<< . .

. 3
( : 33)

. . >>

The understanding as a capacity to judge
I use the expression ˜˜capacity to judge™™ to translate the German
Vermogen zu urteilen. Kant uses this expression when he introduces his
table of logical functions of judgment in the Transcendental Analytic of
the Critique of Pure Reason. There he justifies defining the understanding
as a capacity to judge in the following way. The understanding is a
capacity for concepts. But we form concepts only for use in judgments.
And all forms of judgment govern possible forms of syllogistic inference.
The understanding, then, or the intellect as a whole2 “ our capacity to
form concepts, to combine them in judgments, and to infer true judg-
ment from true judgment in syllogistic inferences “ is nothing other than
a ˜˜capacity to judge™™ (Vermogen zu urteilen) (A69/B94).3
I want to stress several important points here. First, this Vermogen zu
urteilen is different from the Urteilskraft, or power of judgment, that Kant
defines as the capacity to subsume particular instances under general
rules. Either we have the rule, and we look for instances of the rule (this
is the ˜˜determinative™™ use of the power of judgment, for which the
canonical example is of course the subsumption of given appearances
under the categories). Or we have particular objects and we look for the
rules under which they might fall (this is the ˜˜reflective™™ use of the power
of judgment, as described in the Introduction to the third Critique).4 But

˜˜Intellect as a whole™™ because it includes the capacity for concepts (i.e. the understanding in
the narrow sense), the capacity for subsuming objects under concepts, or power of judg-
ment (Urteilskraft), and the capacity for syllogistic inferences, or reason. These three aspects
of the exercise of the intellect, which correspond to the three main chapters in logic
textbooks of the time (1 “ concepts, 2 “ judgments or propositions, 3 “ inferences) are all
made possible by the fact that the intellect is a capacity to judge, a capacity to form
judgments according to the elementary forms laid out in Kant™s table.
In translating Vermogen zu urteilen as capacity to judge, I differ from Kemp Smith (Kant™s
Critique of Pure Reason, New York: St. Martin™s Press, 1965) and Guyer and Wood (Kant™s
Critique of Pure Reason, Cambridge: Cambridge University Press, 1998), who translate it as
˜˜faculty of judgment.™™ I prefer ˜˜capacity to judge™™ because it avoids the dubious faculty-
psychology and stresses instead mental capacities to act in determinate ways (in ordering
See AAv, p. 179; AAxx, p. 211.

defining the intellect, in all its guises (concept formation, subsumption of
instances under concepts or rules, syllogistic inference) as a capacity to
judge is explaining what it is about the understanding that makes it
capable of all the functions described above, including forming rules in
the first place. According to Kant, all of these can be traced back to the
fact that the intellect is a capacity to combine concepts (universals) in the
elementary ways (according to the elementary forms) described in
Kant™s table of logical functions, or forms, of judgment.5
Allison objects to my privileging in this way Kant™s description of the
understanding as a capacity to judge. Kant, he says, defines the under-
standing in many other ways as well: as a faculty of concepts, as a faculty
of rules, as spontaneity, as apperception. I agree. I also agree that the
characterization of the understanding as a Vermogen zu urteilen belongs
specifically to the context of the metaphysical deduction of the cate-
gories. But this does not make it any less important. For what it provides
is a definition of the original capacity from which all aspects of the
understanding are developed. Indeed from the argument I just
recounted it follows that concepts and rules are generated by the under-
standing as a capacity to judge. The understanding as spontaneity,
namely as the activity of producing rule-governed, reason-giving com-
binations of representations, is an activity of the Vermogen zu urteilen. And
in the Transcendental Deduction “ more clearly in B than in A “ Kant
argues that the identity and unity of self-consciousness (¼ apperception)
is the identity and unity of an act of judging, according to the forms Kant
has expounded in his table of logical functions of judgment.6 So

As I understand it, if there is a distinction to be made between function and form of
judgment in Kant™s usage of the terms, it should be a distinction between a rule-governed
act of combining representations (the function of judgment, or judging) and its result (the
form of judgment, namely the ways in which concepts are ordered in a judgment “ a
proposition). At A70/B95, Kant writes: ˜˜If we abstract from all content of a judgment and
consider the mere form of the understanding [Verstandesform] in it, we find that the function
of thought in the judgment can be brought under four titles, each of which contains three
moments under it.™™ Cf. A68/B93: ˜˜I call function the unity of the act of ordering distinct
representations under a common representation.™™ On this point, see KCJ, p. 78. Note also
that the point I am making in emphasizing that for Kant, understanding as a whole is a
capacity to judge, is broader than the point I made in the introduction to KCJ (pp. 7“8)
according to which the Urteilskraft could be understood as the actualization of the Vermogen
zu urteilen as a capacity, or an as yet unactualized potentiality to form judgments. The point
I am stressing now is that all aspects of the understanding (the early modern™s intellectus) as a
capacity, namely the capacity to form concepts, the capacity to subsume objects under
concepts, the capacity to form syllogistic inferences, are imbedded in this original char-
acterization of the understanding as a capacity to judge.
On this point, see KCJ, pp. 64“72.

although I would certainly not claim that characterizing the understand-
ing as a Vermogen zu urteilen is sufficient to account for all aspects of the
understanding as expounded in the Critique of Pure Reason, let alone the
second and third Critiques, I am claiming that all aspects of the under-
standing, in order to be properly understood, need to be traced back to
this original capacity to form judgments.
Now, reducing the intellect to a capacity to judge (specified according to
the elementary forms described in the table) is an extraordinarily import-
ant move to make. It is Kant™s response to the classical question: are there
innate representations? For Kant, there are no innate representations, but
there are innate capacities “ intellectual/discursive capacities of concept
forming and ordering, sensible/intuitive capacities of distinguishing and
ordering individuals. The cooperation of these two capacities in acts of
judging is, according to Kant, what makes us capable of recognizing the
numerical identity of individual objects through time as well as of recog-
nizing empirical objects under concepts of natural kinds. Both capacities
rest on the fact that the cooperation of the understanding, as a capacity to
judge, and sensibility, as a receptivity characterized by specific forms or
modes of ordering, generates categories according to which we can repre-
sent the numerical identity of objects and reflect them under concepts.
I will return in a moment to this issue of the ˜˜generation™™ of the categories.
In KCJ, I have analyzed in great detail Kant™s conception of logical
forms as expounded in his table of logical forms of judgment. My purpose
in doing this was to understand why he thought that just these forms of
discursive thought were minimally necessary for any recognition of objects
under concepts to occur. It is in this context that I have talked about an
˜˜objectifying function™™ of the logical forms of judgment. Allison agrees
with me on this point, and he also agrees about the caution one should
exercise in interpreting the point: it does not mean, of course, that for
Kant any judgment is true. What it does mean is that the logical form of a
judgment is what makes a judgment capable of truth or falsity, because it is
that by virtue of which the judgment expresses the relation of our repre-
sentations to independently existing objects. However, Allison also thinks
that, in my account, the forms of judgment end up ˜˜usurping the objecti-
fying function usually assigned to the categories.™™ But this is not so. What I
say “ in the very passage Allison quotes in support of his claim “ is that only
in the light of the objectifying function of the logical forms of judgment can
we also understand that of the categories themselves.7

See KCJ, p. 12, referenced in n. 2 of Henry Allison™s comments: see ˜˜Categories,™™ p. 79.

What does this mean, and what is the specific ˜˜objectifying™™ function of
the categories, as distinct from that of the logical forms of judgment?
Kant answers this question in the section of the Transcendental Analytic
that immediately follows the table of logical forms of judgment and
introduces the table of categories. The same function, he says, that
gives unity to concepts in judgment also gives unity to the mere synthesis
(or combination) of representations in intuition. The categories express
just those forms of unity of synthesis of representations in intuition
(A79/B105). So the logical forms of judgment are forms of the unity of
the combination of concepts in judgment. The categories ˜˜universally
represent™™ forms of the unity of the combination of representations in
intuition. What they add to the logical forms of judgment is thus the
unity of intuitions under the latter. But they are concepts of a synthesis of
intuition achieved by the very same function that unites concepts in
judgments: the function of the understanding, namely of the capacity
to judge, Vermogen zu urteilen.
The logical forms of judgment are forms of analysis, in the peculiar
sense Kant gives to this term, where analysis does not mean primarily
analysis of concepts (although it also means that), but analysis of a
sensible given in order to form concepts (cf. A76/B102). The categories,
on the other hand, express forms of synthesis of the sensible given.
There is, admittedly, something puzzling about the fact that forms of
synthesis are supposed to originate in forms of analysis. Allison
expresses just such puzzlement when he says: ˜˜I fail to see how forms
of analysis (the logical forms of judgment) might be equated with forms
of synthesis (the categories).™™8 But actually this tells only part of the story.
The whole story is this: it is insofar as they are themselves forms of
synthesis (forms of synthesis or combination of concepts) that forms of
judgment are also forms of analysis (analysis of the sensible given with a
view to forming concepts of objects to be combined “ synthesized “ in
judgments). This is why Kant writes in the section of the Metaphysical
Deduction cited above:
The same understanding, therefore, and indeed by means of the very
same actions through which in concepts, by means of the analytical unity,
it brought about the logical forms of a judgment, also brings, by means of
the synthetic unity of the manifold in intuition in general, a transcen-
dental content into its representations, on account of which they are

Allison, ˜˜Categories,™™ p. 72.

called pure concepts of the understanding that pertain to objects a priori,
a point that could not be derived from general logic. (A79/B105)

˜˜By means of analytic unity™™ means: by means of a unity reached by way
of analysis. Judgment is a synthesis (of concepts) by means of analysis (of
the sensible given). Categories are concepts of the synthesis of intuition
necessary for the analysis of this same intuition that allows concepts of
objects to be formed and synthesized in judgments. So, if you like, the full
process is: synthesis (of intuition) for analysis (into concepts) for synthesis
(of these concepts in judgment). The categories universally represent the
unity of the original synthesis of intuition for analysis for synthesis (of
concepts). I think Sally Sedgwick may be missing this point when she
attributes to me the view that ˜˜the kind of unity necessary for combining
representations in judgment [Kant] calls ˜analytic unity™™™ or again when
she says that analytic unity is ˜˜the unity which combines concepts into the
various forms of judgment,™™ as opposed to the synthetic unity that ˜˜must
be produced in the sensible manifold before any such combination of
concepts can occur.™™9 Kant™s view, as I understand it, is that the combin-
ation of concepts is itself synthetic unity. It is synthetic unity (of concepts)
obtained by means of analytic unity (namely by means of the analytic unity
of consciousness that attaches to all common concepts: see B134n).
The difficulty Allison points out when he says he ˜˜fails to see™™ the
relation between analysis and synthesis as I tried to outline it is a very
important one and has weighed heavily on the reception of Kant™s
critical philosophy. To name only one example, this difficulty motivated
Hermann Cohen, the founder of the Marburg neo-Kantian school, to
dismiss Kant™s metaphysical deduction of the categories altogether and
instead to read the Critique of Pure Reason in backward order, from the
System of Principles, and even from the Metaphysical Foundations of
Natural Science, to the table of the categories, dismissing Kant™s argument
about logical forms and categories altogether. He could make no sense at
all of the argument about synthesis and analysis, in part because he
thought that when Kant talked about ˜˜analytic unity™™ he meant analytic
judgments. Then the whole argument of the metaphysical deduction
became, indeed, incomprehensible.10 One of the first to correct the

See Sedgwick, ˜˜Priority,™™ pp. 81“2.
See Hermann Cohen, Kants Theorie der Erfahrung, 3rd edn (Berlin: Bruno Cassirer, 1918),
pp. 242“5. I have given a more detailed account and criticism of Cohen™s view in the
French version of my book: see KPJ, pp. 92“5. In the English version, the reference to
Cohen™s mistake appears only in a footnote: see KCJ, p. 86, n. 10.

error was a Marburg Kantian, Klaus Reich, in his groundbreaking work,
Die Vollstandigkeit der Kantischen Urteilstafel. However, because Reich™s
effort to revive Kant™s argument in the metaphysical deduction was
flawed in very serious ways, it did not gain very much influence.11
Nowadays, as neo-Kantianism is attracting renewed interest in Kant
studies, I suggest that the least we can do is try to learn from its strengths
but not repeat the errors that cost us, to this day, an absolutely central
aspect of Kant™s whole array of critical arguments.
Now, the relationship I just outlined between synthesis and analysis
(for synthesis) should help me clarify what I mean when I say that the
categories, in Kant™s account, have a role to play at both ends of the
cognitive process.

The categories ˜˜at both ends™™: synthesis and subsumption
When, in the well-known letter to Herz of February 1772, Kant raises the
difficulty of understanding how it is possible for a priori concepts to be
applicable to objects that are given, he contrasts this difficulty with the
absence of any such problem where mathematical concepts are con-
cerned. In their case, he says, no such problem occurs for they ˜˜generate
the representation of their object as magnitude, by taking the unit several
times.™™ But how could the same be done when we were dealing not just
with magnitudes but with qualitatively determined, empirical things?12 In
KCJ, I have suggested that this contrast becomes in fact part of the clue to
the solution: some way has to be found to explain how the categories, just
like geometrical or arithmetical concepts, might be concepts under the
guidance of which the very representation of the objects thought under
them might be generated. This is precisely what is indicated by their
definition, in x14 of the Transcendental Deduction, as ˜˜concepts of an
object, by means of which the intuition of the object is considered as
determined with respect to a logical function of judgment™™ (B128).

Cf. Klaus Reich, Die Vollstandigkeit der Kantischen Urteilstafel (Berlin: Richard Schoetz,
1932); Engl. transl. J. Kneller and M. Losonsky, The Completeness of Kant™s Table of
Judgments (Stanford: Stanford University Press, 1992). Reich™s book has been subjected
to close scrutiny in recent studies of Kant™s table of judgments. See Reinhard Brandt, Die
Urteilstafel. Kritik der reinen Vernunft A67“76/B92“201 (Hamburg: Felix Meiner Verlag,
1991); Engl. trans. Eric Watkins, The Table of Judgments: Critique of Pure Reason A67“76/
B92“201 (North American Kant Studies in Philosophy, 4 [1995]). And Michael Wolff, Die
Vollstandigkeit der kantischen Urteilstafel. Mit einem Essay iiber Freges ˜˜Begriffsschrift™™
(Frankfurt-am-Main: Vittorio Klostermann, 1995).
See AAx, p. 131.

This characterization of the categories means two things. (1) To have a
category is to have a rule for ordering sensible manifolds (and for us
human beings, this means manifolds of spatiotemporal elements) in such
a way that they can be reflected under (empirical) concepts of objects
according to logical functions of judgment. For instance, to have the
category of substance is to have the rule: look for something that remains
permanent while its properties change. To have the category of cause is
to have the rule: look for something real that is such that whenever it
exists (˜˜is posited™™) something else follows. (2) To have a category is to
have a concept under which we can think an object as ˜˜in itself deter-
mined™™ with respect to a logical function of judgment.
Under the first description, categories guide synthesis. Under the
second description, objects are subsumed under them. These are the
˜˜two ends™™ of the cognitive process I mention in my book: first synthesis
(the categories are rules for synthesis); then subsumption (as any other
concept, categories are ˜˜universal and reflected representations™™ under
which objects are subsumed).
I have suggested that these two roles of the categories are apparent in
Kant™s explanation of the difference between judgments of perception and
judgments of experience, in the Prolegomena.13 Consider Kant™s example
of a judgment of perception that eventually becomes a judgment of
experience. ˜˜If the sun shines on the stone, then the stone grows warm™™
is a judgment of perception. ˜˜The sun warms the stone™™ is a judgment of
experience. How do we form judgments of perception, and how do we get
from judgments of perception to judgments of experience? Kant™s answer
is that first we perceive the repeated conjunction of light of the sun and
warmth of the stone. Then we form the hypothetical judgment: ˜˜If the sun
shines on the stone, then the stone grows warm.™™ And finally we come to
the conclusion that light of the sun and warmth of the stone are ˜˜in
themselves determined™™ with respect to the hypothetical form of judg-
ment: the connection exists not just ˜˜for me, in the present state of my
perception™™ but ˜˜for all, always.™™ It is not a ˜˜mere logical connection of
perceptions™™ but a connection in the objects themselves. We then subsume
the logical connection under the ˜˜concept of an object, by means of which
its intuition is determined with respect to the logical form of hypothetical

Prolegomena to any Future Metaphysics that will be able to come forward as science, ed. and trans.
Gary Hatfield (Cambridge: Cambridge University Press, 1997; rev. edn 2004). For the
distinction between judgments of perception and judgments of experience, see xx18“20,
AAiv, pp. 297“302.

judgment™™ (the concept of cause) and we say: the sun warms the stone.
This is the subsumption under the category. It occurs at the end of the
process that goes through the stages just described: perception of temporal
conjunction of events, reflection of this conjunction according to the
hypothetical form of judgment, finally subsumption of the hypothetical
connection under the concept of cause.14
What about the first use, the synthesis according to the categories?
Where does it come into this picture? In the Prolegomena, Kant asks:
what is it that allows me to subsume what is initially a mere logical
connection of my perceptions under the category of cause? And he
answers: I have explained this in the Critique of Pure Reason.15 Now what
he has explained in the Critique of Pure Reason, as far as the concept of
cause is concerned, is that the very experience of an objective succession is
possible in the first place only under the supposition that there is ˜˜some-
thing upon which it follows, according to a rule.™™ In other words, the
experience of an objective succession is possible only under the presup-
position that objects are ˜˜in themselves determined with respect to the
logical function of hypothetical judgment,™™ namely subsumable under
some concept of causal connection. This is how the concept of cause “
the ˜˜concept of an object, by means of which its intuition is considered as
determined with respect to the logical form of a hypothetical judgment™™ “
guides the synthesis of our perceptions for the experience of an objective
succession. This synthesis eventually makes possible the analysis of the
repeated experience into a hypothetical judgment. If we add to the
empirically tested hypothetical judgments the anticipations made possible
by the application of mathematical methods, in the context of the unity of
experience as a whole (the unity of our experience of appearances in one
space and one time), we eventually come to the conclusion that a parti-
cular connection of empirical events is ˜˜in itself determined with respect to
the form of hypothetical judgment.™™ That is to say, an event is ˜˜in itself
determined™™ (as an empirically given event) under the antecedent, the
other is ˜˜in itself determined™™ (as an empirically given event) under the
consequent of a hypothetical judgment “ and in thinking this we subsume
the connection of the two events under the concept of cause.16

See Prolegomena, x20 n, AAiv, p. 301.
See Prolegomena, x22, n. 15, AAiv, p. 305n.
For a detailed analysis of this process, see KCJ, ch. 7, pp. 167“80, and ch. 11, pp. 355“75.
See also ch. 2 in this volume, pp. 58“62 and ch. 6, especially pp. 172“6.

I believe there is a misunderstanding when Allison attributes to me the
view that categories play no role at all in judgments of perception (but
instead are present in them only under the guise of logical forms of
judgment). In my understanding of Kant™s view, they play the first role
outlined above (they guide the synthesis of a sensible manifold), just as
they play this role in any cognitive effort to relate representations to
objects they are the representation of. But they do not play the second
role outlined above (we do not subsume intuitions or perceptions under
them). This is because in a judgment of perception, we are not in a
position to assert that the object of intuition thought under the concepts
combined in our empirical judgment is ˜˜in itself determined™™ with respect
to the connection we are thinking, and thus subsumable under a category.
Now, in these two roles (guide for synthesis, universal representation
under which objects are subsumed) I maintain that according to Kant,
categories are generated by the combined use of our intuitive and dis-
cursive capacities. I now want to say something in response to Sedgwick™s
worries about this point.

As Sally Sedgwick correctly points out, I emphasize the fact that for Kant,
not all comparison is a comparison of concepts, or even a comparison of
objects geared toward the formation of concepts. There is also a strictly
˜˜aesthetic™™ comparison, one that occurs only in sensibility.17 But even
more importantly, I insist that there is for Kant a pre-discursive act of
synthesis of sensible manifolds, which is the necessary condition of the
comparison of these manifolds, a comparison that leads to forming con-
cepts that will be combined according to the logical forms of judgment.
The synthesis is governed by rules: a priori rules that guide the syntheses
to just those forms of combinations that will make it possible to compare,
and thus reflect sensible manifolds according to logical forms of judg-
ment. Those a priori rules are the schemata of the categories. In compar-
ing sensible manifolds that have been synthesized according to those
a priori rules, we generate empirical rules for apprehension, rules that
will be thought under empirical concepts. So, for instance, we have the
a priori rule: ˜˜Look for what can be recognized as remaining one and the
same thing while its properties change™™ (this is the schema for the category

<< . .

. 3
( : 33)

. . >>