second free transcendent choice this account appears to be incompatible with my interpretation.
It is also incompatible with Kants account in the second Critique when he talks about the single
phenomenon of ones character and Kants account earlier in the Religion of a single inexplicable fundamental maxim that must ground all others. If one does allow a revolution in ones
fundamental maxim, then ones unity of character is also threatened in the way I discuss later.
For these reasons I do not think that Kants position on a revolution in character can be taken to
be any more than an idea of reason we hold as support for our efforts to improve ourselves
morally. I do not think it forms part of his metaphysics of free choice.
201
There is a more directly philosophical reason for holding that only one free
transcendent choice is possible for Kant. The unity of identity of an individual
dictates that only one ultimate choice be posited per individual. A moral agents
empirical character is a single phenomenon, and hence has a unity. The unity
of that character must stem from a single cause, that is, there must be a single
decision that results in the single result. Other kinds of causes, such as the concurring causal power of God in creating nature as a whole, would of course be
allowed. But the relevant cause, the decision of the power of choice, can be only
one. To see why, assume for a minute that there is more than one such cause.
Transcendent choices of this kind are decisions that have no explicable basis,
that is, they are themselves the ultimate explanation and cannot be traced to
any other ground. They are, in effect, unrelated to one another, for if they were
related to one another, that relation would be their ground. If there is more than
one such ultimately inexplicable decision, any unity of character would be
merely random since the independent decisions might conict with one another;
were every particular empirical action to have such a corresponding nonnaturally
free intelligible choice, the odds of lack of unity approach certainty. Even if they
did not conict, there would only be a contingent unity of character since if there
were something that necessitated the consistency of these decisions, it would be
a ground of the decisions, but no grounds are allowed for them. Kants conception of the intelligible character as a transcendent cause of the empirical character
seems to require a unity that does not come about merely contingently. Thus, in
order to have a unity of character, only one transcendent free decision is allowed.
This single nonnatural decision is by itself incompatible with metaphysical
naturalism; however, it does at least show that the strong sense of empirical
reality can be given to nature in space and time as a causal system can be
preserved even with the admission of nonnatural freedom. One need not
conceive of nature as mere appearance, existing only as representations, with
ontological priority given to things in themselves. There is no need to assume
that every act in nature has a distinct transcendent cause, or that transcendently
free beings reect upon all the particular circumstances of particular acts in
nature. The freedom that Kant requires is provided by a general decision that is
conceived to occur independently of and in some sense prior to all particular
acts. One is reminded of Platos Myth of Er in which individuals are said to
have an opportunity before birth to choose a life, but after they choose that life
they were condemned to live it out without change.15 Just as Platos myth
15
Henry Allison also notes that this interpretation of Kants conception of freedom resembles the
Myth of Er (Allison 1990, p. 139). He then offers a nonmetaphysical interpretation that he
thinks will avoid the problem. My approach is different because when I maintain the ontological
priority of nature in space and time, I see no room for a transcendental freedom of the faculty of
choice in nature.
202
203
object that this empirical imputation assigns praise and blame to decisions or
actions outside the agents causal power since they are themselves determined in accordance with laws of nature. This is the empirical character,
which is itself integrated into the causal nexus of nature. Kant can justify this
imputation by reference to the single nonnatural decision that forms the
intelligible character and that is said to ground the particular decisions of
the agents empirical power of choice. Transcendentally the agent is responsible for all particular deeds in reference to this one intelligible deed even
though each particular deed is empirically determined in her power of choice
in nature.
Thus it is a valid interpretation to hold that Kant accepts that there is only
one nonnaturally free decision by the power of choice. He has a good philosophical reason to do so, given his conceptions of things in themselves in
relation to nature. The unity of the empirical character is a result of the unity of
the intelligible character, itself determined by a single timeless free decision.
For moral purposes regarding the practice of praise and blame (moral attribution) as well as for determining the application of the principle ought implies
can, we can conceive of any particular act performed by an individual as
causally determined in nature yet as a reection of the single nonnatural
choice. Every particular act performed by an individual is traced to her intelligible character, a timeless principle that is the transcendent cause of that
individuals entire existence in nature.
From one to none: freedom as a postulate
The interpretation I have given shows that the transcendental freedom of
the power of choice can be understood to consist of only one nonnatural
act of the person in herself that results in the entirety of the persons
empirical character and thus all of her actions in nature. For moral responsibility and because ought implies can, we are required, for moral purposes, to assume that somehow ontologically prior to our existence in
nature we ourselves each made a decision regarding what kind of person
we will be morally.
This short section will suggest that even this requirement is susceptible
to a metaphysical naturalism. Since freedom of the will is one of the three
postulates of pure practical reason in the Critique of Practical Reason
along with immortality of the soul and the existence of God (5:132), it is
susceptible to the same interpretation I gave of the postulate of God in
Chapter 5. Kant rarely discusses the postulate of freedom of the will and
the ambiguity of Wille allows that Kant might be referring to the freedom
of pure practical reason as source of the moral law. In this section I am
204
assuming that one can read Kants postulate in reference to the freedom
of choice.16 In this interpretation the postulate is a concept that is held in
the mind for practical, immanent purposes yet is not taken to have any
object as referent. The postulates are limited to concepts in the mind that
are not intended to be existential claims about independent, nonnatural
objects.
I will not repeat the detailed arguments I provided in Chapter 5. In essence
one can consider the claims postulated for practical purposes as concepts
which function to support moral life without ontological implications. Kant
explains this relation as referring to the moral law and only for its own sake in
the Critique of Practical Reason (5:133). For the issue at stake here, the single
free transcendent act of a nonnatural power of choice can be understood as a
mere concept that human beings possess in order to picture for themselves the
inexplicable claim of ultimate responsibility for action (Kant calls it inscrutable, unerforschlichen, 6:21). In the same way that the concept God
functions as a kind of placeholder for any solution for the moral requirement
that the highest good is possible, the concept of a nonnaturally free decision
functions as a placeholder for any solution to the moral requirements of ought
implies can and of moral imputation. No claim to the reality of any nonnatural
cause is made in the postulate, and thus this claim is compatible with a
metaphysical naturalism.
Freedom, however, appears to have a different status than the other two
postulates of immortality of the soul and the existence of God. In the preface to
the second Critique Kant says that freedom is real and that it grounds the
possibility of the other two postulates:
Now, the concept of freedom, insofar as its reality is proved by an apodictic law of
practical reason, constitutes the keystone of the whole structure of a system of pure
16
Consider that the kind of freedom referenced in the postulates as the ground of the moral law
might not refer to the freedom of the faculty of choice but instead to the transcendental freedom
of reason. The claim that freedom functions as the cornerstone of speculative as well as practical
reason cannot refer to the freedom of choice to act for or against the law because that kind of
freedom has nothing to do with speculative reason. Instead the transcendental freedom of
reason, which Kant does at times refer to using the term Wille and that I discussed in
Chapter 4, is what makes the moral law possible by showing that the structure of reason itself
provides principles and ideas and is valid independently of laws of nature. This transcendental
freedom of reason would be able to ground both practical and speculative reason. To say that
this freedom of reason is proved real through the moral law is to say that the fact of reason as
our experience of the categorical imperative is bound up with a claim of its validity. To the
extent that the fact of reason is acknowledged, the freedom of reason is acknowledged as well.
This account is compatible not only with Kants discussion of the fact of reason in the second
Critique (and Chapter 3) but also with Kants insistence in the solution to the Third Antinomy
that it is reasons active role in producing ideas and principles that provides the basis for a claim
to the transcendentally free intelligible character of reason.
205
reason, even of speculative reason; and all other concepts (those of God and immortality), which as mere ideas remain without support in the latter, now attach themselves to
this concept and with it and by means of it get stability and objective reality, that is,
their possibility is proved by this: that freedom is real, for this idea reveals itself through
the moral law. (5:34)
Kant goes on in the next paragraph to differentiate freedom from God and
immortality in that the rst is a condition for the moral law while the latter two
are conditions only for an object of a will determined by the moral law (the
highest good). Being a condition for the moral law itself means that the
requirement does not arise upon reection after the fact of a moral decision
as immortality and God do but is intrinsic to the moral decision itself.
Immortality and God arise only upon the further step of reection upon the
place of morality in the greater scheme of things, a step that lies outside the
inner context of decision making under the moral law.
These considerations give freedom a different status than the other postulates. Freedoms reality is proved though the moral law, freedom is itself a
condition for the moral law, freedom grounds the objective reality of God and
immortality, and freedom even functions as the cornerstone of pure reason as a
whole. Despite the identication of freedom as a postulate, then, it appears that
it might not be possible to interpret it merely heuristically. I think that even
with this special status freedom is best interpreted in roughly the same manner
as immortality and God.
The nature of practical philosophy itself as I showed in Chapters 2 and 3
supports such a reading with regard to freedom. The subject matter of
practical philosophy is free acts considered from the agents perspective as
free decisions. When faced with actual decisions, the moral agent in nature
does not question whether her decisions are free but must view herself as free
in the process of deliberation. The agent does not think about any kind of
freedom and in the moment of deliberation does not have any need for the
nonnaturally free power of choice. Since practical reason is supposed to
concern the ought rather than the is, there need be no ontological claim
at work.
That point, however, is valid only under the priority of action aspect of the
primacy of practical reason. The primacy of practical reason, I showed in
Chapter 3, has two different kinds of meaning in Kants philosophy. The
priority of action stresses the role of practical reason in determining the agent
to right acts. I take this to be operative in Kants examples of particular agents,
and in relation to freedom it is exemplied by the agent-perspective on
deliberation. When deliberating, moral agents always act under the idea of
freedom and while in the grip of the activity of deliberation cannot consider
their decisions as predetermined. Acting under the idea of freedom in this way
has no ontological implications.
206
But the status of freedom is different under the second meaning of the
primacy of practical reason, which I call the priority of belief. The priority
of belief holds that when reecting upon the conditions of human morality
ones concern is with reconciling the claims, that is, the postulates, arising
from the practical use of reason with the limitations on speculative claims
allowed by the rst Critique. I argued that the priority of belief actually
operates under a higher priority for speculative or theoretical reason because
it attempts to incorporate the concepts raised by practical reason into a coherent world view. The agent-perspective assumption of freedom in practice has
no place here. Freedom is instead raised as a condition for the existence of the
moral law when one later reects on the nature of the moral experience and
determines its conceptual requirements.
This origin of the concept of freedom, however, is the same as that of the
other two postulates. All three would arise upon reection after the moral
experience. Since I showed in Chapter 5 that the postulates are to be considered valid not as referring to any objects but only held as concepts in
support of morality, freedom can have that status provided that it can be seen
as stemming from reection upon moral experience in a similar way.
My interpretation of the fact of reason allows for this. The fact of reason is
our experience of the categorical imperative and helps to constitute our moral
experience. When faced with it in actual decision-making we do not question
either its validity or our ability to choose to follow it this is our stance
given the priority of action. But upon reection on that moral experience we
can ask both questions: what would make the categorical imperative really
valid, and whether we could freely choose whether or not to follow it. Our
answer can recognize that the categorical imperative would not be possible
were it not for the transcendental freedom of reason and its normative force
for us would not be coherent were it not for the transcendental freedom of
choice. But since these answers are wrapped up in the question of attempting
to understand our experience of the categorical imperative, we can even say
that our claim that freedom is a condition for the moral law is postulated on
behalf of our experience of the categorical imperative. The reality of
freedom proved by the moral law is simply the practical recognition that in
order to make the experience we have of the fact of reason coherent with our
entire worldview we would need to use the concept of freedom. But the
demand to make our experience coherent still rests on the status of that
experience. And the status of that practical experience is from an agent
perspective. The transcendental freedom of the power of choice is held as a
concept for immanent purposes rather than taken to refer to an actual
property of moral agents. The nonnatural free decision of the power of
choice, then, is subject to the same naturalist interpretation as the other
postulates.
207
Concluding remarks
This chapter has assessed the ontological requirements for the free power of
choice in Kant in order to offer a metaphysically naturalistic interpretation of
Kants moral theory. In addition to the compatibilist view of freedom in nature
that emphasizes the decisions of the empirical power of choice, Kant insists on
a transcendental understanding of freedom of the power of choice. I looked at
two reasons he offers the principle that ought implies can and the requirement of moral responsibility and showed how both can be interpreted to
require only one nonnatural transcendentally free decision. Instead of a metaphysics that requires a nonnatural transcendent free choice for each particular
act, Kant requires only a single act to ground the character that then determines
all the particular acts performed during ones lifetime.
I examined the interpretations of this single transcendentally free act that
best t with metaphysical naturalism. Instead of a focus on how any particular
individual action can be independent of determinism in nature, the focus on
freedom of choice for the metaphysically naturalist Kant ought to be on how
ones overall character can be understood to be independent of determinism in
nature. I think that such an emphasis makes Kants moral theory less burdened
with metaphysical difculties than alternative interpretations that focus on the
transcendental freedom of each particular act distinctly from every other act.
The single-free-act interpretation allows for a view of nature not unlike a
traditional view in which an independent, timeless God creates the whole of
nature, which then exists as real and as causally closed (other than that act of
creation). The multiple-free-act interpretations face the difculty of explaining
how the allegedly timeless transcendentally free cause of the free choice of the
agent in herself is related to the temporally situated natural causal chain of the
empirical agent in nature while still retaining actual empirical determinism as a
real property of a real nature in space and time. More particularly, they must
explain how a timeless agent in herself can be affected by temporally situated
sensations and desires in nature in a way that preserves the timelessness of the
free agent while individuating the free decision to that exact context in nature.
The interpretation of Kant as requiring only a single timeless act vastly
simplies this metaphysical relationship. One still has to picture the near
divine power of an individual agent making a single general choice of what
kind of person to be, but once that admission is made, nature in space and time
would be seen as real and causation according to laws of nature entirely valid
for all events in nature include decisions made by the empirical human power
of choice.
This single nonnatural act is itself still not compatible with metaphysical
naturalism. The naturalist has available a way to avoid such an ontology by
stressing that freedom of the will is a postulate of practical reason. The
208
The postulates of God, immortality, and freedom of choice have in the past two
chapters been shown not only to be fully compatible with a metaphysical
naturalism but also to be best understood using moral idealism. This chapter
will examine the nature of moral value in particular and moral experience in
general. These matters concern the ways in which morality frames our phenomenal experience as spatial and temporal beings in nature. The discussion of
value will show that value is not a property of anything independent of the
moral agent but an ordering of nature by reason. The status of this ordering,
and with it moral experience in general, will be shown to be the ultimately
inexplicable product of reason that fashions the idea of an unconditioned in
morality for its own purposes as in theoretical philosophy it fashioned the ideas
of soul and God as representations of the unconditioned.
My review of the various elements of Kants ethics that individually needed
to be assessed in terms of realism or idealism included two kinds of value. The
value of subjectively chosen ends has already been seen as fully compatible
with both metaphysical naturalism and empirical moral idealism, since the
value of these contingent ends stemmed only from the choice of the empirical
moral agent to pursue them. The value of humanity as an end in itself,
however, appears not to depend upon the empirical moral agent. The categorical imperative in the formula of humanity commands moral agents always to
treat humanity as an end in itself. Reason offers no option to moral agents
regarding the value of humanity as it does regarding the value of freely chosen
ends. I will argue that, despite the necessity and objectivity characteristic of
absolute value, the proper understanding of the value of humanity is in terms
of moral idealism, and that it is compatible with metaphysical naturalism.
This chapter will provide in the rst section a direct discussion of the nature
and status of absolute value in Kant. I will show that one version of the
argument from autonomy succeeds in showing that Kant must reject value
realism. Since value must not precede the autonomous moral law, it must not
be any intrinsic property of things. I will also argue that there is no place in
Kants philosophy for any intrinsic value properties. Moral value is both
transcendentally and empirically idealist. My second section will look in more
209
210
detail at the order of reason. To say that humanity has absolute value is to say
that the order that reason imposes on moral agents choices includes the
requirement that humanity is a limiting condition when choosing subjective
ends. I look at Kants terms intelligible world and moral world to show
that they mean the moral order that reason imposes as a formal ordering of the
objects of nature. Moral experience is understood in terms of the formal rather
than material view of nature. My third and nal section looks at the ultimate
limits that Kant places on the intelligible order of reason by focusing on the
topic of the outer boundary of practical philosophy given in the last sections of
Groundwork III. I conclude by summarizing the nal status we ought to
understand practical reason to have as a faculty that we human beings use as
part of the rst-person view of experience.
Realism and idealism about moral value
The highest value in Kant is the value of humanity, said to be an end in itself.
There is dispute over precisely what Kant means by humanity,1 but I will
largely forgo entering into those disputes and simply refer to humanity as of
absolute value. This section will argue that the absolute value of humanity is
not an intrinsic property but is best understood as dependent upon the order of
ends imposed by pure practical reason and so conforms to moral idealism
rather than moral realism. I will rst look at the relation between value and
autonomy before showing that moral value cannot be an intrinsic property of
objects but instead is an order imposed by reason. Finally I cap off the section
with a discussion of transcendental idealism of value in relation to empirical
realism.
Autonomy and value
A contentious argument in support of the moral idealists claim that even the
absolute value of humanity must be a product of rather than a basis for the
moral law is based on the nature of autonomy.2 Kant divided all moral theories
1
A few examples sufce to show the scope of the debate. Christine Korsgaard holds that it is
rational nature, specically the ability to dene ends for oneself (Korsgaard 1996a). Allen Wood
takes humanity to be the human capacity for pragmatic reasoning (Wood 2008) (See also
footnote 5 to this chapter ). Paul Guyer advocates freedom over rationality as the ultimate value
for Kant (Guyer 2000). Richard Dean argues that humanity is the good will as the only thing
good without qualication (Dean 2006). I believe that the argument I make about the status of
the general concept of absolute value in Kant is compatible with most of these and other
particular conceptions. In the discussion in the text I will note when my claims appear to apply
differently to different conceptions of humanity.
I made this claim very succinctly in (Rauscher 2002, p. 496). For two responses to this argument,
see (Stern 2012, pp. 12122), who nds it too sketchy, and (DallAgnol 2012b, pp. 1619), who
211
into the autonomous and the heteronomous depending on whether the will is a
law to itself or whether some property of the objects of volition is the basis
of the law of the will (4:440). Autonomy is the autonomy of reason considered
as a faculty that is structured so that it produces the moral law of its own.3
Because of this focus on the nature of reason considered as such, autonomy so
understood is a transcendental not an empirical conception; individual choice
with leeway would be an empirical conception of autonomy. The issue here is
whether autonomy understood in this transcendental sense can be the basis for
a claim that the absolute value of humanity must be a result of determination
by the moral law (transcendental idealism of value) or may precede it (transcendental realism of value).
The argument from autonomy for interpreting Kant as a moral idealist
stresses that any basis for morality independent of pure practical reason as
the will of a transcendental moral agent would have to count as heteronomous
and could not be Kants position.4 Since only rational agency is left as the
possible basis for the moral law, and idealism is dened as the claim that all
moral principles, properties, and objects are dependent upon the moral agent,
the proper interpretation of Kants moral theory is idealistic.
thinks it misunderstands the nature of reason (my direct response to the latter is (Rauscher
2012)). The argument is also presented specically in relation to value by Robert Johnson
(Johnson 2007, pp. 14041), who compares it to a Euthyphro-style argument. Others who stress
autonomy as precluding independent criteria for right include (Schneewind 1997), (Reath 2006),
and (Sensen 2013).
By autonomy Kant does not mean individual leeway in choosing what laws to follow or even
individual power in obligating oneself to the law. There is no individual leeway the moral law
is taken to express the structure of reason itself and obligation accompanies the categorical
imperative by means of its imperatival form as the fact of reason. Robert Stern argues that
autonomy can be understood in terms of obligation. He argues that Kant could be a moral antirealist by taking moral obligation to be autonomous. He argues that obligation is idealist and
autonomous because the concepts of duty and obligation apply only to nite, sensible moral
agents. For perfectly rational agents there is no obligation because all of their actions are
automatically determined by pure practical reason. Finite, sensible moral agents are obligated
in virtue of their nitude, not by any kind of choice whether to obligate themselves but simply by
their nature. Stern takes Kant to be a value realist but to allow for this kind of anti-realism for
obligation (Stern 2011b, pp. 8991). But this is a strange limitation for anti-realism. Human
beings are taken to have in common with purely rational agents their reason, and thus in virtue of
their rational nature all rational agents are subject to the moral law. For we nite sensible rational
agents there is simply another added aspect to our natures, namely, the matter of adding sensible
incentives and not of adding anything moral. The manner in which nite moral agents are subject
to the moral law, namely through duty and obligation, stems from their nature. And the manner
in which purely rational agents are subject to the moral law, namely as sole determinant of their
actions, also stems from their nature. Why should one kind of conception of a moral agent
ground an anti-realist claim while the other grounds a realist claim? If the nature of a particular
kind of moral agent grounds an anti-realist claim in one case, it ought to in both.
Henry Allison makes this claim as well when he analogizes autonomy and heteronomy to
transcendental idealism and transcendental realism (Allison 1990, pp. 99100).
212
Those who advocate this defense of an idealism about value stress that any
absolute value of humanity as an end would count as a property of the objects
of volition (4:109) for these purposes and thus be heteronomous rather than
autonomous. Realists can respond by claiming that the value of humanity as an
end in itself is not an object of this sort because it is not something to be sought
or attained by the will the way that happiness is, thus autonomy, the will being
a law to itself, is compatible with that law recognizing the intrinsic value of
humanity as a limiting value.5
The realist response recognizes that the term object of volition at stake
here can mean two different things. First, it can mean the particular object or
goal to be created or attained in action, such as cooking and consuming a
meal.6 Kant apparently has this conception of an object in mind when he
denounces heteronomous moral principles as those that embody the principle
I ought to do something because I will something else (4:441). They are all
based upon some goal, understood as itself valuable to the agent, to be attained
by adherence to a moral principle: self-love, the pleasure of moral feeling,
obedience to God, or even the attainment of individual perfection on the basis
of a (vague) concept of it in reason, as he lists them in the Groundwork
(4:44243). A second meaning of the term object of volition is that the will
would have to treat certain objects as limitations on its actions.7 Here the
object of volition is not merely what is desired but is anything that is the
5
Allen Wood goes further by claiming that autonomy is based on the absolute value of humanity,
the idea being that the value of humanity as the rational will itself is what obligates moral agents
to submit to the dictates of this rational will valued in itself: The idea of autonomy identies the
authority of the law with the objective value constituting the content of the law. It bases the law
on our esteem for the dignity of rational nature, which makes every rational being the moral
legislator (Wood 2008, p. 106). If the argument for value idealism based on autonomy is
successful, as I will advocate, then both the moderate realist position that autonomy does not
preclude an independent absolute value and the more extreme realist position of Wood that
autonomy necessitates an independent absolute value would be shown false. I will not offer a
separate refutation of Woods position. I would like simply to point out that one might doubt that
absolute value of humanity is autonomy as Wood here suggests. Elsewhere in the same book,
citing Kants Anthropology lectures and his Religion, Wood identies humanity as an end in
itself as the absolute value with pragmatic and not moral reasoning, although Wood glosses this
by arguing that humanity in this sense is coextensive with the capacity to give oneself moral laws
(Wood 2008, pp. 889, 94).
Robert Johnson separates what I am treating as one kind of object into two kinds: those to be
produced by the will and those to be realized in action. He takes humanity to be an end of the
second kind because he thinks we can realize humanity by realizing our wills as good wills, or as
he also describes it, actualizing what is only potential (Johnson 2007, pp. 14546). I do not think
these two ought to be differentiated because in both cases the ends are brought into existence
through the action decided upon by the human power of choice. The claim that humanity can be
considered an end to be realized is also irrelevant in this context because the value of humanity,
its being a good, would not be based upon the action that realizes humanity in particular
situations.
Johnson lists this as his second of three types of objects of the will and recognizes that in this
sense humanity would be an object of the will (Johnson 2007, p. 146). Johnson correctly argues
213
content of the deliberation by the will, including the means that might be used
to attain chosen ends. An agent might have a particular reason to avoid certain
means. A desire to full a promise to save the last piece of cake for someone
would limit an individual from using that cake to offer food to a third person.
Kant appears to have this conception of humanity in mind when in the formula
of humanity moral agents are obligated always to treat humanity in a certain
way, as an end, no matter what any other end they might be pursuing (4:429).
In this second understanding of object of volition humanity is not so much to
be promoted or enlarged as it is to be recognized and taken into account.
If the argument for moral idealism based on autonomy proceeds by using
the rst sense of object of volition, then it will not succeed. Certainly
Kant never identies the good of humanity as some good to be produced.
Kant provides what he takes to be a comprehensive list of all possible kinds
of objects as a determining ground of the will not only in the Groundwork
as cited earlier but also in the second Critique as well (5:39), and humanity as an end in itself is not a part of either of these lists.8 But since this
kind of end is not the only possible understanding of the term object of
volition, the argument fails to exclude the other kind of end as a value
independent of the categorical imperative. Further, some interpretations of
humanity that might be understood as taking humanity to be such a value
treat it as something to be promoted in the sense that our actions ought to
enlarge the existence of humanity, understood as freedom or as rational nature,
because of its value; in this way they make Kant into a kind of quasiconsequentialist.9 If humanity functioned as an object of the will in this sense
it would obviously violate the principle of autonomy because it would set
out an object to be promoted, a good, prior to the categorical imperative
in precisely the way that Kant attributes to heteronomous moral principles.
that the value of humanity understood in this sense need not precede the moral law but can issue
from a command of reason.
Oliver Sensen suggests that this lack of inclusion of the absolute value of humanity in Kants list
of ends (which Sensen identies here in terms of the good) is evidence that Kant cannot have
thought that humanity can function as a value grounding the categorical imperative (Sensen
2011, p. 16). This leap is too quick. Kants omission of the value of humanity from these lists
means that humanity would not be the type of end to be created or attained through action, but it
is still possible that humanity could be a different kind of end grounding the categorical
imperative in a different way. There are other reasons to reject the latter claim that I will
discuss.
I take Paul Guyer, who sees absolute value as freedom (Guyer 2000), and David Cummiskey,
who understands Kants valuing of rational nature to allow for a consequentialism (Cummiskey
1996), to offer interpretations of this type. Their conceptions of value are not precisely equivalent to the kind of goods that human beings seek and that Kant condemns as the foundation for
heteronomous moral principles. But there is a sense that the good they identify as humanity is to
be produced. Their conception is similar to Johnsons claim (see footnote 6 to this chapter) that
humanity might be understood as something to be realized.
214
11
12
Wood makes this point as well by differentiating the ends to be produced, which are the subject
of the moral law, and the objective end of humanity, which he takes to ground the principle of
morality (Wood 2008, p. 85).
(Wood 2008, p. 85, my emphasis). A motive to obey the categorical imperative is consistent
with autonomy, just as the feeling of respect as an empirical motivation to perform right acts is
consistent with the purely formal rational origin of the moral law.
He describes this relation as the fundamental principle expresses, or in turn rests on, a
fundamental value but he also describes it in terms of a motive for the categorical imperative
(Wood 2008, p. 55). Humanity can function as that motive because it is the end that rational
beings qua rational beings have already as their subjective end subjective because it is the end
of each particular rational being rather than reason itself and rational agents are thus already
able to consider it in relation to their maxims for actions. Humanity is not an objective motive
until reason declares it to be (4:42829). In a footnote to this passage Kant defers defending his
claim that humanity has this objective status until Groundwork III (the nal section of this
chapter will examine the status of Kants claims in Groundwork III).
215
priori and both stem from the nature of reason. Kants argument for the
formula of humanity rests on the claim that one cannot have a categorical
imperative without a necessary end. There are, however, different ways to
understand the necessity of the end. First, necessary does not imply independent of the agent in a transcendental sense. Kants conception of the a
priori necessary and universal transcendental conditions for the possibility
of experience in the Critique of Pure Reason separates necessity and independence from the transcendental subject of experience. In parallel to this claim,
Kants requirement that there be a necessary end does not dictate that the end
exists independent of the moral agent. Second, the necessity of the end can
have two possible scopes: a necessary end of maxims and a necessary end of
the moral law itself. These two different scopes for the necessity of the end
allow for the categorical imperative itself to be merely formal (4:416) while at
the same time this formal categorical imperative can command an end for
maxims. A look at Kants summary of the formulas of the categorical imperative indicates that Kant intends the necessary end to be an end for maxims: he
says that all maxims have, namely . . . a form, which consists in
universality, . . . a matter, namely an end, and in this respect the formula says
that a rational being, as an end by its nature and hence as an end in itself, must
in every maxim serve as the limiting condition of all merely relative and
arbitrary ends (4:436, rst and last emphases are mine).
The nal result is precisely that ones maxims are limited by the value of
humanity as an end in itself. The way that the value of humanity functions is
not as the legislative basis for the formulation of humanity that a priori limits
the ability of reason to formulate the categorical imperative itself, (which
would violate autonomy). Rather, the very fact that reason is legislating a
categorical imperative raises the issue of the need for some corresponding end
that would operate in relation to maxims, and reason determines that this end is
humanity as a property of rational beings.
Another consideration in rejecting Woods claim that the independent value
is prior to the determination of the moral law is to note that Wood stresses
Kants argument for the principle of humanity, which is just one formula of the
categorical imperative, over the broader context of Kants initial argument for
the categorical imperative as such in Groundwork I and Groundwork II up
through the rst formulation at 4:421. But if one reads the argument for the
principle of humanity in light of Kants more general claim about the merely
formal nature of the categorical imperative, one can see that the value of
humanity is not understood as the basis for formulating the moral law but
only as a consequence of reasons legislation of that law. In that argument Kant
stresses that the categorical imperative concerns merely the form and not the
matter of action (4:416) and requires only conformity with law and not any
condition limiting law as such (4:412, 4:42021). It is not clear why Kant
216
would think that he is able to formulate the categorical imperative without any
reference to any end valuable in itself yet in a manner sufcient to derive and
state the categorical imperative if, as Wood has it, the categorical imperative is
grounded on this independent value in itself. Only after Kant has formulated
the categorical imperative does he then claim that the will requires an end valid
for all rational beings as a motivating ground for the wills selfdetermination (4:427). Kant holds that practical principles are formal if they
abstract from all subjective ends (4:427), that is to say, the categorical
imperative can still be formal if it abstracts not from every end but only from
subjective ends. The formal moral law can be compatible with an objective
motivating ground but not a subjective incentive. It is important to keep this
understanding of ground as motivating ground in mind when interpreting
Kants crucial following paragraph:
But suppose there were something the existence of which in itself has an absolute value,
something which as an end in itself could be a ground of determinate laws, then in it,
and in it alone, would lie the ground of a possible categorical imperative, that is, of a
practical law. (4:428)
When Kant uses the term ground here, it is best to read him as meaning
motivating ground of rather than basis for formulating the categorical
imperative. And in that case Kant is saying that any command to a rational
being to act in some particular way in general, to adopt a maxim that can
serve as a law requires that some end be taken into consideration. Since the
categorical imperative is a general command regarding the tness of maxims,
it requires a general end that reason can provide a priori. Humanity as a feature
of rational beings as such is this general end.
The successful argument from autonomy to the moral idealism of value rests
not upon a claim that humanity would be a heteronomous end to be produced
but upon a claim that it would be an a priori constraint on the legislative
autonomy of reason. An independent value of humanity as the basis of the
moral law would violate autonomy if that value were to be seen as shaping the
actual legislation of the categorical imperative by reason. If instead the value of
humanity is understood as objective, and so an end for all rational beings as
such, only when reason promulgates the categorical imperative and so requires
such an end as a constraint for particular maxims, then autonomy is not
threatened.
These considerations do not preclude some kind of empirical realism about
the value of humanity. The considerations Kant employs in the Groundwork
are at the transcendental level of an analysis of moral agency itself. The
autonomy of reason rules out that the value can precede the legislation of the
moral law but allows that the value is a precondition for the moral laws
existence: Kant says, If, then, there is to be a supreme practical principle and,
217
with respect to the human will, a categorical imperative, it must be one such
that, from the representation of what is necessarily an end for everyone
because it is an end in itself, it constitutes an objective principle of the will
and thus can serve as a universal practical law (4:428). A possible metaphysics of value compatible with autonomy would have the transcendental ideality
but empirical reality of an independent value of humanity. In the next section
I will show that the very idea of an intrinsic moral value of things, including
empirical things in nature, is inconsistent with other elements of Kants
philosophy. I also give a further reason to reject a transcendental realism about
value. Thus no empirically real value property is possible. The following
section will show that absolute value in Kant is only a formal ordering of ends
and not a property of objects.
Absolute value as a nonnatural property
Absolute value, I will show, is in an important way like the contingent value of
chosen ends and is not an intrinsic property of objects or even of moral beings.
In brief, just as contingently chosen ends have value only because particular
moral agents order their ends to reect their desires, humanity in rational
persons has value only because reason orders ends according to its moral law.
Since practical reason is what denes moral agency, the value of humanity is
dependent on moral agency and is thus transcendentally ideal and not transcendentally real. A transcendentally real absolute value would be intrinsic to
objects as they are independently of any relation to transcendental moral
agency.
Certainly some of Kants language suggests that humanity possesses value
as an intrinsic property of rational beings. Kant says that a rational being is
something the existence of which in itself has an absolute value and exists
as an end in itself (4:428, modied). But there is just as much textual evidence
that Kant understands value as dependent upon the moral law or upon reason.
He says that nothing can have a value other than that which the law determines for it (4:436, modied). That the former passages should be understood
under the rubric of the latter passage is suggested by the sentence immediately
after this last quotation, where Kant adds, But the lawgiving itself, which
determines all value, must for that very reason have a dignity, that is, an
unconditional, incomparable value (4:436, modied, my emphasis). The law
determines the value of everything, including itself as a condition for the
possibility of any other value.13
13
Johnson stresses that we should take Kant at his word that nothing means just that,
nothing and acknowledge that the value of lawmaking that determines value is a product of
the lawmaking (Johnson 2007, p. 143).
218
Dueling passages from Kant will not by themselves resolve the interpretive
issue.14 There are several reasons to hold that Kants understanding of the
value of humanity is not a property of objects or of beings independent of the
perspective of the moral agent. First, practical reason provides norms while
theoretical reason concerns ontology. Second, human beings would lack
knowledge of any independent value. Third, even if practical reason were to
be able to determine the existence of a value property in things, it would be
part of the practical point of view and not an independent ontology. Fourth,
even if the thing or property of absolute value were to exist independent of a
property of rational beings, the value of that thing or property would as
normative not t into the ontology of nature but would have to be part of the
practical point of view.
First, existence is prima facie not constituted by normative properties. The
divide between theoretical and practical reason as that between ontology and
normativity needs to be taken seriously as a basis for further understanding of
either. Practical reason might be able to determine that rational beings ought to
be treated as ends in themselves but cannot determine that rational beings
really are ends in themselves. This point is both basic and controversial. My
Chapter 2 on the nature of philosophy showed that the practical is concerned
with determination of free actions and not with the nature of what is.
Second, an independent absolute value could not exist in nature. Oliver
Sensen argues against value realism in part with an epistemological argument
that if value were independently real, human beings would have no access to it
and it would be able to play no part in our moral lives, but his reasoning also
works to show the ontological difculty.15 The ontology of physical nature is
constituted by the laws of nature applied to intuitions. Clearly absolute value is
not something that human beings can sense through outer intuition. Nor could
it be something that is a property of an outer object accessed through inner
intuition via feeling, for several reasons: it would lack necessity and universality as a product of feeling. Feeling also reveals to human beings more about
themselves than about the object because it can at best show that there is
something in the object that causes us to have certain feelings; it no more
reveals the relevant properties in the object any more than aesthetic feeling
reveals that objects possess beauty. Without laws of nature to connect the inner
intuition of feeling to any outer property of value we cannot include any value
property in an ontology of nature based on inner intuition. One might be able
to intuit the value of oneself through inner intuition, although Kant does not
14
15
For the assessment of relevant passages by a realist, see (Stern 2011b, pp. 2640) and by an
idealist, (Sensen 2011, pp. 213, 3951).
(Sensen 2011, 1920). In my paragraph I modify and expand his approach.
219
claim anything like this access to value and it would not make value an
independent property.
Kant does seem to allow for something like a direct moral awareness for
human beings in inner sense. The fact of reason is an experience of moral
obligation in the guise of the categorical imperative as the activity of reason.
This direct awareness, however, does not provide a way for moral agents to
access value properties in nature. The fact of reason itself is not an awareness
of value but of the categorical imperative. At best moral agents would be able
to reason to some independent value on the basis of their awareness of the
categorical imperative. I treat this possibility in more detail in the following
paragraphs. Here I will note that the fact of reason is also said to be the basis
for practical knowledge of freedom (freedom is indeed the ratio essendi of the
moral law, the moral law is the ratio cognoscendi of freedom (5:4)) but not
value. Further, the fact of reason, as discussed in Chapter 3, is part of the
practical point of view and is not intended by Kant to reveal any ontology. For
these reasons value cannot be an intrinsic property of empirical agents in
nature. There can be no empirical value property in nature on this basis.
Lack of value in nature does not rule out the existence of an independent
value of moral agents in themselves independent of nature, that is, transcendentally real as a nonnatural thing in itself, just as reason could in theory
determine that there must be a God. This leads to the third overall reason that
value cannot be an intrinsic property, namely that there would be no way that
moral agents could know such a property through reason and it would thus be
irrelevant for morality. Reason would have to know this property directly or
indirectly through a proof. Direct knowledge would be either active or passive.
Reason is certainly capable of direct knowledge of things that it actively
creates, such as ideas and principles, but if the reason of the transcendental
moral agent were to create the value, that value could not be an independent
intrinsic property of anything. And if value is understood as a product of the
activity of reason, it would thereby be ideal rather than real because it is
dependent on reason. Reason also has no way to passively access the independent property; any passivity on the part of the moral agent, and more
broadly on the part of cognitive beings like us, implies an activity on the part
of the thing to be known and must come to us through the mediation of our
intuition, bringing the argument back to the level of intuition and nature. So
reason would not be able to directly access any transcendentally independent
value property.
The only mode left for reason to know about an independent value property
is indirectly by proof. Kant spent a decade of his life and nearly 900 pages of
text working out his arguments in the Critique of Pure Reason that reason
cannot know anything by a priori proof. The best that reason can do is to create
ideas a priori that can be used heuristically to help to systematize our
220
221
persons because their nature already marks them out as an end in itself
(4:428, my emphasis).
To say that something is an end in itself is to say that it is valuable and vice
versa. This value as end in itself cannot be an intrinsic property of objects.
Being an end, like being a means, is not an intrinsic property of anything. Ends
are only determined by beings who are able to set ends. Empirical moral agents
set ends and by doing so mark them out as valuable. To say that a moral agent
values something is to say that the moral agent holds it as an end (or a means to
the end). Even when a moral agent recognizes herself as an end in itself she
does so qua moral subject in relation to herself qua moral object. The same
relation works at the transcendental level. Pure practical reason as the faculty
of a transcendental moral agent holds humanity to be an end in itself; this
constitutes the value of humanity. Reason does not insist that humanity is an
end in itself because it has some other value. The value of humanity is just
being held by reason as an end. Value as end in itself is not an intrinsic
property of beings with humanity but demands a relation to practical reason.
To some realists this conclusion makes Kants identication of humanity as
having value seem arbitrary. They might say that if humanity does not already
have value, then there would be no nonarbitrary reason for the categorical
imperative to assign it the status of the most valuable end. Thus they would
reject the entire project of examining how reason can create and assign value.
I think that this worry itself is misguided and that, in fact, in two ways the
charge of arbitrariness applies better to a value realist. First, if they are
allowing that there is a reason for the thing of value to have that value, then
their claims arbitrarily assume that it is incorrect for reason to assign value to
an object based on some nonvalue properties that the object has but that it
would be correct for the value itself in the object to be based on nonvalue
properties. Kant always supplies reasons why humanity is of value: because
the nature of rational beings marks them out as an end in itself (4:428), because
humanity possesses autonomy as the ability to give laws (4:43536), and so
on. The value is based on some nonvalue properties regardless of whether
reason assigns that value or it is intrinsic to the object. Since even if value is an
intrinsic property to objects, it still rests on some other features that the object
possesses, then why not allow reason to use the same basis for its assignment
of value to the object? The value realist does not have a nonarbitrary reason to
disallow reason from building value on nonvalue properties yet allow intrinsic
value properties to rest on nonvalue properties.
But perhaps the value realist would deny that there is any basis for the value
property possessed by humanity but instead claim that value is sui generis. In
this case, a second kind of arbitrariness arises. It would be arbitrary if there
were no reason for humanity to have value. In particular if it were the case that
humanity simply was valuable without further explanation possible, then there
222
would be an amazing coincidence that humanity, which includes the selflegislative ability of reason, would be the value that reason is forced to
recognize as the highest value. Reason would be in the state of commanding
all rational beings to respect the highest value, which just happens to be an
activity of reason itself. The value realist position can be described this way
because the value realist in this second kind of case would be denying not only
that reason is the source of assigning value but also that there is any explanation why humanity has value. And further, when the arbitrariness of intrinsic
value is described this way, it makes it clear why reasons assigning value to
humanity is not arbitrary. If reason assigns value to humanity, which includes
the exercise of reason, then reason is assigning value to its own manifestation.
What could be more appropriate for reason to assign value to than itself? Thus
there is a nonarbitrary reason for the faculty of reason to assign value to
humanity. A charge of arbitrariness for the idealist claim about reason as the
source of value is baseless. The tables can instead be turned against the value
realist who is implicated in not having a nonarbitrary reason for the claim that
humanity is of value.
In the following section I explain in more detail what value is in terms of the
order that reason imposes upon the empirical ends of moral agents through the
transcendental order that separates ends in themselves from contingent ends.
The order of reason
The key to understanding what absolute value is lies in stressing that practical
reason itself provides a normative order to things by means of the categorical
imperative. Autonomy implies that the will, as practical reason, itself determines its own law. The law in turn determines what is of absolute value, but
that means only that the actions and ends that are set by the law are ordered by
reason. Value is merely the formal ranking of ends by reason through its
promulgation of moral law. Kant refers to this ordering by reason as an
intelligible order, the formal ordering that practical reason provides to nature
through the categorical imperative. Any absolute value has to be understood in
relation to this formal ordering of ends by reason.
To understand this moral order of reason as a merely formal aspect one
should compare it to the individual empirical moral agents ordering of
contingent ends, such as having a certain job, eating fresh papaya, playing
the piano, or helping a certain charity. Each individual moral agent chooses
and ranks ends that she desires to pursue, thereby creating a subjective order
among the objects and events in nature that constitute either the means for ends
or those ends themselves. The value of these ends is not a property of empirical
objects or events but is merely a shorthand for their place in the formal order
that the agent imposes on nature: for an end to have any subjective value at all
223
is for that agent to include it somewhere in her hierarchy of ends, the greater or
lesser value of ends is their place in that agents hierarchy of ends. Kant puts
the point this way in the Groundwork: The ends that a rational being proposes
at his discretion as effects of his actions (material ends) are all only relative; for
only their mere relation to a specially constituted faculty of desire on the part
of the subject gives them their value (4:429, modied). Kant understands
value in a strictly deationary way as simply the fact that an agent has that end
and ranks it against other ends.
The same conception of value works at the transcendental level for the
absolute value of humanity in relation to the transcendental moral agent.
Instead of the myriad of particular contingent ends of individuals, the transcendental moral agent must abstract from content and face two possible kinds
of ends, necessary and contingent; the latter are simply the possible contingently chosen ends of moral agents. The necessary end is deemed by reason to
be humanity. Pure practical reason imposes its own order on the objects in
nature by ranking these two kinds of ends in the abstract. The ranking is
simple: humanity is always to be ranked higher than any contingent end and it
thus has value as an end in itself. This is the import of the formula of humanity,
but one clearer place where Kant describes this is in his claim that humanity
has dignity beyond all price (4:43435). A price is xed in relation to other
objects, some of which would be considered more desirable than others and
thus command a higher price. The dignity of humanity beyond all price means
that nothing could be the equivalent or be above humanity in price.16 No
appeal to any intrinsic value property is needed. To say that human beings are
to be treated as ends in themselves is to say that when comparing various ends,
there is none higher than that of humanity on the scale provided by reason.
16
Oliver Sensen has recently argued that Kants use of the term dignity reects its traditional
meaning as standing raised up above others and does not refer to any intrinsic property (Sensen
2011, pp. 14373). Sensen makes a detailed assessment of traditional and contemporary uses of
the term dignity and shows that the contemporary understanding of dignity as an intrinsic
value possessed by human beings was not operative in Kants time and is not reected in Kants
use of the term. The traditional meaning of dignity is expressed in the sense of a particular rank
that distinguishes one from others. A dignied person might avoid behavior she feels is beneath
her dignity. Human beings were said to have dignity within nature in virtue of their possession
of certain abilities that animals did not have, such as reason. Dignity did not indicate any
intrinsic value; rather, it indicated a special status that might then make certain treatment
appropriate, although the status and elevation does not require any moral valuation. Two
specic examples of this nonmoral kind of elevation are the dignity of a teacher and the
dignity of mathematics, both of which are terms Kant used (Religion, 6:162 and Pure Reason
A464/B492). This sense of dignity is relational rather than intrinsic. Something is dignied in
relation to other things that do not have the same elevated status. Nothing possesses dignity
intrinsically. Kants use of dignity in his moral writings, Sensen shows, is in line with the
traditional meaning and so need not indicate a reference to any intrinsic moral value. (Sensen
2009). I am indebted to Sensens work on dignity and value in formulating my similar
conception.
224
Reason does not recognize any preexisting value of humanity but reason
assigns that value to humanity through its ordering of ends.
All value, even absolute value, is thus for Kant ideal rather than real. In a
transcendental sense all value depends upon the transcendental moral subject
as a rational being subject to the categorical imperative imposed by reason
itself. In an empirical sense, absolute value is objective in that every rational
being must recognize it, but there is no independently empirically real value
property in persons or objects in nature. This idealism of value is also fully
compatible with a metaphysical naturalism in which the order imposed by
reason is merely a formal ordering of objects and persons in nature. The order
of reason is a reection of the formal character of the categorical imperative.
Through the categorical imperative pure practical reason imposes its own
intelligible order on nature without adding any content to nature except the
actions that reason itself causes.
This order imposed by reason is the same as the order Kant cites in the Third
Antinomy when he says that reason does not give in to those grounds which
are empirically given, and it does not follow the order of things as they are
presented in intuition, but with complete spontaneity it makes its own order
according to ideas in particular declaring certain actions to be necessary and
possible through the causality of reason (A548/B576). It is the same order
mentioned in Groundwork III as the idea of another order and another
lawgiving than that of the mechanism of nature (4:458). Kant also describes
an intelligible order in the second Critique (5:42, 5:8687, and 5:106). As an
ordering of objects, this order of reason is an imposition of form onto the
content of nature. The order of value is not intrinsic to nature but is determined
only through reason and applies to nature. The intelligible order of things is
not an order of intelligible things but an order of things in our experience,
particularly in relation to rational beings.
In the Canon of Pure Reason Kant denes the moral world as the world
as it would be if it were in conformity with all moral laws (as it can be in
accordance with the freedom of rational beings and should be in accordance
with the necessary laws of morality) but which is conceived thus far merely
as an intelligible world . . . [and] is therefore a mere, yet practical, idea which
really can and should have its inuence on the sensible world, in order to make
it agree as far as possible with this idea. (A808/B836). This intelligible world
is in essence an idea of the order of reason actualized. The formal order
imposed on the content of nature would result in a fully moral world in which
all of reasons value ordering is respected and all rational beings follow the
dictates of the categorical imperative. Kant calls the intelligible world an idea.
I discussed the nature of ideas of reason in Chapter 5, but it is worth noting that
in his initial discussion of ideas in the Critique of Pure Reason Kant uses
normative examples of ideas. Platos ideas were, Kant says, preeminently
225
practical. The examples of virtue and of a republic show that ideas can function
as paradigms or standards for actions (A31516/B37173). Through reason
these ideas can become efcient causes (of actions and their objects), namely
in morality and thus can have real effects on the world (A317/B374). The
moral world that Kant calls an idea in the Canon must be understood along
these lines as archetype, but it is more than simply a model since it also
includes the self-conception of moral beings as members of such a world.
In the Canon the moral world is also said to include moral beings. Kant goes
on to say of this moral world:
The idea of a moral world thus has objective reality, not as if it pertained to an object of
an intelligible intuition (for we cannot even think of such a thing), but as pertaining to
the sensible world, although as an object of pure reason in its practical use and a corpus
mysticum of the rational beings in it, insofar as their free power of choice under moral
laws has thoroughgoing systematic unity in itself as well as with the freedom of
everyone else. (A808/B837)
The moral world is not merely the order of ends or results of these actions but
also includes the rational beings themselves united in a kind of mystical body.
A full idea of a moral world, however, does not picture the rational beings
outside of or independent of their places in nature but only insofar as their free
powers of choice are in a systematic connection. The connection among
rational beings in choosing their actions systematically in accordance with
moral laws is just as formal as the lawful ordering by reason itself. Thus the
conception of an intelligible world is one that also includes rational beings
acting as moral agents in the world.
It is clear that moral agents at least represent themselves as being members
of the intelligible world when considering their actions in nature. Kant even
holds that rational beings in the sensible world are also by virtue of their
obligation to the moral law already members of the intelligible world as a
moral world. In the second Critique Kants declares that We are indeed
legislative members of a kingdom of morals [Reich der Sitten] possible
through freedom and represented to us by practical reason (5:82). The moral
world can be created through free acts in the sensible world; insofar as humans
actually do follow the categorical imperative they actually bring about the
partial transformation of the sensible world into a moral world as ordered by
reason.
The Critique of Practical Reason claries this use of the intelligible world
as a partial realization of the order of reason in the sensible world by distinguishing the concepts of the archetypal and ectypal worlds. The direct application of the moral order to the sensible world provides it with the form of an
intelligible world. This law is to furnish the sensible world, as a sensible
nature (in what concerns rational beings), with the form of an intelligible
226
Thomas Auxter argues that the ectypal world performs the function I have assigned to the
archetypal world, namely, that of providing an idea of a morally perfect world to reason (Auxter
1982, pp. 634). His reading relies on a claim that the archetypal world as an idea of reason
cannot include any sensible content and so requires an ectypal world to apply this abstract idea
of moral perfection to sensibility in general. The ectypal world is therefore for Auxter not an
interpretation of actual experience along moral lines but only an idea of a morally perfect world.
On this reading, however, no room is left for the archetypal world as an idea of reason. The
archetypal world, as a moral idea of reason excluding all sensible content, would consist merely
of form and would be the categorical imperative itself (5:33), but would no longer be an idea of
a world. Further, Auxter quotes but does not take into account Kants remark that the ectypal
world must exist in the sensible world (5:43).
227
by the particular moral agent who chooses that end, the objective value of the
necessary end of humanity is not in nature but is merely an intelligible
property conceived by reason as a part of its determination of the categorical
imperative. To say that value is an intelligible property is only shorthand for
saying that an individual (for subjective value) or reason itself (for objective
value) places it above other things in the order of ends each impose on
nature.
The transcendental moral idealist can still insist that there is a kind of
empirical realism about value, although as an objective standard and not as
an independent property. The basis of this claim would lie in the status of
reason in moral agency. A successful transcendental argument for moral
agency would show that all beings with a certain ability, such as the ability
to deliberate prior to choosing to act, would have to have other certain
characteristics, such as a particular structure to the deliberative process which
for Kant would be at least agent-perspective freedom governed by reason
through the moral law as a categorical imperative. Moral agents who are
thereby necessarily governed by the faculty of reason would also thereby
necessarily share the moral experience of an intelligible world, and the
intelligible world and its value properties could be said to be empirically
real in that sense. There would be no intrinsic value or good or right in
objects, but the source of those moral properties in pure practical reason
would give them a legitimacy beyond the individual empirical moral agent,
even beyond all possible empirical moral agents, just as I said in Chapter 4,
that one might consider pure practical reason as empirically real. The key to
this transcendental structure for Kant would be that for all beings, the
capacity to have moral experience would require that the being possess an
objectively valid faculty of reason. The possession of objectively valid reason
would be a fact in nature and the requirements of reason for value
understood as the formal ordering by reason could be said to be empirically
real as part of moral experience.
That empirical reality of an objectively valid reason relies upon a successful
argument for its transcendental necessity for moral agency. But as discussed in
Chapter 4, Kant denies in the second Critique that a deduction that will prove
the reality of the moral law is possible. Even more revealing is that even when
Kant appears to have offered what he took to be a successful deduction in
Groundwork III, his own reections on it actually reveal that he neither
intended nor believed that he had done any more than shown that empirical
moral agents must operate with a certain self-conception that is itself ultimately inexplicable and, further, that it is only imposed by the subjective needs
of human reason. This concession points to an empirical idealism for value as
an objective standard as well as for other elements of moral agency, as my next
section will show.
228
18
See, for example, (Wood and Schnecker 2004, p. 198f) (Schnecker is responsible for the
material on Groundwork III). Schnecker ends his analysis at the end of the Deduction (4:455).
The notable exception is Henry Allison, whose recent commentary includes a thorough
discussion of them (Allison 2011). I will note my agreements and disagreements with his
interpretation in further notes.
229
caution is needed. He often asserts that human beings are members of the
world of understanding (4:453) or things in themselves (4:451), or beings
whose pure activity must be distinct from the sensible world (4:451). The basic
argument Kant provides to escape from the famous apparent circle invokes
transcendental idealism as a ground for division of the self into that which
appears and that which does not appear (4:451452). Human beings, he says,
possess a purely active faculty of reason that justies us in holding that we do
not entirely belong to the world of sense but must also belong to an independent intelligible world. Kants section How is a Categorical Imperative Possible? invokes what looks like an ontological claim: the world of
understanding contains the ground of the world of sense and so too of its
laws (4:453). All these and many other passages point to a theoretical
conclusion about the actual nature of human beings, an ontological claim
about the nature of reality. Many commentators stress these passages in their
interpretations of Groundwork III.19
The proper way to understand the deduction in Groundwork III is to see it in
context. The context comes, rst, from the initial sections of Groundwork III
that set up the problem to be resolved; these sections receive detailed attention
in the commentaries and are familiar enough to everyone to leave unexamined
here. The second part of the context is in the section on the boundary of
practical philosophy given after the deduction itself. Particularly the material
in the last part of Groundwork III shows that the deduction is not to be
understood in ontological terms but instead as an ultimately unsatisfactory
defense of an idea of reason adopted as a self-conception by moral agents,
namely, that moral agents possess an objectively valid pure practical reason.
The term boundary can refer to two different things when understood
from a certain perspective: an inner and an outer boundary. One can consider
the atmosphere of earth, for example, to have an inner boundary where it
begins as the ground and water end and an outer boundary as the atmosphere
19
Dieter Henrich sets the tone in a series of articles in which he claims that Kant seeks to provide a
justication for morality he deems moral insight which must be essentially ontological.
(Henrich 1994). His direct work on Groundwork III is (Henrich 1975). Dieter Schnecker
makes what he calls Kants onto-ethical principle the lynchpin of his ontological interpretation of the argument (Schonecker 1999) and (Wood and Schnecker 2004). Paul Guyer sees
Kant as offering a metaphysical argument (Guyer 2009), where Guyer gives a strongly
metaphysical interpretation of the argument without delving into the material regarding the
boundary of practical philosophy. In earlier work Guyer allows that Kant backs away from a
claim that he has given a theoretical argument and instead assigns freedom the status of an idea
(Guyer 2007a, p. 167). Henry Allison stands out as questioning the validity of any metaphysical
reading. Allison is careful in sifting through the various claims and terminology in Groundwork
III to separate what he sees as a metaphysical reading of transcendental idealism from what ts
his own two-aspect interpretation. Allison admits that much of Kants language appears to
invoke ontology while arguing that a more plausible interpretation is nonmetaphysical (Allison
2011, pp. 34244).
230
thins out and outer space begins. One can picture oneself rising from the sea,
piercing the boundary between water and air, rising through the air, and
eventually leaving air behind and entering empty space. Similarly, although
a national border delimits territorial extent the same way in every location, an
individual relates to that boundary differently when entering and when leaving
a country, as would happen to a tourist driving through one country to reach
another. The specic term used by Kant, boundary of practical philosophy
will also refer to two different things. First, it can refer to a boundary between
theoretical or speculative philosophy and practical philosophy, the conceptual
division where theoretical claims leave off and practical claims begin. This
I will call the inner boundary; Kant refers to it in paragraph 5 of the section
(4:456) when he questions where the boundary of practical philosophy begins.
Second it can refer to the extreme legitimate use of practical philosophy itself
beyond which practical philosophy cannot venture. This latter is what Kant
calls the outermost boundary. I will look at the inner boundary and its
implications for the deduction before turning to the outer boundary and Kants
nal judgment on the deduction.
The inner boundary of practical reason
The general structure of Kants section on the boundaries of practical philosophy is this: the initial paragraphs (15) lay out the distinction between the
nature of the practical and the speculative to show that the practical begins
when ontological speculation ends and practical use of the result of speculation
begins. The middle paragraphs (68) discuss the nature of the practical as
positing an order of things different from that of nature, that is, the legitimate
space for the practical beyond its beginning boundary. The nal paragraphs
(915) concern the determination of the outermost boundary and discussion of
particular questions that can never be answered because they transgress the
boundary.20
Kant provides an initial basis for the inner boundary in the rst paragraph
when he argues that there is a difference between pure concepts of the
understanding, namely here causality, and ideas of reason, namely freedom.
The basis is that the former can be conrmed in experience and the latter not.
20
In his commentary on the Groundwork, Allison holds that only the nal third of this section
concerns the boundary issue. This claim fails to take seriously the way in which the rst
paragraphs of this section begin the argument by setting out the distinction between theoretical
and practical philosophy in relation to the use of reason. The preparation both settles the inner
boundary of practical reason and lays the ground for determination of the outermost boundary
as dependent upon the nature of reason and its ideas. Allison does see the middle part of this
section as concerned with the legitimate domain of practical philosophy (Allison 2011,
p. 348).
231
He concludes this paragraph with the claim that freedom is only an idea of
reason, the objective reality of which is in itself doubtful (4:455). This
distinction is important for two reasons. First, it immediately raises the question regarding how, and whether, the freedom discussed in Groundwork III
and related ideas can be conrmed if not through experience. This issue is at
the heart of Kants rejection of a deduction in the second Critique, and in the
Groundwork Kant concludes that they cannot be conrmed but only ascribed
to oneself as part of our self-conception as rational agents. Second, the
identication of this freedom with an idea of reason already casts into doubt
the interpretation that assumes that Kant is making an ontological claim in his
deduction. In fact, looking back at the conclusion of Kants deduction in
Groundwork III, one sees that Kant already invoked the term idea: categorical imperatives are said to be possible because the idea of freedom not
freedom itself makes us members of an intelligible world; and that beyond
our sensibly affected will we add the idea of the same will but belonging to
the world of understanding, again not a claim that our will is actually such a
pure will in any sense (4:454, my emphasis). The ideas of reason in the
Dialectic of the Critique of Pure Reason were said not to have any reality of
their own but to be illusory, and any positive use of the ideas would come
merely from their immanent use in aiding the understandings task of cognizing our experience. We should expect, then, that Kant is working toward a
claim that the self-ascription of freedom to our wills does not itself involve
claims to reality but instead will aid in comprehending the nature of our
experience as active beings.
The distinction between sensibly conrmed concepts of understanding and
ideas of reason, which are incapable of being conrmed, is not itself the inner
boundary of the practical because speculative philosophy also uses ideas of
reason. What, then, is the proper inner boundary for practical philosophy? 21
Let us compare the roles of speculative and practical philosophy to nd it.
As already noted, Kant presents both speculative and practical reason as
ranging beyond the eld of nature itself, that is, beyond the concepts of the
understanding, to the realm of ideas. The boundary between the sensible and
intelligible is not the inner boundary of the practical.
Kant has also presented them as facing the same difculty: the clash between
freedom and natural causality. The thesis and antithesis arguments in the Third
Antinomy of the rst Critique exhibit the purely speculative origin of this
issue; the practical origin of the issue is shown in the preceding paragraphs of
Groundwork III. The appearance of this issue in both practical and speculative
philosophy shows that the issue itself cannot provide the boundary.
21
Kant uses the terms speculative philosophy and speculative reason, and correspondingly
practical philosophy and practical reason, interchangeably in Groundwork III.
232
Kant also notes that both practical and speculative philosophy insist on
resolving the issue, although with a difference that one might mistake for a
boundary. Neither practical nor speculative reason can simply leave the
problem unsolved. Speculative reason is said in Groundwork III to need to
resolve the problem in order to clear the way for practical philosophy.
Practical philosophy must resolve the problem in order to defend morality
against the fatalist who otherwise would chase away moral science itself
(4:456) and from similar external attacks (4:457). This discussion appears
to introduce an asymmetry that could provide the basis for the boundary:
perhaps the boundary of practical reason begins with the insistence that the
clash between freedom and nature be resolved; perhaps speculative reason
by itself has no interest in this matter. But this is not true: in the Critique of
Pure Reason Kant insisted that speculative reason resolve the antinomies to
defend reason itself from stubborn dogmatism and a hopeless skepticism
(A407/B434). Further, if the insistence on solving the problem were to lie
merely with practical philosophy, that fact alone could not constitute a
boundary but merely be an effect of whatever it is behind the inner
boundary that motivates the insistence. Even if Kant intended to claim in
the Groundwork, contra the rst Critique, that only practical reason has an
interest in resolving the conict, it would provide no more than a hint at the
actual boundary.
Kant offers another difference between practical and speculative philosophy
that points indirectly toward the inner boundary: he assigns speculative but not
practical philosophy the task of resolving this issue (4:456). This is not itself a
boundary: Kant is quite explicit that it cannot yet be said here that the
boundary of practical philosophy begins (4:456).22 By this he means that
the task of resolving the problem is not within the boundaries of practical
philosophy. One might say that this task is, however, the nal geographical
feature in the territory of speculative philosophy.
What then constitutes the inner boundary of practical philosophy? It must
concern whatever motivates practical reason to insist on a resolution to the
clash between freedom and nature and also relate to the exclusive ability of
speculative reason to try to resolve that clash. It must lie in the difference
between the types of question asked by speculative and practical reason, and
the resulting different ways that these questions can be answered. It must then
be the fact that while speculative reason considers what is, practical reason
asks what ought to be. Practical reason then has the motivation to resolve the
22
Jens Timmermann treats this paragraph as if Kant were questioning whether the outer boundary
of practical philosophy begins here, then notes that the beginning is more appropriate
(Timmermann 2007, p. 146).
233
issue because human beings ought to be free in order to be moral, yet it does
not have the capability of resolving the issue by itself because it cannot
determine what is the case.
Here, then, is the inner boundary of practical philosophy. The inner boundary of practical reason in Groundwork III must be the point at which mere
speculation about what is recedes and the moral ought requires the possible
truth of some claim about freedom that entails overstepping the bounds of
nature. This boundary begins where a self-ascription for moral purposes
clashes with nature. One might say that the inner boundary of practical reason
is determined by a) the line between experience as subject to concepts of the
understanding and the content of ideas of reason that go beyond experience,
and b) the requirements demanded by the ought. When one takes up the
practical perspective of the ought in a way that results in an apparent clash
with the concepts of the understanding, one has crossed the inner boundary
and entered the territory of practical philosophy.23
The specic self-ascription that Kant discusses is that of an independence
of reason from subjectively determined causation so that we may consider
ourselves as subject to the demands of reason itself (4:457). I provided
reasons for this interpretation in Chapter 4. This independence of reason
would be possible only if we could consider ourselves in a different order
of things and in a relation to determining grounds of an entirely different
kind than the order of nature and its determining grounds in desires and
sensible stimulations. This is precisely the order of things at stake in my
discussion of value and moral experience in the rst two sections of this
chapter. Since the practical self-ascription of a free will (i.e., causality by
reason) apparently clashes with causality in nature yet is inextricably linked
to the claim that we ought to have a free will, it clearly lies in the territory
of the practical.
The import of this determination of the inner boundary is that it helps to
explain the status of the conclusion of the deduction. When answering the
question How is a categorical imperative possible? (4:454), Kant explains
that the idea of freedom makes us members of the intelligible world, and that
we use the idea of the will belonging to the world of understanding. That these
are mere ideas and not actual empirically valid concepts of the understanding
places them in the realm of ideas the objective reality of which is in itself
doubtful (4:455). We human beings, for practical purposes, must think of
ourselves as endowed with a free will that responds to the dictates of pure
reason that itself gives the law. And the territory in which we adopt practical
23
The discussion about the inner boundary of practical philosophy concerns pure practical
philosophy and not empirical or anthropological practical philosophy, which can be considered
part of nature.
234
ideas is the territory of practical philosophy. Having seen that territory from
the perspective of its inner boundary, Kant turns to its other frontier: the outer
boundary of practical philosophy.
The outer boundary and the incomprehensibility of the practical
The inner boundary shows where practical philosophy starts and to that extent
provides part of the characterization of the nature of practical philosophy. The
outer boundary will complete that characterization by showing the limits of
practical philosophy. As limited, I will show, practical philosophy is necessarily incomplete. There are two reasons for the incompleteness of the ideas of
reason for practical justication.24 First, they cannot themselves be explained
and to that extent remain skeletal assertions rather than fully eshed out
justications. Second, because they are ideas of reason, they are subject to
the restrictions on reason itself regarding its search for the unconditioned. Both
of these reasons point to limitations for the results of the deduction in
Groundwork III.
Kant identies one aspect of the outer boundary in one sentence: By
thinking itself into a world of understanding practical reason does not at all
overstep its boundaries, but it would certainly do so if it wanted to look or feel
itself into it (4:458, modied). Although we are justied on practical grounds
in adopting the idea of a free will and pure reason, we do not know that we are
intelligences, still less do we know how we could be intelligences. To adopt
the idea of the intelligible world is not to hold to its reality. Later in the same
paragraph Kant insists that the intelligible world is only a standpoint that
reason sees itself constrained to take for practical purposes, and that the idea
of the intelligible world (the idea of another order and another lawgiving
than the natural order) is conceived under its formal condition, that is, law as
universality of a maxim (4:458). To go beyond this mere practical standpoint
as idea, and its only applicable practical use in defending the validity of reason
in commanding the will to follow its formal condition of universality, is to
pierce the outer boundary of the practical.
Specically, any attempt to explain how this reason and this will are
possible violates the boundary. Kants explanation of this restriction links
practical reason to speculative reason again, since both attempt to provide
24
Henrich identies two ways that Kant presents the outermost boundary: rst, as the incomprehensibility of how freedom is possible within the intelligible world, and second, as the
incomprehensibility of any relation between freedom in the intelligible world and an interest in
the sensible world (Henrich 1975, pp. 745). He admits that Kant treats these issues as identical
(4:45960) but takes them to be distinct because only one concerns the validity of and the other
motivation for the moral law. But this distinction is irrelevant since Kants point is that nothing
about the intelligible world is comprehensible, whether it relates to the sensible world or not.
235
25
26
In his extensive notes to his translation of the Groundwork into Portuguese, Guido de Almeida
similarly notes the difference between a claim that it is necessary for human beings to
presuppose freedom on the one hand and the possibility of that freedom which is presupposed
on the other. In other words, Kant allows that practical philosophy requires us to adopt ideas
that themselves refer to things that are incomprehensible and perhaps not even possible (Kant
2009, pp. 42728).
Allison describes this conclusion as an admission that Kant has shown only one necessary and
not all the sufcient conditions for the possibility of the categorical imperative (Allison 2011,
p. 359). I would stress that even this one necessary condition is only assumed as possible
because it cannot itself be sufciently explained.
236
apparent here, but the similarity is more important. Both insist on absolute
necessity. The import is explained by Kant as follows:
Now, it is an essential principle of every use of our reason to push its cognition to
consciousness of its necessity (for without this it would not be cognition on the part of
reason). It is, however, an equally essential limitation of this same reason that it can see
neither the necessity of what is or what happens nor the necessity of what ought to
happen unless a condition under which it is and happens or ought to happen is put at the
basis of this. (4:461)
Practical reason and speculative reason share the property of seeking a condition for everything in their explanations. In order to nd an end to this restless
search, they both create an idea of something that is unconditioned, or necessary. Kants explanation in the Critique of Pure Reason of reasons search for
the unconditioned (A32138/B37796) is treated in the Groundwork as applying to practical reason as well. All of reasons ideas are to be understood as
attempts by reason to provide an unconditioned explanation for something.27
Practical reasons idea of an absolutely necessary moral law is mentioned in
this paragraph, but it is not the only practical idea that Kant considers at the
time he wrote the Groundwork. Kant makes a similar point in the Naturrecht
Feyerabend lecture, given summer 1784 when he was writing the Groundwork. The initial fth portion of this lecture course, labeled as Introduction,
provides an overview of some of the claims of the Groundwork, in particular
showing the importance of freedom and the way in which universality and
ends-in-themselves are applied to questions of political right. At one point
early in the lecture Kant argues that in practical philosophy there must be
something necessary, the necessity of which is inconceivable. But rather than
an unconditioned necessary law as in the Groundwork, Kant here discusses an
unconditioned necessary end:
That something must exist as an end in itself and that not everything can exist merely as
a means is as necessary in the system of ends as Ens a se [a being in itself] is in the
series of efcient causes. A thing that is an end in itself is Bonum a se [a good in itself].
What can be considered merely as a means has value merely as a means when it is used
as such. Now for this there must be a being that is an end in itself. One thing in nature is
a means for another; that continues on and on, and it is necessary in the end to think of a
thing that is itself an end, otherwise the series would have no end. (27:1321)
27
Allison claims that the position Kant argues for in the concluding remarks is unique to the
Groundwork (Allison 2011, pp. 36162). Allison claims that the quest for the unconditioned in
practical philosophy does not appear in either the rst or second Critique. Allison is overlooking the key role of the limits of explanation to the intelligible character of reason in the
resolution to the Third Antinomy. As I show later on and have already discussed in Chapter 4,
while Kant does not name pure practical reason as such, he does describe free causality in terms
of the causality of reason and argues that the causality of reason is simply a given that cannot be
questioned further.
237
Although Kant discusses a necessary end in the system of ends rather than
a necessary law, the intent is similar.28 As in the Groundwork, Kant
compares a practical necessity with a speculative concept of a necessary
being as cause. Both are conclusions to which reason is drawn when
considering a series of conditions. In the case of morality it is a meansends series of justications; in the speculative it is the cause-effect series.
In parallel with the Groundwork we should expect Kant to invoke the
incomprehensibility of this necessity as a function of the nature of reason.
Kant does precisely that:
But how a being can be in itself just an end and never a means, is as hard to conceive as
how in the series of causes there would have to be a necessary being. And yet we must
accept both because of the needs of our reason to have everything complete. It lies in the
nature of human reason that it can never have insight into something except as
conditioned, never have insight into something without a ground, and there is no
ground for ens and bonum a se [being and good in themselves]. (27:1321, my
emphasis)
In his Feyerabend lecture, then, Kant is using the same approach as in the
Groundwork by showing how both speculative and practical reason, qua
reason, posit something necessary and unconditioned for the sake of completeness, yet remain dissatised with the result because the necessary being or end
is not fully explained through any conditions of its own.29
These passages from Feyerabend and the Groundwork are in fundamental
continuity with the nal paragraphs of the resolution to the Third Antinomy
where Kant says that there can be no further question of what causes reason to
have the intelligible character that it has. The two unanswerable questions
regarding the causality of reason are rst, why has reason not determined
itself otherwise? and why has [reason] not determined appearances
28
29
Paul Guyer also quotes from this passage from Feyerabend but interprets in in light of his claim
that freedom is the thing valued in itself (Guyer 1998, p. 33).
Without citing and probably without being aware of these passages from the Feyerabend
lecture, Christine Korsgaard attributes a similar kind of reasoning to Kant regarding the way
that pure reasons search for the unconditioned is worked out in both theoretical and
practical reason (Korsgaard 1996a, pp. 117, 119). She mistakenly, however, emphasizes a
difference in that theoretical reason inevitably runs into antinomies while practical reason
does not, which she takes to allow practical reason the liberty to use the idea of the
unconditioned positively. It is true that theoretical reason runs into antinomies when it
demands an unconditioned, but this is true only when the demand is to nd the unconditioned in experience (taken as a transcendentally real world). Reasons search for the
unconditioned also causes it to demand an absolute subject (soul) and absolute being
(God); in addition the solutions to the dynamical antinomies posit an unconditioned cause
and unconditionally necessary being. The parallel between theoretical and practical is more
important: the ideas of an unconditioned cause, unconditioned subject (soul), unconditioned
end, or unconditioned law are all created by reason as placeholders for its own satisfaction
and cannot be said to have any corresponding object in nature.
238
There is one difference between this conclusion and those in Feyerabend and Groundwork. In
the latter Kant makes an analogy between the practical problem of searching for a condition for
a necessary end or necessary law and the speculative search for a being in itself or supreme
cause of the world. In the rst Critique, however, Kant makes an analogy between reason as the
unquestionable free cause and space as the outer form of human intuition. (Kant makes a similar
analogy between moral law and the forms of intuition in a note dated during the 1780s [R7201,
19:275].) In both cases, Kant says, we can give no reason why they and no others ll the role
they do. This analogy is much weaker than the others and misleads the reader. Space is not in
any sense a condition in a series whereas reason as free cause is a condition in a series of causes
and effects. One is not driven by the nature of reason to ask why human beings have the forms
of intuition that they do have; but one is so driven to ask for further conditions for any cause in a
series. The fundamental point of the analogy, however, still holds, namely that just as there is no
explanation why we have space as a form of intuition instead of some other, to us inconceivable, form, there is no explanation why we have reason as opposed to some other, to us
inconceivable, faculty.
239
Concluding remarks
At the end of Groundwork III, Kants nal position on his achievement in the
deduction amounts to a warning not to take the conclusion to be more than a
practical idea that human beings can use to help to justify their moral lives,
provided that they do not venture too far in their quest for that justication.
Even in the place where Kant is supposed to have provided his strongest
argument in favor of a theoretical proof of the reality of morality, he does
not conclude more than that human moral agents must conceive of themselves
as beings endowed with an objectively valid reason. Only as a self-conception,
only from within the empirical moral agents own practical point of view is this
idea even operative, and then without any ontological implications that would
conrm empirical reality to the objective rationality of other moral agents.
In very broad terms, we should expect no different from Kant than the
restriction of morality to the empirical moral agent and thus to an empirical
moral idealism. Since morality is a product of reason, we should expect that
morality would face the same boundaries of explanation that theoretical
knowledge faces given the limitations of reason. More important, we should
be prepared to accept with Kant that reason alone cannot provide knowledge of
real things, which is the same as admitting that we should not assume that the
universe is fundamentally rational. The main point of the Critique of Pure
Reason is that reason cannot divulge reality to us. We can be guaranteed to
know the structure only of the kind of world we can experience, that is, nature
in space and time. The categories of the understanding and the forms of
intuition can be known to be actual structures of our experience only because
they can be shown necessary for our kind of experience. They have no
conrmable validity beyond that limitation. Reason can add to the empirical
knowledge of the world only the systematic organization of concepts of the
understanding with the caveat that this systematic organization reects the
structure of reason and not the structure inherent in the objects of experience.
Reasons limitations are greater outside the boundary of experience. Reason
cannot provide knowledge of the nature of things in general nor prove one way
or the other whether there is a soul, a God, or a free cause. Were the universe
itself rational, were it structured by reason, then the faculty of reason would be
able to condently and correctly provide answers to these questions. But it
cannot.
Nor would the practical use of reason be able to provide this access to things
in general or even things in our experience. Unlike the concepts of the
understanding, the principles of reason do not even purport to refer to objects
in nature, and so are not amenable to any empirical conrmation or any
independent standard. If concepts of the understanding do not match nature,
nature will so to speak push back. Reasons moral principles govern actions
240
31
32
Formosa reaches this conclusion with regard to the realists talk of the moral law stemming
from rational nature and the constructivists talk of the constitution of self-legislating beings.
Just one paragraph earlier he had identied the realist discussion of the moral law as lying
in the nature of things ([Formosa 2013, 189] quoting [Kain 2004]). When realists talk of
rational nature, they tend to mean not simply a type of being that happens to exist in nature
with some particular way of structuring its own experience but instead a being who is able
fundamentally to discern what morality is in the universe because the faculty is not in any way
arbitrary.
Allen Wood admits that his interpretation of Kants deduction of the moral law shows that for
Kant there is no difference between the actuality of the moral law and our belief in that actuality
(Wood 1999, p. 381, n. 30). This admission seems inconsistent with Woods value realism.
242
Postscript
used in such a way, and I have used in this book the example of the
transcendental justication of causal relations in nature and the requirement
that any being who would represent that nature objectively must have a
cognitive system that processes sensations using causal relations. But there
are also signicant portions of Kants theoretical philosophy that involve
ontological claims absent in the practical philosophy, in particular the nature
of the spatiality and temporality of things. My claim is restricted to Kants
practical philosophy.
Nor am I claiming that I have presented precisely the historical Kants
position. Given the scope of Kants ambitions and the inconsistent use of
some of his terminology no interpretation can claim to comprehend all of
Kants stated claims and arguments. I offer an interpretation that encompasses
the main claims in Kants ethics and shows how Kant either did or could have,
using the resources available within his practical philosophy, presented an
entirely metaphysically naturalistic ethics. Relatedly, I do not make comprehensive claims about his philosophical development but only review the
development of some issues to draw out some material that I think has bearing
on interpreting his core critical works.
I do offer a plausible and defensible interpretation of Kants ethics along
metaphysically naturalistic lines. Given Kants distinction between physics
and psychology as two disciplines that study matter and thinking nature
respectively, his denial that the latter can rise to the level of a science, and
his rejection of a reduction of the psychological to the material, it would appear
that empirical psychology would not fall under metaphysical naturalism. But
given Kants equally strong insistence that empirical psychology is a possible
discipline short of proper science, his allowance of inner intuition as a source
of content governed by time as a form of intuition as well as the pure concepts
of the understanding, his strict empirical determinism regarding decisions
considered as appearance or phenomena, I allowed that the subject matter of
empirical psychology should be included in Kants conception of nature for
these ontological purposes despite lack of epistemic access.
With regard to entities, the elements of Kants moral theory that appear to
require an appeal to things in themselves independent of nature in space and
time are the transcendentally free faculty of pure practical reason itself, the
existence of God and immortality of the soul as postulates, and the transcendentally free faculty of choice. The elements that appear to require nonnatural
properties are the values of ends, both chosen by particular moral agents and
necessary as ends in themselves.
The transcendentally free faculty of pure practical reason (G) is not itself an
object per se but as a faculty it would have to reside in a being, and commonly
this being is understood as a person in herself independent of nature. I argue
that pure practical reason can t within nature because it is understood as the
243
244
Postscript
With regard to moral properties, the value of humanity as an end in itself (C)
would appear to be incompatible with Kants ontology of nature. Contingent
values of particular ends (A) depend upon empirical moral agents desires and
are thus clearly within nature because they are dependent on the empirical
subject. The value of humanity, however, is often assumed to be a necessary
property of the objects rational beings independent of the empirical subject.
Kants ontology of nature has no place for such intrinsic moral properties.
I showed that the value of humanity is dependent upon the transcendental
moral subject (a being with transcendentally justied practical reason) in a way
parallel to the dependence of contingent values on empirical subjects. To have
absolute value is only to have the highest rank in practical reasons ordering of
ends. Practical reason issues this ranking through the categorical imperative, so
the value is not inherent in objects as any intrinsic property but is only a formal
framework for experience provided by reason.
In these ways I have shown that Kants ethics is compatible with a metaphysical naturalism. I take this result to be important for two reasons. In
contemporary philosophy Kantians may defend many of the claims they derive
from Kant without worry that those claims are incompatible with nature as
now conceived. The second and more important reason is that my interpretation shows how Kants concern was to defend the moral life of human beings
as natural beings. Traditionally Kant has been seen as insisting upon supernatural abilities or entities in ethics free agents in themselves independent of
space and time relying on God and their own immortal soul to complete their
full moral lives. Even more generally, Kant is taken to deny knowledge of
supersensibles in theoretical philosophy in order to allow practical philosophy
to ll this conceptual space with the supersensible entities it requires. My interpretation reduces the claims about supersensibles greatly because I show that
the most important faculty for morality, practical reason, is comprehensible
entirely within nature, and that the other claims to supersensibles, as postulates, are not taken to have supersensible reference at all. This points to a
general interpretation of Kant as trying to comprehend the place of human
beings in the universe. And it conforms to an emphasis on the human point
of view and the limitations of reason, a topic that also grounds my claim that
Kant was a moral idealist.
Kant as moral idealist
The question whether Kant is a moral realist is more nuanced than the question
about naturalism. Realism operates on two levels: empirical, having to do with
the existence of actual moral agents, and transcendental, a consideration of the
requirements for any possible moral agent. In my rst chapter, I dened moral
realism and moral idealism in this way:
245
246
Postscript
247
successful deduction of the validity of morality. This by itself could still allow
for a view that, although we cannot prove that morality is transcendentally
valid, we must still believe that it is, and thus believe that morality is empirically real. The second point is that one can question whether Kant understands
the practical point of view to be making any claims about empirical reality in
light of its relation simply to the determination of action. On one side of these
two pivot points is the interpretation of Kant as a transcendental idealist and
empirical realist, on the other side is the interpretation of Kant as only an
empirical idealist. An analysis of these points will clarify the extent to which
Kant can be considered a realist at all.
At the rst pivot point is Kants denial of any possible transcendental
deduction of the validity of the moral law, based upon the claim that there is
no set of objects against which the moral law can be conrmed. The theoretical
transcendental deduction ranges over intuitions and so the categories can have
empirical reality as the structural framework for objects. Kant holds that there
are no corresponding independent objects to provide this conrmation for the
moral law and thus a transcendental justication for the moral law is unavailable. I argued that Kant had available to him appropriate independent objects,
namely, free acts of reason themselves. Actual empirical decisions of the
power of choice would have to implement the systematization provided by
the moral law in order for them maximally to embody freedom understood as
rationally determined decisions; systematicity among free actions would be
impossible without guidance from the moral law. I also argued that such a
view could ground an evolutionary account of the moral law in which any
possible organism that would evolve in order to make agent-perspective,
deliberate decisions would have to embody the faculty of reason to provide
this moral law. In this way Kant could have provided a transcendentally ideal
but empirically real pure practical reason. Since he claimed that he could not
provide a successful transcendental deduction of the moral law, he would have
to be interpreted as only an empirical idealist regarding reason.
Another avenue is open to the realist to preserve the transcendental ideality
of reason. The denial of a successful transcendental deduction provides only
the epistemic conclusion that empirical moral agents could not know whether
their faculty of reason was justied, but moral agents would still be free to
believe that reason is justied. This would retain the transcendental ideality
and empirical reality of practical reason as the ontology of morality even if it is
subject to some epistemic doubt. Analogously, an agnostic would admit that
Gods existence cannot be proved but would not deny that the subject of belief
would be a real God. If successful, this ontological claim would retain the
transcendental idealism and empirical realism of reason.
As an interpretation of Kant this move is questionable because of the second
pivot point: the nature of belief and the practical point of view. In a broader
248
Postscript
249
Works cited
Abela, Paul. 2002a. Kants Empirical Realism. Oxford, New York: Clarendon Press;
Oxford University Press.
2002b. Kantian Walls and Bridges: Challenging the Integrationist Model of the
Relation of Theoretical and Practical Reason. British Journal for the History of
Philosophy 10(4): 591615.
Adickes, Erich. 1927. Kant Und Die Als-Ob-Philosophie. Stuttgart: Fr. Frommans
Verlag (H. Kurtz).
Allison, Henry E. 1986. Morality and Freedom: Kants Reciprocity Thesis.
The Philosophical Review 95(3): 393425.
1990. Kants Theory of Freedom. New York: Cambridge University Press.
1996. Idealism and Freedom: Essays on Kants Theoretical and Practical
Philosophy. Cambridge, New York: Cambridge University Press.
1997. We Can Act Only Under the Idea of Freedom. Proceedings and Addresses
of the American Philosophical Association 71(2): 3950.
2004. Kants Transcendental Idealism: An Interpretation and Defense. Revised and
Enlarged Edition. New Haven: Yale University Press.
2007. Comments on Guyer. Inquiry: An Interdisciplinary Journal of Philosophy
50(5): 48088.
2011. Kants Groundwork for the Metaphysics of Morals: A Commentary.
New York: Oxford University Press.
2012. Essays on Kant. 1st ed. Oxford: Oxford University Press.
De Almeida, Guido Antonio. 2012. Critique, Deduction, and Fact of Reason. In Kant
in Brazil, North American Kant Society Studies in Philosophy, edited by Frederick
Rauscher and Daniel Omar Perez. Rochester: University of Rochester Press,
pp. 12754.
Ameriks, Karl. 1982. Kants Theory of Mind: An Analysis of the Paralogisms of Pure
Reason. Oxford, New York: Clarendon Press; Oxford University Press.
2003. Interpreting Kants Critiques. Oxford, New York: Clarendon Press; Oxford
University Press.
Auxter, Thomas. 1982. Kants Moral Teleology. Macon, Ga.: Mercer University Press.
Bagnoli, Carla. 2011. Constructivism in Metaethics. In The Stanford Encyclopedia
of Philosophy, edited by Edward N. Zalta. http://plato.stanford.edu/archives/
win2011/entries/constructivism-metaethics/ (April 1, 2013).
2012. Kants Contribution to Moral Epistemology. Paradigmi: Rivista di critica
losoca, pp. 6979.
2014. Starting Points: Kantian Constructivism Reassessed. Ratio Juris 27: 31129.
250
Works cited
251
252
Works cited
Works cited
253
1994. The Concept of Moral Insight and Kants Doctrine of the Fact of Reason.
In The Unity of Reason: Essays on Kants Philosophy, edited by Richard Velkley.
Cambridge, MA: Harvard University Press, pp. 5587.
Hill, Thomas E. 1989. Kantian Constructivism in Ethics. Ethics 99(4): 75270.
2012. Virtue, Rules, and Justice: Kantian Aspirations. Oxford: Oxford University
Press.
Hffe, Otfried. 2010. Kants Critique of Pure Reason: The Foundation of Modern
Philosophy. Dordrecht: Springer.
Hudson, Hud. 1994. Kants Compatibilism. Ithaca: Cornell University Press.
Hussain, Nadeem, and Nishi Shah. 2006. Misunderstanding Metaethics: Korsgaards
Rejection of Realism. Oxford Studies in Metaethics 1(2006): 26594.
Hutter, Axel. 2003. Das Interesse Der Vernunft: Kants Ursprngliche Einsichtund
Ihre Entfaltung in Den Transzendentalphilosophischen Hauptwerken. Hamburg:
F. Meiner.
Johnson, Robert N. 2007. Value and Autonomy in Kantian Ethics. In Oxford Studies
in Metaethics, edited by Russ Shafer-Landau. New York: Oxford University Press,
pp. 13348.
Kain, Patrick. 2004. Self-legislation in Kants Moral Philosophy. Archiv fr
Geschichte der Philosophie 86(3): 257306.
2005. Interpreting Kants Theory of Divine Commands. Kantian Review 9(1):
12849.
2006. Realism and Anti-Realism in Kants Second Critique. Philosophy Compass
1(5): 44965.
Kant, Immanuel. 2009. Fundamenao Da Metafsica Dos Costumes. translated and
edited by Guido Antonio de Almeida. So Paulo: Discurso Editorial.
Kaufman, Alexander. 2012. Rawls and Kantian Constructivism. Kantian Review
17(02): 22756.
Kielkopf, Charles F. 1997. A Kantian Condemnation of Atheistic Despair:
A Declaration of Dependence. New York: Peter Lang.
Kinnaman, Ted. 2008. Kants Nonfoundationalist Grounding of Reason. In Recht
und Frieden in der Philosophie Kants: Akten des X Internationalen KantKongresses. Band 5, edited by Valerio Rohden, Ricardo Terra, Guido de Almeida,
and Margit Rufng. Berlin: Walter deGruyter, pp. 34352.
Kitcher, Philip. 1993. The Evolution of Human Altruism. The Journal of Philosophy
90: 497516.
Kleingeld, Pauline. 1998. The Conative Character of Reason in Kants Philosophy.
Journal of the History of Philosophy 36(1): 7797.
2010. Moral Consciousness and the Fact of Reason. In Kants Critique of
Practical Reason: A Critical Guide, edited by Andrews Reath and Jens
Timmermann. New York: Cambridge University Press, pp. 5572.
Korsgaard, Christine. 1989. Personal Identity and the Unity of Agency: A Kantian
Response to Part. Philosophy & Public Affairs 18(2): 10132.
1996a. Creating the Kingdom of Ends. New York: Cambridge University Press.
1996b. The Sources of Normativity. New York: Cambridge University Press.
2003. Realism and Constructivism in Twentieth-century Moral Philosophy.
Journal of Philosophical Research 28: 99122.
2008. The Constitution of Agency: Essays on Practical Reason and Moral
Psychology. New York: Oxford University Press.
254
Works cited
Works cited
255
256
Works cited
Works cited
257
2010. The Primacy of the Practical and the Idea of a Practical Postulate. In Kants
Critique of Practical Reason: A Critical Guide, edited by Andrews Reath and
Jens Timmermann. New York: Cambridge University Press, pp. 16896.
Wolff, Robert Paul. 1974. The Autonomy of Reason: A Commentary on Kants
Groundwork of the Metaphysic of Morals. New York: Harper & Row.
Wood, Allen W. 1970. Kants Moral Religion. Ithaca: Cornell University Press.
1984. Kants Compatibilism. In Self and Nature in Kants Philosophy, edited by
Allen W. Wood. Ithaca: Cornell University Press, pp. 73101.
1999. Kants Ethical Thought. New York: Cambridge University Press.
2007. Comments on Guyer. Inquiry: An Interdisciplinary Journal of Philosophy
50: 46579.
2008. Kantian Ethics. New York: Cambridge University Press.
Wood, Allen W., and Dieter Schnecker. 2004. Immanuel Kants Grundlegung
Zur Metaphysik Der Sitten: Ein Einfhrender Kommentar. 2nd ed. Paderborn:
Schningh.
Zammito, John. 2008. Kant and Naturalism Reconsidered. Inquiry 51: 53258.
Index
as if interpretation, 155
Abela, Paul, 34, 83
Adickes, Erich, 170
Allison, Henry, 21, 34, 989, 102, 108, 112,
11921, 133, 157, 183, 187, 195, 201,
211, 22830, 2356
Ameriks, Karl, 37, 56, 104, 122
amoralism, 43
anomalous monism, 134
anthropology, 17, 20, 35, 38, 578, 76, 96,
212
architectonic, 512, 56, 58
atheism, 149
autonomy, 2335
and concept of God, 176
and structure of reason, 125
and value, 21017
not arbitrary, 2212
denition of, 210
Auxter, Thomas, 226
Bagnoli, Carla, 256
Baumgarten, Alexander, 73
Beck, Jacob Sigismund, 18990
Beck, Lewis White, 87, 989, 108, 149
Bell, David, 64
biology, 35
Bird, Graham, 28, 301, 33, 64
Bittner, Rdiger, 25
Bojanowski, Jochen, 12, 16
Boyd, Richard, 14
Byrne, Peter, 151
categorical imperative, 445, See also moral
law
and empirical reason, 119
and evolution, 1434
and knowledge of value, 220
and particular duties, 37
and postulate of freedom, 206
and priority of action, 8891
and rational agency, 26
258
Index
and value, 41
as decision procedure, 245, 39
as nature of agency, 267
compatible with moral realism, 13, 23
in mathematics, 23, 51
in political philosophy, 42
in theoretical philosophy, 23
origin of term, 1011
seen as only alternative to realism, 11, 37
critique. See also method
as division in philosophy, 56, 5968
as project of theoretical reason, 914
Cummiskey, David, 213
DallAgnol, Darlei, 14, 32, 37, 98, 210
Davidovich, Adina, 163
Davidson, Donald, 134
de Almeida, Guido, 98, 235
De Caro, Mario, 29
Dean, Richard, 210
deduction
empirical, 603
metaphysical, 132
of moral law, 247
of the moral law, 21, 65, 96, 1302, 143,
145, 22731, 2345, 238
transcendental, 20, 45, 603, 132
Denis, Lara, 149
determinism, 356, 11420, 18793
and practical freedom, 111
dignity, 223
divine command theory, 16, 177
domain
of ethics as free acts, 736, 205
of philosophy as ultimate end of human
reason, 535, 72
required for a system, 52
separate for theoretical and practical, 701
duty
no direct duties to God, 172
no duty to believe in God, 87
particular duties, 379
pure and empirical, 58
Dworkin, Ronald, 10
Dye, James Wayne, 31
empirical idealism
moral, 202, 240, 2468
and contingent ends, 37, 245
and duties, 379
and fact of reason, 1034
and God, 1779, 245
and moral law, 45
and necessary ends, 3942, 246
and practical reason, 2478
259
and reason, 457, 1456, 247
and value, 224
theoretical, 1920
empirical realism. See also metaphysical
naturalism, nature
moral, 202, 2468
and duties, 379
and free choice, 201
and method in ethics, 648
and moral law, 45
and necessary ends, 3942
and reason, 457, 71, 1435
and value, 21722, 227, 246
theoretical, 1920, 30, 33, 64
and category of causality, 64
ends
end in itself ultimately inexplicable, 2367
objective or necessary, 3942, 209, See also
humanity, value of; value, absolute
of human reason as domain of philosophy,
535
order of. See reason, order of
subjective or contingent, 37, 209, 222
that are also duties, 3940
error theory, 1216, 27, 248
Esteves, Julio, 132
ethics
as discipline in philosophy, 50
transcendental method in, 648, 1302
whether systematizable, 523
evolution
and moral law, 1434
and non humanlike altruism, 62
and transcendental method, 623
of categorical imperative, 247
experience
and fact of reason, 1012
moral, 202, 648, 85, 92, 96, 98100,
1302, 170, 206, See also practical
point of view
expressivism, 1315, 267
fact of reason, 32, 445, 96106, 206
and awareness of value, 219
and freedom, 1046
and metaphysical naturalism, 104
and moral agent, 1034
and realism and idealism, 1034
as act or deed, 1001
as consciousness of moral law, 667
denition of, 100
limits empirical realism, 144
Falkenburg, Brigitte, 34
Formosa, Paul, 12, 14, 25, 37, 240
Frster, Eckart, 59, 1756
260
Index
Hffe, Otfried, 34
Hudson, Hud, 187
humanity. See also value, absolute
as end in itself, 3942
value of, 402, 20917
and arbitrariness, 2212
assigned by reason, 223
based on natural property, 2202
not an intrinsic property, 21718, 2202
Hume, David, 31
Hussain, Nadeem J. Z., 26
Hutter, Axel, 55, 58
hypothesis, transcendental, 149, 158, 171
compared to belief, 15960
compared to idea, 1536
compared to postulate, 162
not endorsed in use, 1556
idea, 42, 119, 230
and boundary between practical and
theoretical, 231
and domain of philosophy, 535
and transcendental realism, 142
as heuristic device, 154, 219, 238
compared to hypothesis, 1536
of God. See God, concept
of hell not coherent, 171
required for architectonic unity, 512
idealism
empirical. See empirical idealism
moral. See moral idealism
transcendental. See transcendental idealism
immortality, 423, 70, 83, 149, 203
independence
as metaphysical separation, 15, 125
as unchangeable by another, 1267, See also
reason, structure of
intelligible world, 1368, 167, 2247, 229,
2335
as moral ideal, 2245
realized in nature, 2257
intuition, 23, 345, 117, 218
forms of. See also space, time
Johnson, Robert, 21113, 217
judgments, 11920, 1336, See also reason,
freedom of
Kain, Patrick, 25, 37, 39, 43, 102, 13941, 170,
177, 240
Kant, Immanuel
as reective rather than active person, 94
religious beliefs of, 149
Kielkopf, Charles, 101
kingdom of ends, 26, 38, 73, 245
Index
Kinnaman, Ted, 56
Kitcher, Philip, 12
Kleingeld, Pauline, 55, 978, 1001
Korsgaard, Christine, 11, 21, 234, 26, 41, 143,
210, 237
Krasnoff, Larry, 10
Kuehn, Manfred, 149
Larmore, Charles, 38
logic
as division in philosophy, 56, 58
Loparic, Zeljko, 29
Louden, Robert, 38, 169
Macarthur, David, 29
Manchester, Paula, 52
mathematics, 23, 31, 35, 51, 58, 66, 128, 161,
223
Meerbote, Ralph, 134, 187
metaphysical naturalism, 27, 2933, 36, 64,
2414
and duties, 39, 243
and empirical power of choice, 1856
and empirical reason as cause, 11420
and fact of reason, 104, 243
and free choice, 243
and freedom of choice, 457, 201, 203
and God, 14950, 177, 243
and moral law, 243
and necessary ends, 402
and obligation, 43, 243
and priority of action, 96
and reason, 2423
and structure of reason, 1434
and transcendental freedom, 125
and transcendental method, 603
and transcendental method in ethics, 648
and value, 2224, 244
denition of, 29
in theoretical philosophy, 33, 241
limited to ethics, 323
metaphysics
as division of philosophy, 56
method
and logic, 56
and systematic organization of cognition,
523
in ethics
compared to chemistry, 24, 656
transcendental, 9, 21, 32, 5968, 76
transcendental, 29, 107, 174
Mieth, Corinna, 128, 1323
mind, 27
and evolution, 623
nature of, 17
261
moral agent
and belief contrasted with opinion and
knowledge, 15660
and belief in postulates, 85
and causality of reason, 1203
and fact of reason, 1013
and rst-person freedom, 756
and God, 1779
and objectivity, 19, 648
and priority of action over belief, 8891,
936
and reason, 457, 756
and truth, 13
as being in itself, 28, 33, 67, 136, 198203
as subject rather than object, 1516, 47
empirical, 1618, 202, 39, 648, 968,
1034, 11420, 12932, 15660, 179,
1846, 189, 201, 222, 239
in constructivism, 247
in denition of moral realism, 14
obligation tied to, 434
possibility of non humanlike, 18, 145
praise and blame, 1938
purely rational, 98100, 129
transcendental, 1618, 202, 648, 968,
1034, 1302, 179, 211, 2224, 245
moral anti-realism, 11, 211
problems with term, 1112
moral idealism, 68, 1456, 2448, See also
empirical idealism, transcendental
idealism
and freedom of choice, 47
and obligation, 43
and postulates, 423
and priority of action, 94
denition of, 14, 244
separates reason and reality, 240
use of term, 12, 27
moral law, 445, See also categorical
imperative
consciousness of, 667
different from categorical imperative,
98100
possibility of deduction of, 1302
relation to concept of God, 15770
reveals freedom, 92, 11213, 1359, 188,
194, 205
stems from structure of reason, 125
ultimately inexplicable, 2356
why a priori, 95
moral nonrealism, 11, 13, 16, 27, 38
denition of, 14
moral principle, property, or object
contrasted with one required for morality,
15
262
Index
empirical, 54, 57
practical, 44, 568, See also reason, practical
and action, 72, 946
as distinct set objects, 701
as pragmatic use of knowledge, 74
contrasted with theoretical, 6873
domain as free acts, 6876
domain not empirical, 75
inner boundary of, 22934
outermost boundary of, 2289, 2348
structure of, 558, 91
theoretical, 44, 568, 228, See also reason,
theoretical
and knowledge, 72
as distinct set of objects, 701
contrasted with practical, 6873
transcendental, 5960, 174
physics, 30, 345, 61, 75, 167
Pierce, Drew, 169
Pihlstrm, Sami, 60
Plato, 29, 201, 224
Pollok, Konstantin, 31, 34
postulate, 83, See also freedom, God,
immortality
and idealism, 1779, 245
and primacy of practical reason as belief,
848, 93
as positing existence to theoretical concepts,
1623
constructed by practical reason, 1667
denition of, 152, 1612
immanent not transcendent, 15770, 204
theoretical use of term, 1523, 162
practical point of view, 25, 28, 756, 8891,
946, 1013, 1336, 1456, 239,
2478
primacy of practical reason, 8194, 162
and integration model, 83
dened as priority of action, 82
dened as priority of belief, 81
priority of action, 82, 8897, 102, 105,
2056, See also reason, practical
priority of belief, 8190, 97, 103, 105,
149, 206, 248, See also reason,
practical
Proops, Ian, 98
psychology
empirical, 31, 33, 357, 44, 62, 74, 956,
104, 11618, 1846, 242
moral, 26, 111
rational, 122
questions of reason, 578, 157
philosophy
denition of, 51, 534, 723
Railton, Peter, 11
Rawls, John, 10, 21, 24, 98
naturalism, 2736
and moral law, 44
and postulates, 423
contemporary, 2930
epistemological, 29
Kant's use of term, 27
metaphysical. See metaphysical naturalism
denition of, 241
methodological, 2831, 956
denition of, 29, 241
relation to realism, 9
semantic, 29
nature, 30, 336, 71, 84, 168
empirically real, 613, 11420
order of, 11819, 152, 18993, 2334
reason in, 11618, 1203
Index
realism
moral. See moral realism
reason, 36, 82
active nature of, 712, 100, 137
and system of cognitions, 512
as transcending nature, 122
blamed for wrong action, 1947
causality of, 67, 10930, 138, 1948, 224,
237, 243
and judgments, 1336
in nature, 10910, 11920
transcendentally free, 10910, 12030
ultimately inexplicable, 237
critical use of distinct from practical or
theoretical, 91
denition of, 109, 1203
empirical, 36, 107, 11420, 1846
freedom of, 47, 11423, 1306, 183,
203
human, 535, 92, 151, 164, 228, 237,
240
intelligible character of, 113, 116, 1223,
191, 1956, 237
interest of, 8194, 232
one faculty with different uses, 69
order of, 42, 11819, 152, 2227
practical, 43, 457, 157, See also
philosophy, practical
and action, 946
and concept of God, 1667, 1701, 1757
and empirical idealism, 1456
and is/ought distinction, 6970, 1456,
2324
and value, 218
creates it own objects, 712
meaning of Wille, 1834
primacy of. See primacy of practical
reason
pure, 28, 57, 116, 12230
regulative use of, 34, 36, 1534, 157, 168,
174, See also ideas, postulates
requires an unconditioned, 2358
speculative. See reason, theoretical
structure of, 108, 12230, 13946, 194, 196,
239, 242
theoretical, 946, See also philosophy,
theoretical
and boundary of practical, 2313
and concept of God, 1656
and critical works, 914
and is/ought distinction, 6970, 1456,
2324
and value, 218
higher priority of regarding belief, 848
must conform to practical belief, 814
263
objects given to, 712
same as speculative, 68
Reath, Andrews, 25, 389, 42, 211
Reinhold, Karl Leonard, 196
relativism, 43, 146
Ritchie, Jack, 29
Rohden, Valerio, 24
Rosenthal, Jacob, 128, 1323
Rotenstreich, Nathan, 85
Sayre-McCord, Geoffrey, 1214, 255
Schmidt, Claudia, 36
Schneewind, J. B., 38, 211
Schnecker, Dieter, 2289
science, 28, 302, 57, 61, See also biology,
chemistry, physics
social, 35, See also anthropology, history,
psychology
Sensen, Oliver, 211, 213, 21819, 223
Shah, Nishi, 26
Skorupski, John, 30
Smith, Joel, 29
space, 9, 17, 1920, 33, 35, 238
Studlin, Carl Friedrich, 57
Stern, Robert, 21, 37, 43, 143, 188, 210,
218
Stevenson, Leslie, 158
Strawson, P.F., 63
Street, Sharon, 1213, 16, 24
Stroud, Barry, 60, 634
Sullivan, Peter, 29
Sussman, David, 101
system
nature of systematic organization, 523,
124
structure of ethics, 523, 12930
teleology, 35, 54, 57, 167, 169
theology, 28, 114, 177
time, 9, 17, 1920, 33, 36, 113, 203
timelessness, 12230, 146, 181, 1867, 191,
198203
Timmermann, Jens, 132, 232
transcendental argument. See deduction,
transcendental
transcendental idealism
moral, 202, 457, 2468
and deduction, 247
and duties, 379
and fact of reason, 1034
and God, 1779
and method in ethics, 648
and necessary ends, 3942, 246
and reason, 1435
and value, 21117
264
Index