Autonomy of Reason and Practical Cognition
Thomas Buchheim
Kant’s philosophy is arguably the most decisive and powerful philosophical attempt ever to
assert freedom as the essence of man that distinguishes him exclusively. No other being,
whose existence we are objectively aware of, has according to this Kantian conception, if
remotely true, traits that are similar to freedom, and thanks to it man determines everything
that goes beyond mere nature.
However, Kant pays dearly for this determining power of human freedom insofar it is not
situated within a common context with nature but can only become real practically – in
human actions –, so that it cannot be cognized theoretically as objective reality or given
property of man. Moreover, it concerns practically only what is morally practical, whereas for
example artistic or creative freedom, but also freedom of action, which determines the
political and liberalist concept of human freedom remain out of consideration. For they are, if
we want to trust Kant, nothing more than meshes of natural processes that occur under
heteronomous laws that have at most indirectly to do with freedom. For that reason, I myself
do not think and cannot endorse Kant’s conception when it comes to freedom, even if I deeply
admire the almost indefeasible fortress that Kant built for the practical power of freedom in
his sense.
1. Two arguments against Kant’s conception of freedom
The almost indefeasible fortress of Kantian freedom has been subject to heavy attacks through
one of its indispensable contravallation, where it seems to be most vulnerable, not only by
Kant’s contemporaries, but also in most recent times, namely concerning the problem of how
freedom, morally determined, is related to freedom of immoral actions, whereby ‘immoral’
has to be understood as counter- or anti-moral, that is condemnable or evil, not as merely
amoral or morally neutral. Defining freedom morally implies the necessity of conceiving
everything that is immoral as being unfree or not free. This problem is well-known and
discussed in the context of the problem of “positive evil”, that is an evil that is not conceived
privatively as a mere occurring or relative defect of something good, but as a freely
committed, freely intended and only therefore imputable evil. Kant writes on this issue in his
famous note to the First Section of his Religion:
2
[1] “If the good = a, the opposite contradicting it is the not-good. Now, this not-good
is the consequence either of the mere lack of a ground of the good, = o, or of a
positive ground antagonistic to the good, = -a; in this latter case, the not-good can
also be called positive evil. (With respect to pleasure and pain there is a similar
middle term, whereby pleasure = a, pain = -a, and the State in which neither of the
two obtains is indifference, = o.) Now, if the moral law in us were not an incentive of
the power of choice, the morally good (the agreement of the power of choice with the
law) would be = a, and the not-good, = o; the latter, however, would be just the
consequence of the lack of a moral incentive, = a x o. In us, however, the law is
incentive, = a. Hence the lack of the agreement of the power of choice with it (= o) is
possible only as the consequence of a real and opposite determination of the power of
choice, i.e. of a resistance on its part, = -a; or again, it is only possible through an evil
power of choice. And so between an evil and a good disposition (the inner principle
of maxims) according to which the morality of an action must be judged, there is no
intermediate position“ (RBR AA 6, 22 f. Fn.)
Kant’s conception of freedom is therefore committed to be able to consistently think such a
positive evil as freely acted subversion of the morally good. For if immoral actions fell out of
freedom, and if hence our actions were under heteronomous laws of nature, then also the
reverse, namely morally good actions would not be imputable to the individual agent’s
realization of freedom but to his merely unimpeded autonomy of universal practical reason.
Karl Leonhard Reinhold and Kant’s early critics coined the term “intellectual fatalism” for
this condition of will, which cannot be assumed to be free. Their argument goes as follows:
Where the universal autonomy of reason is confronted with few heteronomous obstacles,
there the morally good realizes itself as a kind of fortune, whereas in all other cases the
heteronomous conditions remain prevalent that we call “immoral” or “evil” with regard to
human beings. Strictly speaking, however, these conditions are actually not “evil”, since
heteronomous conditions are not imputable to the free decision of the subject. What is
currently discussed under the label “moral luck” – a concept introduced by Bernard Williams
– resembles the intellectual fatalism in different ways.
In recent times, Reinhold’s striking argument against Kant’s conception of autonomy has
been introduced in the contemporary debate by Jörg Noller, who has explored the different
ways of how the post-Kantian philosophy, Kant included, attempted to escape intellectual
fatalism.
3
Already Fichte, in his System of Ethics from 1798, had famously attempted to escape this
dilemma by opposing an intellectual “laziness” or “inertia” to the morally reflecting
understanding of each subject. Those who overcome this laziness conscientiously to sufficient
extent act morally good out of individual freedom; those who follows his laziness all too fast
and without full moral consciousness act morally evil in the same free manner. However,
Michelle Kosch has recently pointed to the fact that this kind of evil would not be any more
an individual and positive opposition to the morally good (although Fichte calls this laziness
explicitly “positive” (System of Ethics, ed. Breazeale/Zöller, 189), and identifies it with
Kant’s radical evil), but rather understood in a privative way that, according to Kosh (and
here I agree with her), would not any longer merit the label of a “positive evil”: „[...] the
intellectual element which it was so important to preserve (to distinguish human evil from
mere animality) really does have to be present in limited quantity (or with limited
effectiveness) in order for evil to be possible.” And so it wasn’t “an alternative to the picture
on which evil is a deficiency, and particularly a deficiency in the exercise of the intellect”.
Michell Kosch has used this argument as an objection to Schelling’s attempt to establish a
positive evil, but Kant is, as mentioned earlier, equally committed to consider the moral
alternative to good as being positive or stemming from practical reason, if freedom itself is to
be defined morally. What kind of solution does Kant offer, in order to prevent Fichte’s
abandonment of individual evil and endorsement of a non sufficiently reflecting action as
privative immoral, or to escape intelligible fatalism, according to which the morally good
alone is grounded in the full autonomy of reason, whereas the positive evil remains utterly
erratic and incomprehensible?
In her above mentioned talk, Michel Kosch has interpreted these alternatives in terms of a
dilemma of two “claims”, one of which would obviously have to be denied, although both
should actually be preserved both from a Kantian and a morally rational point of view:
(a) the rationality of an action is a necessary condition for the freedom of this action.
(sofern eine böse Handlung frei ist, muss sie notwendigerweise vernünftig sein)
and
4
(b) the rationality of an action is a sufficient condition for the moral goodness of that
action. (vernünftig zu sein ist für sich hinreichend, um eine Handlung als moralisch gut zu
qualifizieren)
It seems that we need either to give up the connection between rationality and freedom of an
action or to deny the relationship between rationality and morally goodness. Or, as a third
option, we could argue that evil differs from good in terms of its lower degree of rationality.
However, in this case the moral definition of freedom would become problematic.
In what follows I want to examine whether and how Kant can escape this dilemma and both
objections of intelligible fatalism on the one hand, and the impossibility of a positive evil on
the other hand. The core of the solution that Kant claims in my eyes consists in the argument,
that practical rationality and theoretical understanding and all logical operations being
directed towards cognition are not of gradually but of vectorially distinghished relevance,
depending on whether it logically apprehends the right or wrong. To logically apprehend the
wrong is in its full sense a mode of understanding and rationality, even if one is not aware of
its being failed or wrong. Reason is already rational before it rationally conceives either what
is right or what is wrong. And for that reason also the logical and rational conception of
failure is necessarily a mode of reason. Rationality would not be what it is if it were not the
same in the case of failure and in the case of truth.
2. Good and evil as objects of practical reason
The section in which Kant discusses this internal difference of practical reason and on which
he builds his conclusions can be found in the second chapter of the Critique of practical
reason, a centerpiece of Kant’s moral philosophy, which remarkably is one of the
hermeneutically least understood parts of Kant’s conception of freedom.1 Whereas the First
Chapter deals with the “Principles of Pure Practical Reason” and constitutes the “autonomy”
of pure practical reason (= Theorem IV, AA 5, 33-39), the Second Chapter deals with “the
concept of an object of pure practical reason”. Objects are generally and especially in Kant
something that is cognizable under certain transcendental conditions (like for example
1
Recently two dissertations on this section (Stephan Zimmermann: Kants ‚Kategorien der Freiheit’,
Berlin/Boston 2011 and Heiko Puls: Funktionen der Freiheit. Die Kategorien der Freiheit in Kants ‚Kritik der
praktischen Vernunft’, Berlin/Boston 2013) and a new collected volume on the categories of freedom (Stephan
Zimmermann Hg.: Die ‚Kategorien der Freiheit’ in Kants praktischer Philosophie. Historisch-systematische
Beiträge, Berlin/Boston 2016) have substantially contributed to a better understanding.
5
intuition and functions or categories of the understanding.2 And in fact Kant speaks in this
Second Section of practical “cognition” (KpV AA 5, 66.8) of good and evil as objects of
practical reason:
[2] „by this it happens that, since all precepts of pure practical reason have to do only
with the determination of the will, not with the natural conditions (of practical ability)
for carrying out its purpose, the practical a priori concepts in relation to the supreme
principle of freedom at once become cognitions […] (KpV AA 5, 66.3-8)
However, it is something quite different to have certain principles under whose obedience one
operates rationally, and to cognize objects on which one directs oneself rationally and in
accordance with those principles. The latter is not contained in mere autonomy or selflegislation of reason per se. The object that provides the content of my will is not already
good insofar as I or practical reason gives itself a purely formal principle autonomously, and
otherwise: Practical reason is not the autonomous master of its principles insofar as the object
that provides the content of my volition is good. From the beginning of his Second Critique
Kant argues in the quite opposite direction: Will’s determination according the law of reason
must not be obtained from what characterizes an object of my volition. No possible object
must be the ground of the determination of the will but rather the universal form of the law is
meant to be the sufficient reason for it. If this is the case, and if both the goodness of the
object and the autonomy of the principle of practical reason are independent, then what
matters is to cognize whether and how something that I can will, that is an action, is in
accordance with this autonomous law of reason or not. According to Kant, this cognition is a
practical cognition which produces the objects of practical reason (good and evil), or more
precisely: the relevant object according to the concepts for good and evil – other than in the
case of speculative or theoretical reason, without being dependent on a preceding empirical
intuition of these objects at all:
[3] “the practical a priori concepts in relation to the supreme principle of freedom at
once become cognitions and do not have to wait for intuitions in order to receive
meaning; and this happens for the noteworthy reason that they themselves produce
the reality of that to which they refer (the disposition of the will), which is not the
business of theoretical concepts.” (KpV AA 5, 66.8-11)
2
Cf. especially Jochen Bojanowski: „Kant über praktischen Gegenstandsbezug“ in dem erwähnten Sammelband
(S. 107-128) and idem.: Kant’s Categories of Practical Reason as Categories of Practical Cognition, in: Kantian
Review 20 (2015), S. 211-234.
6
For practical cognition of its object (what, which action do I have to want in my situation
most likely?), reason needs no empirical intuition of such objects, since it rather produces the
object (that it wants) by means of a pertaining cognition. However reason needs, as Kant goes
on, certain categories (= practical concepts a priori), in whose application on empirically
underlaid representations it formulates the object so that is cognizable in its status in relation
to the highest principle of freedom. A cognition, be it practical cognition or theoretical
cognition like in the First Critique cannot exist without a formulation of the object by means
of categories. This performance of formulation is not exercised by the universal reason as the
provider of autonomous principles, but it is rather the matter of an individual subject of
cognition. Kant calls the relevant categories, that is root concepts of our practical reason, as
“categories of freedom with respect to the concepts of the good and evil”. The do not concern
pure practical reason (as does the principle of autonomy), but “practical reason in general”
(AA V, 66), that is reason within a subject that has also to consider empirical conditions and
that is directed towards possible actions and intentions within the empirical world. Kant
writes:
[4] „One quickly sees that in this table freedom is regarded as a kind of causality which, however, is not subject to empirical grounds of determination - with respect to
actions possible through it as appearances in the sensible world, and that
consequently it is referred to the categories of their natural possibility," while yet
each category is taken so universally that the determining ground of that causality can
be taken to be also outside the sensible world in freedom as the property of an
intelligible being, until the categories of modality introduce, but only problematically,
the transition from practical principles in general to those of morality, which can only
afterwards be presented dogmatically through the moral law.“ (KpV AA 5, 67)
Dealing with these categories and considering actions that are possible through freedom as
appearances in the empirical world, each subject’s practical reason cognizes the relevant
maxim or rule in relation to the supreme principle of freedom as being permitted or forbidden
or at least exceptionally allowed or forbidden or as duty of contrary to duty, maxims or
directions of actions that are either rules of volitions or of omission or of making an exception
of actions that are in turn either directed to the personality or merely to the condition of
person or mutually to a person with regard to the condition of another person. Reason always
begins with a formulation of such actions, informed by empirical representations, that is
sufficiently specific in order to be caused in a determined way, but also sufficiently general,
in order to allow a supersensible causality; for example to insult someone, which requires a
certain tone of voice or certain words that are to be produced, although in this case their cause
7
would be to imputed to my free causality, and although people insult others without being
aware of it (there exists something like a pathological, automatically triggered compulsion to
insult).
To cognize that such an insult is allowed for me here or even required, however, does not
mean in any case that this is also true and in a moral sense good and right. The question of
truth related to the cognition gained through the formulation is rather, as Kant puts it,
“problematically introduced” with my practical judgment. On closer consideration it could be
morally reprehensible having insulted the person, as if she had to accuse herself of moral
misconduct that she is not willing to see. Even I myself could realize it if I did not have
already accepted an evil maxim and affirmed and prolonged it with this deed. By cognizing
something as something, I cannot ensure that it is what I cognize, even if we speak of a
“cognition” in the definite sense of the word only if what we grasp is really true and the same
as how we conceive of it. All possible acquisition of cognition is only be possible by passing
through possibly error; cognition is problematic as long as it cannot be dogmatically
demonstrated out of valid laws and theorems.
In case of practical rational cognition, however, we do not speak of “wrong” or “mistaken”,
but of “misguided” or “damnable” or “evil”. This becomes obvious, at least according to
Kant, since even the criterion of truth and of what is right can be found in a person’s reason
(other than in case of theoretical cognition) so that there is a need of a certain “positive
antagonist” in our operations of cognition (that is the necessary performance of formulation)
in order to arrive at a faulty result, that is in practical terms: to produce an evil disposition of
the will.3
3
This antagonist that leads us to a faulty result in practical cognition is what Kant calls in his Metaphysics of
Morals an „incapacity“ (MdS AA 6, 227). However, it is not the case that we always need to understand the
word “incapacity” as a deficiency or a restriction of a capacity, and hence as something merely privative; for
example we speak of the “incapacity” to acknowledge the merits of someone else or the “incapacity” to show
gratitude, where the word is used for a inner refusal that is internal to the capacity. We need to understand the
passage in this sense. The positivity of evil is not suspended by the fact that it must not be used for a definition
of the essence of freedom according to Kant.
8
3. Objective and possible self-corruption in the process of practical cognition
According to Kant, moral cognition means “to subject” all our different intentions of actions
that we want to realize (including acts of omission) to “the manifold of desires to the unity of
consciousness of a practical reason commanding in the moral law, or of a pure will.” (KpV
AA 5, 65.25 f.).
[5] „since actions on the one side indeed belong under a law which is no law of nature
but a law of freedom, and consequently belong to the conduct of intelligible beings,
but on the other side as also events in the sensible world yet belong to appearances,
the determinations of a practical reason can take place only with reference to the
latter and therefore, indeed, conformably with the categories of the understanding, but
not with a view to a theoretical use of the understanding, in order to bring a priori the
manifold of (sensible) intuition under one consciousness, but only in order to subject
a priori the manifold of desires to the unity of consciousness of a practical reason
commanding in the moral law, or of a pure will.“ (KpV AA 5, 65)
This unity of consciousness is to be produced by ourselves by means of those categories, in
which we formulate the intended actions. And this unity can be formed in different ways,
depending on whether the moral law is given the supreme condition in a maxim or not the
supreme but one that is affected by another condition that is not by all means made explicit.
Not all conditions of the of the validity of a sentence need to be made explicit in order to
consider the sentence as valid. And precisely this is according to Kant what constitutes the
evil, namely to acknowledge the validity of a general sentence or law without explaining the
subjective reservations under which it stands despite the approved universality. According to
this picture I cognize an action as quite allowed that could be convicted as illegal and contrary
to duty according to a dogmatic account. For example the safety regulations of oil drilling on
deep sea are in best interest of the oil company that have the license and potential to produce
such oil; which is the reason for the fact that a relaxation of safety regulations must be
allowed in situations that heavily interfere with this interests, in order to sustain their
legitimacy and recognition in general. For otherwise the safety regulations would have to be
lowered, in order to even attract producing companies and hold them profitable. By the mere
formulation in relevant categories of what is allowed, bidden and forbidden, it is given the
rationally justified impression that particular exceptions of the safety regulations be allowed.
And this permission is accepted. It is not just apparent but becomes evident by the
formulation.
9
In this case, the universal rule is corrupted by its formulation with regard to concrete
empirical conditions of a an action that is possible through freedom by an interest, that is the
universal law does not hold the supreme condition of the maxims of action in this reasoning,
although it is acknowledged with regard to its validity. A positive antagonist against the
universal law as an incentive of action realizes itself by the mere formulation of maxims or a
framework of prescriptions and legal actions under universal empirical conditions in certain
categories.
In his paragraph on autonomy (§ 8) of the First Chapter, Kant described this quite clearly as
follows:
[6] “Thus the moral law expresses nothing other than the autonomy of pure practical
reason, that is, freedom, and this is itself the formal condition of all maxims, under which
alone they can accord with the supreme practical law. If, therefore, the matter of volition,
which can be nothing other than the object of a desire that is connected with the law, enters
into the practical law as a condition of its possibility, there results heteronomy of choice,
namely dependence upon the natural law of following some impulse or inclination, and the
will does not give itself the law but only the precept for rationally following pathological
law“ (KpV AA 5, 33)
The kind of formulation of actions under conditions of sensibility allows to put a certain
matter of volition as a rational condition of the possibility of the practical law itself.4 This
leads (due to corrupting, but not lesser reason) to heteronomy of choice (Willkür), namely out
of freedom with regard to the sensible conditions, under which one has to act; it leads to
heteronomy of choice, that is to a form of patronizing along the guidance of an interest or a
matter of volition. However, this introduction to heteronomy is therefore not to a lesser degree
free than a sincerely enduring formulation of the object of practical reason.
Hence we have shown that Kant does not succumb to the argument of intelligible fatalism,
that is that neither the conceivability of a free action that is contrary to the order of the moral
law must be abandoned, nor must he hold that evil actions can only be realized by a kind of
erratic leap into heteronomy and thus non-freedom. Rather, according to Kant, both follows in
the course of an operation of practical reason that is generally indispensable with regard to
practical cognition, whereby it is impossible to identify practical reason as autonomously
4
Fichte has overseen this in his conception of evil or immorality as ultimately being based on a laziness to moral
„reflection“: the formulation that has to be performed in order to arrive at the cognition of what is duty or
permitted and forbidden. In my opinion he explains everything else perfectly – only this slipped his attention.
10
commanding with it being practically cognizing. Pure practical reason cannot command other
than the moral law; but “practical reason in general” can (with regard to sensible
circumstances under which one has to act) cognize other as being feasible than what is
morally right. By the mere imperative, practical reason has not yet cognized anything as being
right. But freedom, under the conditions of sensibility, seizes the opportunity afforded by the
commandment, as Paul says (Rom 7,8), and therefore realizes either freely the morally good
or freely the morally evil.
Since practical reason as law-giving cannot be within the same mode of operation than pure
practical reason as cognizing, and since both morally good and wrong can be cognized as
rationally feasible, both propositions given by Michelle Kosch can be maintained without
contradiction – by a mere clarification of their meaning:
(a) The rationality of an action is a necessary condition for the freedom of this action.
and
(b) The rationality of an action is a sufficient condition of the moral goodness of this
action.
We must merely concede in Kant’s sense that practical reason necessarily operates on two
stages, namely first as autonomously law-giving reason and second as reason that cognizes
what is feasible. Since the latter is rational in two differently directed versions, we are entitled
to say the following:
(a*) The rationality of an action (in the sense of a its two stage form as reason being
law-giving and cognizing what is feasible under sensible conditions) is a necessary
condition for the freedom of this action.
and
(b*) The rationality of an action (in the sense of a its two stage form as reason being
law-giving and cognizing what is feasible under sensible conditions without selfcorruption) is a sufficient condition of the moral goodness of this action.
In contrast to a divine being that unconditionally follows the moral principles, man stands
under empirical conditions of his existence. For that reason he cannot avoid to observe
moral principles other where than in a sensible world and under sensible conditions.
11
For this purpose he needs to cognize what sensible actions satisfy the moral principles most
likely, under which his reason stands a priori. And for this cognition he needs to perform
formulations with regard to possible actions, as they confirm with the moral principles, that
are inaccessible to him. However such formulations are quite accessible by means of
categories of freedom, in order to give a good and evil disposition an expression that is
rationally compatible with the individual consciousness of the moral law. An evil disposition
arises from the positive antagonism of the umbrage that a sensible being takes to the law that
is imposed to it regardless of its sensibility.
Finally I want to return to the general topic of freedom after Kant. I think I have shown that
Kant himself did not hold that human freedom consists merely in the autonomy instead of
heteronomy of pure practical reason, but rather in its preservation in everyone’s practical
cognition of what is most likely to do under his or her circumstances, that is what is in
conformity with this autonomy of pure practical reason. Each cognition, however,
presupposes inevitably a conceptual or propositional formulation of that context on which the
intention of cognition is directed. It is only by virtue of such a formulation that what is
grasped by cognition can either conform with what is true or wrong, that is go wrong. Such a
failed conception of the conditions is based on a somehow distorting representation of them
by that performance of formulation that remains even lesser undetected the less the relevant
cognition is part of a formulated and verified theory or doctrine of even the entire context of
that cognition.
Whoever is not inclined to difficult or controversial areas of practical cognition can always,
by application of the categories of freedom, cognize what is right without distortion of the
conditions. He can also expose it to a more or less obvious distortion so that other appears to
be feasible than in the case of others in his place. But whoever is looking for the cognition of
what is right in difficult or controversial areas of the practical risks to be misled by delusions
that he himself does not realize although others may note them.
12
In any case, in my opinion the Kantian argument is of special importance, although it is often
neglected in today’s interpretations of his moral philosophy, that, despite the admitted
autonomy of our pure practical reason, what is morally right and wrong cannot, as a challenge
of distinction, be subject to negotiation of our freedom and the pertaining reasoning of
practical reason in general, but binds our freedom, in some circumstances without our
awareness. In my opinion, this element of Kant’s moral philosophy has not been taken
seriously enough in contemporary critical discussions. From its neglect follows that we are
quite evil although we consider ourselves to be in perfect accordance with the Kantian
autonomy of freedom.