. 6
( 10)


This at least would agree with Guyer™s account (cf. n. 32). But even so, the idea that we
confirm our interpretation of a perceived succession by relating it to a rule constitutes only
a secondary aspect of Kant™s argument, which depends on the first: perceiving objective
succession is presupposing a rule (as yet unknown).

possible rules according to which a given event can occur. We suppose
rules because our sensory given has been such that we could, by associa-
tions guided by our capacity to form hypothetical judgments, generate
the representation of such rules. Thus it is correct to assert that rules
hold of objects themselves, albeit objects as appearances.
Nevertheless, the argument so understood still allows us to say, at most,
that there has to be a reasonable degree of regularity in appearances for
us to be able to identify any event, or change of state of an object. It does
not allow us to assert that for any given event there is a rule, and thus the
antecedent of a rule, and thus a cause. What we have here is, on the one
hand, a rather loose epistemic principle that says that we should look for
rules to confirm our perception of something™s having happened; and on
the other hand, an even looser ontological principle that says that there is
some degree of regularity in nature. We do not have the universal, strictly
necessary, objective principle: ˜˜Everything that happens presupposes
something which it follows in accordance with a rule.™™ Nor do we have
any warrant that the rule is itself strictly necessary. In other words, to the
first question stated at the end of the first section (do we make the
presupposition, for anything that happens, that it presupposes something
which it follows in accordance with a rule?) the answer is: yes, we do make
the presupposition. This is how we generate the representation of ˜˜some-
thing that happens,™™ or an event, in the first place. But this positive answer
is considerably weaker than the one Kant would like to assert, because
˜˜rule™™ here means just this: rule, regular pattern of recognition that the
event instantiates, not strictly necessary rule, or law. To the second
question “ is the presupposition true? “ we have an even weaker positive
answer: yes, there has to be some degree of regularity in nature for any
identification of an event to be possible. But conformity to a determinate
rule is not warranted in every individual case, and there is no clear sense
in which the rule could be said to be necessary, or described as a law. And
finally, to the third question “ does the supposition warrant the transition
from judgment of perception to judgment of experience, that is, from the
statement of mere regularities to that of law-like connections? “ we would
definitely have to say, no, it does not.
But again, it is clear both from the formulation of the principle of the
Second Analogy, especially in the second edition,34 and from the texts of
the Prolegomena cited in the first section, that Kant intends to prove more

Cf. above, n. 20.

than this. He intends his principle to be asserted as true of every indivi-
dual event; of every individual event he intends to assert that it occurs in
accordance with a strictly universal causal law.
This difficulty has been widely noticed, and is in fact a major reason, as I
suggested earlier, for the lack of agreement among Kant commentators
on the interpretation of the Second Analogy. It seems that Kant™s argu-
ment needs repairing, or at the very least disambiguating. But people do
not agree on how this should be done. On the Buchdahl“Allison line,
Kant™s purpose in the Second Analogy was not at all to prove that empir-
ical objects stand under strictly universal causal laws, even less to provide a
justification for the transition from mere regularities to empirical causal
laws. All he ever intended was to prove that our perception of objective
succession (any perception of particular events) presupposes a general
concept of causal connection which allows us to think the succession as in
some way ˜˜constrained™™ and therefore objective rather than merely sub-
jective and arbitrary. This is sufficient to prove, the view goes on, that the
causal principle is not derived from experience, but is instead a transcen-
dental principle of the possibility of experience and thus also a principle of
the possibility of the objects of experience. But there is no claim on Kant™s
part that we shall find, indeed that there are, any empirical causal laws for
those particular events we are thus enabled to perceive. Causal laws can be
discovered only empirically. That we can and should anticipate them is
prescribed to us not by the understanding and its causal principle,
expounded and proved in the Second Analogy, but by reason and its
regulative idea of a universal order of nature. According to this view, then,
not only does the Second Analogy provide a positive answer only to the
first of the three questions I stated at the end of the first section, but this
answer is seriously weaker than even the formulation I myself reduced it
to. What Kant allegedly proves is not that we assume causal regularities in
nature when we perceive objective succession. Rather, what he proves is
that we are capable of distinguishing objective from subjective succession
only because we have at our disposal a causal concept to ˜˜bind down™™ the
temporal order of our perceptions. This does not involve making any
further assumption about any causal laws or even regularities this succes-
sion itself might instantiate.35
Against the Buchdahl“Allison line, Michael Friedman maintains that
Kant™s intent in the Second Analogy was (1) to defend the universal causal

See Buchdahl, Metaphysics and the Philosophy of Science, pp. 651“5; Allison, Transcendental
Idealism, pp. 222“32; see also Allison, ˜˜Causality and causal laws.™™

principle as an objective principle, and (2) to provide the ground for the
transition from mere empirical regularities to strictly universal causal laws.
That Kant so intends his Analogy (indeed, all three Analogies taken
together) is shown, Friedman argues, by the use he makes of them in the
Metaphysical Foundations of Natural Science. Particularly significant in this
regard is Kant™s explanation of the way in which Newton™s law of universal
gravitation is obtained. In ch. 4 of the Metaphysical Foundations, Kant argues
that the law of universal gravitation is obtained by the application to
Kepler™s laws, which express observed regularities of the planetary motions
in the solar system, of the universal laws of motion stated as axioms in
Newton™s Principia. And in ch. 3 of the Metaphysical Foundations, Kant argues
that Newton™s laws of motion are themselves obtained by application of the
Analogies to the empirical concept of matter. Kant™s use of the Analogies, so
construed, shows eloquently that Kant did intend his causal principle as an
objective principle, and that he did intend it as grounding the transition
from mere regularities (in this case, Kepler™s laws) to strictly necessary
causal laws (in this case, Newton™s law of universal gravitation).36
However, showing that Newton™s law of universal gravitation can be
derived from Kepler™s laws only under the presupposition of Newton™s
laws of motion, and that Newton™s laws of motion in turn are obtained as
strictly universal laws only under the supposition of the Analogies of
Experience, is not explaining what makes such application of a priori
concepts possible, or in what sense we can hold the causal concept to be
actually true of empirical objects. Such an argument shows only that the
causal principle is an a priori presupposition of Newtonian science. Kant
acknowledges as much in the Prolegomena when he analyzes the progress
from judgments of perception to judgments of experience, meanwhile
expressly sending us to the Critique for understanding what makes the
concept of cause an objective concept. Clearly, pointing out the use
made of the causal principle in natural science is in Kant™s eyes sufficient
to justify neither this use nor the causal principle itself.
There remains, then, the possibility that Kant was simply mistaken about
his own proof. He wanted his argument in the Second Analogy to prove an
objective principle asserting the existence of strictly necessary causal laws in
nature “ a principle which alone was compatible with his interpretation of
Newtonian science. But all he could actually propound was the proof of a

See Friedman™s illuminating analysis in ˜˜Kant and the twentieth century,™™ particularly
pp. 33“6. See also his ˜˜Causal laws and the foundations of natural science,™™ esp.
pp. 175“86.

principle asserting the existence of some degree of objective regularity in
nature as a condition of possibility of our experiencing events. Such a
conclusion would be close to the one Strawson reaches, in The Bounds of
Sense, after offering what he takes to be the only acceptable version of an
argument for the Second Analogy, as opposed to what he has denounced as
Kant™s ˜˜non-sequitur of numbing grossness.™™
I shall consider again Strawson™s view and compare it with mine in the
concluding section of this chapter. But first I want to stress that in fact,
Kant does offer an answer to the difficulties I raised. The reason this
answer has been absent from my account so far is that, to establish it, the
discursive model I laid out at the beginning of this chapter and then used
in my reconstruction of Kant™s proof, is not sufficient. The discursive
model has to be completed by appealing to what Kant calls our ˜˜pure™™
intuition of space and time. This is what I now want to consider.

Causality, necessity, and time
All three Analogies have one common premise, which I did not state in the
argument outlined and analyzed above, although it in fact plays an import-
ant role in the argument of all three: ˜˜Time itself is not perceived™™ (cf. B219,
B225, B233, B257). What Kant means by this is that we have no unified
temporal frame of reference within which to coordinate states of affairs and
events, except through our correlating the latter according to rules.
According to the Transcendental Aesthetic, we have an a priori intui-
tion of time. This intuition consists, very roughly, in our representing
time (1) as one time in which all particular temporal relations are deter-
mined; (2) as continuous (or what in contemporary terms we would call
˜˜dense™™: between two points in time there is always a third point whose
position is determined as being ˜˜before™™ the one and ˜˜after™™ the other);
and finally, (3) as a (unified and continuous) time within which every
state of affairs and event is completely determined, namely uniquely
individuated as to its position (all of these features hold for space too).
Now, I want to suggest that these characteristics of time as the intentional
correlate of our a priori intuition are precisely what provides Kant with
the missing link for transforming the potentially weak version of his
answers to the first and second questions we needed the Critique to
answer (see above, p. 157) into a strong version of these answers. Let
me very briefly outline how this is so.
As we saw in the previous section, the argument of the Second Analogy
in its first four steps is meant to prove that we perceive an objective

succession just in case we presuppose an antecedent objective state of
affairs upon which the succession follows according to a rule. Thus when-
ever we perceive such an objective succession we are driven to look for
something that precedes, that can be thought under the antecedent of a
hypothetical rule. Whenever we find that something C regularly precedes
the event: A, then B (say, the sun™s shining on the stone regularly precedes
the stone™s being cold, then warm; warm, then warmer) we take C to be
the cause of the succession, A, then B. More precisely, as Kant makes clear
in the final developments of the Second Analogy, the cause precedes the
full realization of the effect, but is simultaneous with the first initiation of
the effect (cf. A203/B248). Thus the correlation between cause and effect
exists at the very first initiation of the cause, and is continuously preserved
through time as long as the cause (what is thought under the antecedent
of the rule) obtains. And this preservation through time of any correlation
that actually obtains is what makes possible the empirical individuation of
states of affairs and events in time. Now, ˜˜existence at all times™™ is what
Kant describes as the schema of necessity, that is, the sensible feature by
which one may recognize an empirical object as necessary (see A145/B184).
I suggest that the only possible candidates for being something that
possesses necessity in this sense (˜˜existence at all times™™), are the law-
like correlations between states of affairs and events preserved through
time. These law-like correlations are the empirical realization of our a
priori intuition of time as one, continuous time which is the locus of the
complete determination (individuation) of events and states of affairs.
Because it has to be thought as thus preserved through time (for the
unity of time to be empirically realized), the connection between an
event and ˜˜what precedes it, according to a rule™™ should be thought as
a necessary connection.37

In the Transcendental Deduction of the Categories, the argument for the objective validity
of the categories is completed only when Kant states that space and time themselves, as
formal intuitions, stand under the unity of apperception, and thus under the categories.
Thus anything given in space and time is by that alone already susceptible to being thought
under the categories (see x26 in the B Deduction, and chs. 1 and 3 in this volume, esp.
pp. 32“5 and 67“9). Earlier in this chapter I suggested that appealing to the role of our
time-intuition in individuating objects and events in order to complete the argument of the
Second Analogy, repeats an earlier move in the argument of the Transcendental
Deduction: a move from conditions of possibility for thinking an object to conditions of
possibility for the object itself. This is, I suggest, what we see happening here: (5) in my
outline of the argument of the Second Analogy said only that we experience objective
succession only if we presuppose something, upon which it follows according to a rule; (6)
said that objective succession itself presupposes something upon which it follows, etc. The

The argument just outlined provides an interesting view of the differ-
ence between the necessity of the analytic connection of concepts (˜˜If
there is perfect justice, the obstinately wicked are punished™™), and the
necessity of the synthetic connection of events in time (˜˜If the sun shines
on the stone, the stone gets warm™™). What I am suggesting is that for
Kant, in the latter case, the a priori intuition of time makes up for the
lack of an analytic connection of concepts. If we go back to the discursive
model expounded in the first section of this chapter, we can thus say the
following. On the one hand, the modus ponens based on a synthetic and
empirical hypothetical premise is distinct from a modus ponens based on
an analytic premise: we have and can have no complete concept of the
state of the world at any instant t1 when the sun is shining on a stone,
which could yield the complete concept of the state of the world at an
instant t2 such that the stone™s heating up would be contained in it. But
on the other hand, no state of an object would be individuated in time
(have a completely determinate position in time) unless rules of correla-
tion of events in time were true ˜˜at all times.™™ This is no mere epistemic
condition for our knowing objects and events, but a condition for any
object™s being an object for us in the first place: a thing endowed with
recognizable properties, and individuated in space and time. In other
words, according to Kant the preservation at all times of the empirically
attested rules of correlation of events and states of affairs (and thus their
strict necessity) is a transcendental condition for the representation of
objects, i.e. for objects themselves as appearances. For only through such
preservation of empirical correlations through time can the unity, con-
tinuity, and ordering of our pure temporal intuition be realized in
empirical objects of knowledge (appearances).
That the causal principle is a principle of ordering by way of which the
order of our temporal intuition is realized in appearances is just what
Kant says in the course of the third exposition of his proof:
Understanding belongs to all experience and to its possibility, and the
first thing that it does for this is not to make the representation of the
objects distinct, but rather to make the representation of an object
possible at all. Now this happens through conferring temporal order
on the appearances and their existence . . . [Thus] arises a series of
appearances, in which by means of the understanding, the very same

move is justified, according to Kant, by the fact that there would be no formal intuition of
time as that in which any object at all is individuated, unless the conditions for thinking the
objects were realized as conditions for these objects themselves, individuated by their
position in space and in time. I say more on this point in what follows in the main text.

order and continuous [stetigen] connection in the series of possible perceptions is
produced and made necessary as is encountered a priori in the form of inner
intuition “ time “ wherein all perceptions must have their place [my emphasis].

Where does this now leave us with respect to the three questions we
needed the Critique to answer? (1) Do we presuppose the truth of the
proposition ˜˜everything that happens follows upon something else
according to a rule™™? Kant™s response: yes, we do. We would not perceive
any succession as an objective succession unless we did make this presup-
position. (2) Is the principle true, according to which ˜˜everything that
happens presupposes something which it follows in accordance with a
rule™™? Kant™s response: yes, the principle is true. For the complete
determination of the spatiotemporal position of objects and their states
is achieved only by the universal correlation of appearances determining
each other™s state, according to rules. But that they necessarily are
completely determined as to their position in space and time is a priori
warranted by their belonging to one space and one time, the ˜˜pure
intuitions™™ expounded in the Transcendental Aesthetic and the ˜˜formal
intuitions™™ standing under the unity of apperception, according to the
Transcendental Deduction of the Categories. (3) How does admitting
the truth of the principle justify the transition from asserting observed
regularities to asserting that these regularities are strictly necessary
causal laws? Kant™s answer: there is no definitive justification in any
particular case. It is always possible to mistake a mere regularity (a
mere repeated succession of similar events) for a necessary connection
(a succession that occurs according to strictly necessary causal laws). But
what the principle does tell us is that all events do obey such necessary
connections, because without such connections there would be no unity
or continuity of empirically real time, and no complete determination of
empirical events (no individuation in time).38

This is where the regulative idea of a universal order of nature “ mentioned earlier in my
discussion of Allison and Buchdahl “ comes to play an important role. Whenever we do
allow ourselves, in any particular case, the transition from asserting a mere regularity (˜˜if
the sun shines on the stone, the stone gets warm™™) to asserting a causal connection (˜˜the sun
warms the stone™™), not only do we presuppose the truth of the universal causal principle as
an ontological principle in the realm of appearances, as established by the first Critique. But
moreover we suppose, as an epistemic (therefore ˜˜merely regulative™™) principle, that the
empirical regularity we have discovered is so related to the universal order of nature that it
is correct to assign precisely to this regularity the necessary character of a causal law. In the

The burden of the proof of the Second Analogy thus hinges on accepting,
(1) that we do have an a priori intuition of one time as the condition for
there being any object at all for us, and (2) that empirical time-
determinations (time-determinations of empirical objects, as appear-
ances) must exactly map the properties of this a priori intuition, most
notably the unity and continuity of temporal determinations, and the
complete determination (individuation) of objects and their states in time.
Kant™s argument for the nature of space and time is provided in the
Transcendental Aesthetic and completed in the Transcendental
Deduction of the Categories, in the Critique of Pure Reason. Evaluating
any aspect of that argument is beyond the limits of this chapter. Let me
just point out some of the possible outcomes of such an evaluation. We
might conclude that the argument is sound. This would be good for the
˜˜strong™™ version of the Second Analogy. Or, we might deny altogether
that we have anything like an a priori intuition of time (and space) as a
condition for there being any object at all for us. Or, we might accept that
we do have something like the a priori intuition of time (and space) Kant
claims we have (an intuition of time and space as one, continuous, infinite,
and the condition for any experience of object). But we might maintain
that it is a fact about ourselves for which we can give a naturalistic account,
just as we can offer a naturalistic account for the fact, say, that we perceive
ordinary middle-sized objects as belonging to a three-dimensional space
or the sun as revolving around ourselves rather than ourselves as revolv-
ing around the sun. We might then think that even if Kant is correct in his
account of our experience of objective succession, we need to question
rather than endorse the assumptions concerning the supposedly objective
features of the world our ˜˜natural™™ representation of time may lead us to
form (such as the objective necessity of the causal connection).
If so, we might then be left with the weaker version of Kant™s argument,
outlined in the second section. As I noted earlier, the outcome of this
weaker version bears a close resemblance to the outcome of the argument
Strawson, in The Bounds of Sense, says Kant should have upheld but did not
uphold,39 stumbling instead into the pitfall of the ˜˜non-sequitur.™™
However, there are major differences between Strawson™s reconstruction
of what he takes to be the only acceptable argument for the Second

third Critique, Kant adds that making use of such a regulative principle presupposes a
principle of reflective judgment, that of the ˜˜logical purposiveness of nature.™™ See Kant,
Critique of the Power of Judgment, Introduction, iv, AAv, pp. 180“1. See also Critique of
Judgment, First Introduction, vi, AAxx, p. 216.
See Strawson, The Bounds of Sense, pp. 140“6.

Analogy, and the one I think Kant did defend. Strawson™s version revolves
around the question: what does the world have to be like for our experi-
ence of objective succession to be possible at all? His answer is, very
roughly, that for such an experience to be possible, nature must offer a
background of regularity in the correlated persistence and alterations of
objects. Kant™s argument as I understand it revolves around the question:
what activities of our discursive and receptive capacities are necessary for
our experience of objective time-determination to be possible at all? It
thus relies on an elucidation of acts of the mind which Strawson, at least in
The Bounds of Sense, scornfully rejects as belonging to the ˜˜imaginary
subject of transcendental psychology.™™40
But Strawson™s rejection is damaging to our understanding of Kant™s
argument, which has for its indispensable background Kant™s aesthetic
as a theory of sensibility, Kant™s logic as a theory of discursive capacities,
and ultimately Kant™s transcendental psychology as an account of how
we generate, through the exercise of our imagination guided by our
discursive capacities, our representation of a unified world of objects
uniquely identifiable and re-identifiable in space and time. I have sug-
gested that Kant™s argument for a strong version of the causal principle
ultimately depends upon his claims concerning our a priori intuitions of
space and time and the conditions of their empirical realization. Given
that Kant™s theory of space and time is also the most controversial aspect
of the system of transcendental conditions of experience he sets up in the
first Critique, it is no surprise if the point of greatest resistance we reach in
examining his argument for the causal principle is met precisely there.

A P P E N D I X : T H E F I V E E X P O S I T I O N S O F K A N T ™S A R G U M E N T

Kant™s exposition of the argument proper runs from ¶¶1 to 17, where ¶ 1 is
the first paragraph of the proof added in B. The order of the proofs I
recount starts with the first proof in A, thus ¶3, and ends with the proof
added in B, thus ¶¶1“2. I first give an outline of the respective structure of
each version of the argument, and then give textual support for each in
particular. My view is that in repeating the argument like this, Kant is not
just groping for the right formulation. Rather, I suggest Kant proceeds as
follows. (1) In A, he gives a first, detailed exposition of his proof, supported

Ibid., p. 32.

by the now famous example of the difference between (successively) per-
ceiving the (objectively simultaneous) parts of a house and (successively)
perceiving the (objectively successive) positions of a ship (¶¶3 to 6,
A189“94/B234“9). (2) He repeats the proof as an indirect proof (¶¶7“8,
A194“5/B239“40). He then raises an objection in empiricist style: why
suppose that the representation of causal connection precedes experience
rather than being derived from it (¶¶9“10)? Responding to this objection
(¶11) leads Kant to (3) a third exposition of the proof, where the collabora-
tion between the discursive role of the understanding and the intuitive role
of sensibility in perceiving objective succession becomes more prominent
than it was in the previous expositions (¶¶12“16, A198“201/B243“6). This
is important because indeed urging that understanding is necessary for the
very combinations of perceptions in sensible intuition is Kant™s answer to
the empiricist objection. Finally, (4) Kant recapitulates his proof one last
time, in a short paragraph (¶17, A201“2/B246“7). In the B edition, he
prefaces the whole exposition with (5) a very compressed new exposition of
the proof (¶¶1“2, B232“4).
Let me now give the textual support for this reading. The numbering
is mine.41

First exposition (¶¶3 to 6, A189“94/B234“9)
11 The apprehension of the manifold of appearance is always successive.
The representations of the parts follow upon one another.
21 Whether they also follow one another in the object is a second point
for reflection, which is not contained in the first.

(Here comes a long parenthesis on the notion of an object, which Kant
concludes [end of ¶3, A192/B237] with the statement: ˜˜that in the
appearance which contains the condition of this necessary rule of appre-
hension, is the object.™™)
31 Yet I also note that, if in an appearance that contains a happening, I
call the preceding state of perception A, and the following one B, then B
can only follow A in apprehension, but the perception A cannot follow
but only precede B. For instance, I see a ship move downstream. My
perception of its position downstream follows the perception of its posi-
tion upstream, and it is impossible that in the apprehension of this

Translations are mine, although I have tried as much as possible to follow Paul Guyer and
Allen Wood™s translation. I put my own comments on Kant™s argumentative moves in

appearance the ship should first be perceived lower downstream and
afterwards upstream. The order in the succession of perceptions in
apprehension is therefore here determined, and the apprehension is
bound to it. (A192/B237)

(Kant then contrasts this case with the example, previously given, of
perceiving a house, where the order of apprehension is arbitrary: ˜˜In
the series of these perceptions there was no determinate order that
made it necessary when I had to begin in the apprehension in order to
combine the manifold empirically.™™ On the contrary . . . )
41 This rule is always to be found in the perception of that which
happens, and it makes the order of the perceptions that follow one
another (in the apprehension of this appearance) necessary. (A193/B238)

In accordance with such a rule there must therefore lie in that which in
general precedes an event the condition for a rule, according to which
the event always and necessarily follows. (A194/B239)

(41) as stated here contains in effect (4), (5), and (6) in my analysis of
the argument as outlined above: from the fact that (31) I perceive a
succession as objective just in case the succession in apprehension is
order-determinate, and (41) this is so just in case a rule makes the
order determinate, it follows that (51) perceiving an event is supposing
a rule, i.e. (61) the event itself presupposes a rule, or ˜˜in what precedes
an event there must be the condition for a rule.™™

Second exposition (indirect proof) (¶¶7“8, A194“5/B239“40)
12 Suppose nothing preceded an event, upon which the latter must
follow, according to a rule.

(Negation of [41] or of [6] in my outline of Kant™s argument.)
22 Then all succession of perception would be determined solely in the
apprehension, i.e., merely subjectively, but it would not thereby be
objectively determined which of the perceptions must really be the
preceding one and which the succeeding one.

(Repetition of [11] and [21] in the direct proof: the negation of [41] in the
direct proof, or [4], [5], [6] in my outline, leaves us only with [1] and [2].)
32 In this way we would have only a play of representations that would
not be related to any object at all, i.e., by means of our perception no
appearance would be distinguished from any other as far as the

temporal relation is concerned, since the succession in apprehension is
always the same, and there is thus nothing in the appearance that
determines it so that a determinate succession is made necessary as
objective. I shall thus not say that in the appearance two states follow
upon one another, but only that an apprehension follows upon another.

(Negation of [31] in the direct proof: negation of the order-determinateness
of apprehension, and thus of any representation of objective succession.
But the fact is, we do have order-determinateness, and thus repre-
sentation of objective succession as distinct from merely subjective
succession in apprehension. Therefore, premise [12] in the indirect
argument is false. We can thus assert [52]):
52 If, therefore, we experience that something happens, then in so
doing we always presuppose that something precedes it, upon which
it follows according to a rule. For without this I would not say of the
object that it follows, for the mere succession in my apprehension, if it is
not, by means of a rule, determined in relation to something that
precedes, does not justify a succession in the object. Thus it is always
with respect to a rule according to which the appearances are deter-
mined in their succession, i.e. as they happen, by the preceding state,
that I make my subjective synthesis (of apprehension) objective; only
under this presupposition is the experience of something that happens
even possible.

(Note that clearly, according to this formulation, in Kant™s mind the
epistemic point [52] is also the ontological [transcendental] point [62]:
not only do we presuppose something that precedes, but the objective
succession [the event] itself presupposes something that precedes,
according to a rule.
In ¶¶9 and 10, Kant formulates the empiricist objection mentioned
above. In ¶11, he announces that the answer to this objection
depends on a correct understanding of what we do when we relate our
representations to an object [A197/B242]. This introduces his third

Third exposition (¶¶12“16, A198“201/B243“6)
(Note that in this exposition, [3] and [4] in my outline are not clearly
distinguished. This makes even more visible the interdependence
between the rule-governed character of the objective succession
and the irreversibility (order-determinateness) of the subjective

13 In the synthesis of appearances the manifold of representations is
always successive.
23 Now no object at all is thereby represented, since through this succes-
sion, which is common to all apprehension, nothing is distinguished
from anything else.
33 and 43: As soon as I perceive, or presuppose [wahrnehme oder voraus-
annehme], that in this succession there is a relation to the preceding state
out of which the representation follows according to a rule, I represent
something as an event, or something that happens, i.e. I cognize an
object that I must posit at a determinate place in time which after the
preceding state cannot not be otherwise assigned . . . Thus it happens
that an order is given to our representations, in which the present
(insofar as it has come to be) points to some preceding state as an, albeit
still indeterminate, correlate of this event that is given, a correlate which
relates as a determinant [bestimmend] to this given as its consequence, and
connects it with itself necessarily in the sequence of time.

(Here comes a long development on the conditions of time perception
[¶¶13“14, A199“200/B244“5], where Kant explains the respective roles
of understanding and sensibility in our representation of objective tem-
poral succession. Kant then gives what is perhaps his most explicit
formulation of [5] and [6]):
53 That something happens is therefore a perception which belongs to a
possible experience. This experience becomes actual when I regard the
appearance as determined in its position in time, and therefore as an
object that can always be found in the connection of perceptions in
accordance with a rule.
63 Now this rule for determining something with respect to its temporal
succession, is that in what precedes the condition is to be encountered
under which the event always (i.e. necessarily) follows. The principle of
sufficient reason is thus the ground of possible experience, that is, of the
objective cognition of appearances in respect of their relation in the
successive series of time.

Fourth exposition (¶17, A201“2/B246“7)
14 To all empirical cognition there belongs the synthesis of the manifold
through the imagination, which is always successive; i.e., in it the repre-
sentations always follow upon each other.
24 But the succession is not at all determined in the imagination as to its
order (what must precede and what must follow), and the series of

successive representations can be taken backwards just as well as
34 But if this synthesis is a synthesis of apprehension (of the manifold of a
given appearance), then the order is determined in the object, or to
speak more correctly, there is in the synthesis an order of succession that
determines an object.
44 In accordance with this order, something must necessarily precede,
and when this is posited, then the other must necessarily follow. If, then,
my perception is to contain the cognition of an event, i.e. that something
actually happens, it must be an empirical judgment in which one thinks
that the succession is determined, i.e. that it presupposes with respect to
time another appearance, upon which it follows necessarily, or accord-
ing to a rule. . . .
54 Thus the relation of appearances (as possible perceptions) according
to which the existence of that which succeeds (what happens) is deter-
mined in time necessarily and in accordance with a rule by something
that precedes, is the condition of the objective validity of our empirical
judgments with respect to the series of perceptions, and thus of their
empirical truth, and thus of experience.
64 Hence the principle of the causal relation in the succession of appear-
ances is valid for all objects of experience (under the conditions of
succession) since it is itself the ground of the possibility of such an

Fifth exposition (¶¶1“2, added under the title ˜˜Proof™™ at the beginning
of B: B232“4):
(The proof actually begins with ¶2. ¶1 is a reminder of a result from the
first Analogy: all objective change [Wechsel, transition from one state of
affairs (A) to another (B)] is an alteration [Veranderung, change of states
of a permanent substance].)
15 I perceive that appearances follow one another, that is, that there is a
state of things at one time the opposite of which was in the preceding time.
25 Thus I am really connecting two perceptions in time. Now connection
is not the work of mere sense and intuition, but is here the product of a
synthetic capacity of the imagination, which determines inner sense with
regard to temporal relation. But imagination can combine the two states
in question in two ways, so that either the one or the other precedes in
time; for time cannot be perceived in itself, nor can what precedes and
what follows in objects be as it were empirically determined in relation to
it. I am therefore conscious only that my imagination sets the one state

before and the other after, not that the one state precedes the other in
the object; or in other words, through the mere perception the objective
relation of the appearances that are succeeding one another remains
35 Now in order for this to be cognized as determined, the relation
between the two states must be thought in such a way that it is thereby
necessarily determined which of them must be placed before, and which
after, rather than vice versa.
45 But the concept that carries with it a necessity of synthetic unity can
only be pure concept of understanding, which does not lie in perception;
and here it is the concept of the relation of cause and effect, the former of
which determines the latter in time, as its consequence “ and not as
something that might simply precede in imagination, (or not even be
perceived at all).
55 Therefore it is only insofar as we subject the succession of appear-
ances, and therefore all alteration, to the law of causality, that experience
itself “ i.e. the empirical cognition of appearances “ is possible.
65 Consequently the appearances themselves, as objects of experience,
are possible only in conformity with this law.

(Note that this proof follows exactly the order of the premises in my
outline of the argument. In one important respect, however, I find this
exposition less clear than any of the expositions in A: the representation
of the temporal order-determinateness of objective succession is directly
related to the causal principle itself, without the intermediate step of
˜˜presupposing a preceding state, upon which the succession follows,
according to a rule.™™ I think this lack blurs the nature of Kant™s argu-
ment, for it relegates into the shade the logical model I analyzed in part
one. But this model, I argued, in fact plays a prominent role in Kant™s


Kant claimed that human beings™ representation of the world depends on
a system of fundamental categories or ˜˜pure concepts of the understand-
ing.™™ He also claimed that these categories were originally nothing other
than elementary logical functions, which find expression in logical forms
of judgment. Kant expounded these functions in a systematic ˜˜table™™
which then became the architectonic principle not only for the Critique of
Pure Reason, but also for the Critique of Practical Reason and the Critique of
Judgment. In a famous footnote to the Metaphysical Foundations of Natural
Science, Kant claimed that as long as one accepted the two cornerstones of
his doctrine “ the merely sensible, receptive character of our intuitions,
for which space and time are a priori forms and the derivation of cate-
gories from logical functions of judgment “ then it mattered little if the
details of his proofs (in particular, the details of his transcendental deduc-
tion of the categories) failed to carry complete conviction in the eyes of his
readers. For the two main points of his demonstration, as far as he was
concerned, were sufficiently established. Those two points are that (1) we
have a priori concepts of objects originating in the understanding alone;
and (2) these concepts can be applied in cognition only to appearances
(that is, to objects given in accordance with the a priori forms of space and
time), not to things as they are in themselves.1

Cf. Metaphysical Foundations, AAiv, 475“6n.


The problem is that precisely the two purported pillars of the critical
system are what consistently met, very early on, with the most radical
skepticism on the part of Kant™s readers. Kant™s logic is charged with
being archaic, caught within the narrow bounds of Aristotelian predicative
logic. It is also charged with psychologistic fallacy: Kant is mistaken in
supposing that logical forms are in any sense descriptions of acts of our
minds. As for the role he assigns to a priori forms of intuition in grounding
synthetic a priori judgments, Kant is charged with relying on a conception
of arithmetic and geometry made obsolete by the development of non-
Euclidean geometries and modern quantificational logic; he is also charged
with a misguided absolutization of a Newtonian model of natural science
made obsolete by revolutions in nineteenth- and twentieth-century physics.
In the present chapter I shall examine Kant™s claims concerning the
second of the two cornerstones mentioned above: the derivation of cate-
gories from logical functions. To do this I shall focus on one particular
case: the category of community, its relation to the logical function of
disjunctive judgment, and its application to appearances in the so-called
˜˜principle of community,™™ the Third Analogy of Experience. This case is
interesting for two main reasons. First, it is the most difficult to defend.
Kant himself was aware of this, and took great pains to explain why even
in this case, however implausible it might seem, the relation he main-
tained between logical functions and categories does in fact hold. The
general view of Kant commentators, however, is that his defense remains
utterly unconvincing. I shall argue, on the contrary, that the correspond-
ence Kant wants to establish between the logical function of disjunctive
judgment and the category of community is an important and interesting
one, although indeed it is more complex than any other. But this very
complexity is in fact my second reason for focusing on this case: what
makes the category of community difficult to grasp is that it can be under-
stood only in connection with the other two categories of relation (and
even with the previous two ˜˜titles™™ of categories, quantity and quality).
This being so, examining Kant™s argument in this case should also give us
some insight into his overall argument on the relation between logical
functions, categories, and the application of categories to appearances.
This chapter is in four parts. In the first, I shall briefly expound the
relation Kant claims to establish between logical functions of judgment
and categories.
In the second part, I shall examine Kant™s logical form of disjunctive
judgment and its relation to the category of community or universal

In the third part, I shall examine Kant™s proof of the Third Analogy of
Experience, namely his proof that necessarily, things we perceive as
simultaneously existing exist in relations of universal interaction or, in
Kant™s terms, of dynamical community.
The lesson of this examination, I shall suggest, is that neither Kant™s
general claim concerning the role of logical functions of judgment in
generating our representations of objects, nor even his more particular
claim concerning the relation between the form of disjunctive judgment
and the category of community, deserve the summary dismissal they are
often met with. Rather, Kant™s argument provides an intriguing account
of how elementary functions of minds such as ours might be responsible
for the unity of our unsophisticated, ordinary perceptual world, as well
as for the relation between this world and our more sophisticated,
scientific worldview.
Finally, in the fourth and concluding part I shall suggest that paying
close attention to the Third Analogy (the ˜˜principle of community™™) and
not just to the better-known Second Analogy (Kant™s response to Hume
on the concept of cause and its objective validity) give us important
insights into the unity of Kant™s critical system as well as its relation to
its philosophical posterity.

Logical functions and categories: the understanding
as a capacity to judge
In the Critique of Pure Reason, Kant explains that the understanding, or
intellect as a whole “ the intellectual faculty at work in forming concepts,
combining them in judgments, combining judgments in inferences, and
finally constituting systems of knowledge “ the intellect that produces all
this is essentially a Vermogen zu urteilen, a capacity to form judgments.2 In
other words, describing the features of the intellect that make it capable
of forming judgments is by itself describing just those features that also
make it capable of forming concepts, inferences, systems of thought and
knowledge. This is because, as Kant puts it in the section that precedes
his table of logical functions of judgment, if we start with the traditional
notion that the understanding is a capacity for concepts, we soon find,
upon examination, that we form concepts only for use in judgments, and
this use itself involves implicit inferential patterns and their systematic

Cf. A69/B94, A81/B106. Cf. chs. 2 and 4 in this volume.

Kant™s explanation of this point can be summarized as follows.
Concepts, as he defines them, are ˜˜universal and reflected representa-
tions.™™ They are formed by comparing individual objects, focusing on
the common features or marks of these objects and ignoring their
differences.3 A concept is thus a conjunction of common marks under
which one may recognize a class of objects as falling under the same
concept. But this means that forming concepts is forming implicit judg-
ments: for instance, forming the concept ˜˜tree™™ is forming the implicit
judgment, ˜˜everything that has a trunk, branches, and roots, is a tree™™
(and conversely, ˜˜everything that is a tree has a trunk, branches and
roots™™). On the other hand, forming such a judgment is forming the
major premise for a possible syllogistic inference, for instance, ˜˜every-
thing that has a trunk, leaves, and roots, is a tree; this tiny thing here has
a trunk, branches, roots; therefore it is a tree.™™ Judgments and syllogistic
inferences, systematically arranged, give rise to universal hierarchies of
genera and species under which individual things are classified into
natural kinds; thus they give rise to systematic knowledge.
It is by virtue of their form that judgments can thus be the source of
the systematic unity of knowledge. What Kant calls the form of a judg-
ment is the way concepts are combined in judgment.4 When we analyze
the ˜˜mere form™™ of judgment, we have to consider concepts themselves
as to their ˜˜mere form,™™ namely their universality: their being combin-
ations of marks common to a multiplicity of individual objects.5 The
˜˜form™™ of a judgment is thus the way in which concepts, as universal
representations, are combined in it. Kant™s table of logical forms of
judgment6 is a table of just those modes of combination of concepts
that are minimally necessary for the functions of intellect briefly outlined
above to emerge: subsumption of individual objects under concepts,
syllogistic inference, the systematic arrangement of knowledge and

This is true also of the categories, but does not challenge their apriority. On this point, see
above, ch. 1, pp. 26“9; also KCJ, p. 121.
Cf. Jasche Logic, x18, AAix, p. 101. Also Reflexionen 3039 and 3040, AAxvi, pp. 628“9.
Jasche Logic, x2, xx 4“8, AAix, pp. 93“6; Reflexion 2855, AAxvi, p. 547; Reflexion 2859, AAxvi,
p. 549.
On Kant™s notion of a ˜˜function™™ of judgment, see A68/B93. Cf. also A70/B95. If we rely on
Kant™s explanations in these texts, logical function and logical form of judgment seem to be
distinguished as (1) the structure of an act “ a structure that makes the act adequate to
achieving a specific purpose, that of ˜˜ordering representations under a common represen-
tation™™ “ and (2) the result of the act: the mode of combination of concepts, or the ˜˜form™™ of
the judgment resulting from the act. On this point, see above, ch. 4, pp. 92“5.

I now want briefly to review this table, with only the degree of detail
necessary to situate the particular function of disjunctive judgment
within it.
Recall that concepts, in Kant™s logic, are defined as ˜˜universal and
reflected representations™™ (that is, as universal representations formed
by comparing objects, selecting common marks, leaving aside particular
marks by which the objects thought under the same concept nevertheless
differ from each other). So considered, the kinds of combinations con-
cepts may enter into in judgment are exclusively what Kant calls ˜˜concept
subordinations,™™ where the extension of one concept (everything that falls
under the concept) is, as a whole or only in part, included in, or excluded
from, the extension of the other. The first two titles in Kant™s table
(quantity and quality, in their first two moments: universal and particular,
affirmative and negative) describe precisely the four possible cases just
mentioned: inclusion of the extension of a concept in the extension of the
other, or exclusion therefrom (affirmative or negative judgment, As are
B or As are not B), in totality or in part (universal or particular judgment,
all/no As are B, some As are/are not B).7 To these four possible combin-
ations that exhaust the possible cases of concept subordination, Kant adds,
under each of the first two titles (quantity and quality), a form of judgment
that relates concept subordination, respectively, to individual objects
(singular judgment under the title of quantity), and to the unified logical
space within which all spheres of concepts reciprocally limit each other
(˜˜infinite™™ judgment, A is not-B).
The raison d™e tre for the third title, that of ˜˜relation,™™ is more difficult to
elucidate. Kant notes that a judgment, considered according to the
forms of relation, combines two concepts (categorical judgments) or
two judgments (hypothetical judgment, where the connective is
˜˜if . . . then™™) or several judgments (disjunctive judgment, where the
connective is ˜˜either . . . or™™) (A73/B98). This is hardly any explanation
at all. We can do better if we consider the relation of judgment to
syllogistic inference mentioned above. We saw that combining concepts
in a universal categorical judgment (all As are B) was eo ipso obtaining the
premise for a syllogistic inference in which one might attribute the
predicate B to anything thought under the subject-concept A. This is

On these explanations and the privilege given to the point of view of extension in defining
the form of judgment as to its quantity and quality, cf. Jasche Logic, xx21“2. Note that
consideration of the extension of concepts, and of judgment as expressing the inclusion
or exclusion of concepts™ respective extension (Umfang), is also prominent in the explana-
tions Kant gives at A71“2/B96“8.

why Kant calls a universal categorical judgment a rule, and the subject-
concept in such a judgment the condition of a rule (for instance, the
concept ˜˜man™™ functions as a condition of the rule: ˜˜all men are mortal™™).
The term ˜˜condition™™ should here be understood as meaning ˜˜suffi-
cient™™ not ˜˜necessary™™ condition: that some entity X be a man is a
sufficient condition for its being mortal. Or, if X is a man, then X is
mortal. Since being a man is a sufficient condition for being mortal,
subsuming any individual X under the concept ˜˜man™™ provides a suffi-
cient reason for asserting of it that it is mortal.8
However, there are other kinds of conditions of a rule. One is that of
hypothetical judgment, the second title of relation in Kant™s table.
According to this form, a concept is not by itself, on its own, the condition
for attributing a certain mark to an object thought under the concept.
Instead, one can do so only under an added condition: ˜˜If C is D [added
condition], then A is B™™ (and thus any object X subsumed under the
concept A receives the predicate B under the added condition that some
relevant C is D). Kant™s example is the proposition: ˜˜If there is perfect
justice, then the obstinately wicked will be punished.™™ (Implicit possible
subsumption: any individual falling under the concept ˜˜wicked™™ is
doomed to be punished, under the added condition that the state of the
world be one of perfect justice). Or, to take up an example Kant uses in the
Prolegomena, ˜˜If the sun shines on a stone, the stone gets warm™™ (implicit
possible subsumption: any individual falling under the concept ˜˜stone™™
gets warm, under the added condition that the stone be lit by the sun).9
A third kind of condition of a rule is that expressed in a disjunctive
judgment. The proper function of this form of judgment is to recapitu-
late, as it were, the possible specifications of a concept. According to this
form, one divides a concept, say A, into mutually exclusive specifications
of this concept, say B, C, D, E: A is either B, or C, or D, or E. There are
two different ways in which one might consider it as a possible rule for
subsumption, and thus a rule by virtue of which one might attribute
some predicate to any individual thought under the condition of the
rule. One is to say that the subject of the disjunctive judgment, say A, is
the condition of the rule ˜˜A is either B, or C, or D, or E,™™ so that being
thought under A is a sufficient condition for being thought as falling

On the notion of the condition of a rule, see A322/B378; also Jasche Logic, x58, AAix, p. 120.
Reflexionen 3196“3202, AAxvi, pp. 707“10.
Cf. Prolegomena, AAiv, p. 312. And see above, ch. 6, pp. 151“3, for the difference between
Kant™s hypothetical judgment and the material conditional of modern propositional logic.

under either B, or C, or D, or E. But this is not terribly informative.
A more interesting way (corresponding to the classical inferences in
modus ponendo tollens or modus tollendo ponens) is to consider the assertion
of any one of the specifications (B, C, D, or E) of the divided concept A as
a sufficient condition for negating the others, and conversely consider-
ing the negation of all but one as a sufficient condition for asserting the
remaining one: A is B under the condition that it be neither C, nor D, nor
E; A is neither C, nor D, nor E, under the condition that it be B; and so
on.10 Note also the close connection between the forms of disjunctive
and infinite judgment: these forms jointly contribute to the constitution
of a unified logical space within which concepts delimit each other™s
sphere, and thus contribute to the determination of each other™s
About the fourth title, that of modality, Kant explains that it does not
add to the ˜˜content™™ of judgments. What Kant seems to mean is that the
modal determinations of judgment do not determine a specific difference
in the function of judging “ by contrast with quantity, according to which
one subordinates all or part of the extension of two concepts; with quality,
according to which the extension of the subject-concept is included in or
excluded from the extension of the predicate-concept; and with relation,
according to which one states that the predicate-concept can be asserted of
individual objects under the condition that the subject-concept itself be
asserted of them, or under an added condition (expressed in the ante-
cedent of a hypothetical judgment). Instead, the modality of a given
judgment expresses only ˜˜its relation to the unity of thought in general.™™
Correspondingly, Kant™s modality of judgments finds no particular lin-
guistic expression, contrary to quantity (˜˜all™™ or ˜˜some™™), quality (˜˜is™™
simpliciter or ˜˜is not™™) and relation (˜˜is,™™ ˜˜if . . . then,™™ ˜˜either . . . or™™).
Instead, in the examples Kant gives for ˜˜problematic,™™ ˜˜assertoric,™™ or
˜˜apodictic™™ judgments, modality is marked by no particular modifier
but consists, he says, merely in the ˜˜value of the copula™™ in the judgment,

Just as Kant™s hypothetical judgment is different from truth-functional material condi-
tional, so Kant™s disjunctive judgment is different from truth-functional disjunction. First
of all, as we just saw, Kant™s disjunctive judgment is a disjunction of predications: a concept
A is specified as either B, or C, or D, or E (and thus any object falling under A falls under
either B, or C, or D, or E). Second, the disjunction is exclusive, not inclusive: what is
asserted in a disjunctive judgment is that if one of the disjunct predicates belongs to the
subject, then the others do not, and conversely. Thus the meaning of the connective
˜˜either . . . or™™ grounds the forms of inference in modus ponendo tollens and tollendo ponens:
asserting one of the predicates is a sufficient reason for negating the others, negating all
but one is a sufficient reason for asserting the remaining one.

as determined by its place in a hypothetical or disjunctive judgment or in
syllogistic inferences (A74“6/B100“1).
These remarks are certainly too brief to give a full account, even less
an evaluation, of Kant™s table. My hope is that they at least shed some
light on the systematic character and, in the end, the simplicity of Kant™s
table: it displays forms (1) of concept subordination (first two moments
of quantity and quality), (2) under either an ˜˜inner™™ or an ˜˜outer™™
condition (first two moments of relation), which also takes into account
(3) the subsumption of singular objects under concepts (singular judg-
ments, third moment of quantity) and (4) the unity of concept subordin-
ation in a system of genera and species (infinite and disjunctive
judgments, third moments of quality and relation). Finally, (5) the
place of each judgment in other judgments or in inferences (its ˜˜relation
to thought in general™™) determines its modality. It is no whimsical choice
on Kant™s part to have presented these forms as a table. The tabular
presentation makes perspicuous ˜˜at one glance™™ the systematic whole of
elementary logical functions at work for the production of any of the
judgments by means of which individual objects given in sensibility are
subsumed under concepts.
Kant calls analysis the use we make of the understanding according to
the logical forms laid out in his table. By analysis here he does not mean
simply or even primarily analysis of concepts, i.e. the laying out of the
marks that constitute the content of a given concept. He means the
analysis of representations given in sensibility so as to generate concepts
from them, by means of the aforementioned operations: comparing
individual objects, focusing on common features or marks of these
objects and setting aside their differences.11 Now, such analysis presup-
poses that the objects in question are combined together in some way, in
order to be thus compared and subsumed under concepts. And not only
this: they need to be recognized as a plurality of individual things that
remain identical through time.12 For this much more than simply bring-
ing together objects for comparison is needed. What is needed is a
process of generating the representation of these objects themselves as
numerically identical individuals persisting through time. And for this,
our representation of space and time themselves need to be unified and

On this notion of analysis, cf. A76/B102. So considered, analysis consists in the operations
of ˜˜comparison, reflection, abstraction™™ described in Jasche Logic, x6, AAix, p. 94;
cf. Reflexion 2876, AAxvi , p. 555. And above, ch. 1, pp. 21“3.
On this point, see KCJ, ch. 3, pp. 44“52.

ordered. All of these operations of bringing together and ordering
(which I list here in a regressive order, from the derivative to the
primary): (1) bringing together individual things for comparison,
(2) generating the representation of these individual things as numerically
identical and persisting through time, (3) bringing together the
manifold of space and time themselves “ all of these operations Kant
calls synthesis. For any analysis leading to concepts to take place,
synthesis must already have taken place. And given that analysis
proceeds according to the logical functions of judgment, synthesis too
must take place in such a way that what is synthesized becomes
susceptible to being brought under concepts according to the logical
functions of judgment.
This relation between analysis and synthesis, finally, provides the key
to Kant™s definition of the categories. They are, he says, ˜˜universal
representations of pure synthesis™™ or, according to the more extensive
definition of the B edition, they are ˜˜concepts of an object, by means of
which the intuition of this object is taken to be determined with respect
to one of the logical functions of judgment™™ (A78/B104; B128). This
means two things: (1) categories are concepts that guide the syntheses
of spatiotemporal manifolds toward analysis according to the logical
functions of judgment, and (2) categories are, like any other concept,
˜˜universal and reflected representations.™™ What they ˜˜universally
reflect,™™ however, are not empirical features of objects, but just those
syntheses by means of which manifolds given in (pure or empirical)
intuition become susceptible to being reflected under concepts com-
bined according to logical functions of judgment.
I said a moment ago that Kant™s table of logical functions was meant to
make available ˜˜at one glance™™ the system of elementary logical functions
necessary to generate the least empirical judgment by means of which
empirical objects are subsumed under concepts. I also suggested that the
specific role of infinite and disjunctive judgments is to relate all concept
subordination to the unified logical space within which concepts reci-
procally delimit each other™s sphere and meaning. If I am right, this
means that correspondingly, the specific synthesis corresponding to
these logical forms will be a synthesis by means of which the totality of
objects belonging to a common logical sphere is reflected under con-
cepts. The logical form of disjunctive judgment, and the corresponding
category of community, thus provide the general structure, or ordering
function, for the standpoint on the whole in the context of which any
cognitive function is performed.

I now want to show what this means by considering more closely
Kant™s exposition of the relation between logical form of disjunctive
judgment and category of community.

Disjunctive judgment and the category of community
(Gemeinschaft, Wechselwirkung)
There are two ways in which Kant might choose to characterize the form
of disjunctive judgment. He could characterize it by focusing on the
relation of concepts considered in their content, and say that a concept
A is determined, that is, specified, either by the specific mark B or by the
specific mark C “ for instance, an animal is either a human being or a
beast, a rational or a non-rational animal. Or he might characterize the
form of disjunctive judgment by focusing on the extension of concepts
and say that in a disjunctive judgment, one states that a concept A,
considered in its extension or sphere, is divided into two mutually
exclusive and exactly complementary spheres, the sphere thought
under concept AB and the sphere thought under concept AC.
Kant chooses the second description of the form of disjunctive judg-
ment, focusing on the extension of concepts. This is particularly explicit
in the Jasche Logic as well as in the Reflexionen on logic from the critical
period. There Kant pictures the disjunctive judgment ˜˜A is either B, C,
D, or E™™ by the division of a rectangular area A (representing the exten-
sion of the divided concept A) into four regions B, C, D, and E (which
respectively represent the extensions of the species of A). In a disjunctive
judgment, says Kant, any ˜˜X thought under the concept A™™ belongs to
one or the other of the divisions B, C, D, or E. He prefaces this explana-
tion by a comparison between categorical and disjunctive judgment:
In categorical judgments, X, which is contained under B, is also con-
tained under A:
In disjunctive ones X, which is contained under A, is contained either
under B or C, etc.
Thus the division in disjunctive judgments indicates the coordination
not of the parts of the whole concept, but rather of all the parts of its

In the Critique of Pure Reason, Kant draws a surprising parallel between
this logical form and the category of community: just as in a disjunctive

Jasche Logic, x29, AAix, p. 108. Cf. also Reflexion 3096, AAxvi, pp. 657“8.

judgment, the sphere of a concept (its extension) is divided into its
subordinate spheres so that these subordinate spheres are in a relation
of mutual determination while at the same time excluding each other, so
in a material whole, things mutually determine each other, or even in
one material thing or body considered as a whole, the parts are in a
relation of mutual attraction and repulsion:
In order to be assured of this agreement [between the category of
community and the form of a disjunctive judgment], one must note
that in all disjunctive judgments the sphere (the multitude [Menge] of
everything that is contained under it) is represented as a whole divided
into parts (the subordinated concepts), and, since none of these can be
contained under any other, they are thought of as coordinated with one
another, not subordinated, so that they do not determine each other
unilaterally, as in a series, but reciprocally, as in an aggregate (if one member
of the division is posited, all the rest are excluded, and vice-versa). Now a
similar connection is thought of in a whole of things, since one thing is not
subordinated, as effect, to another, as the cause of its existence, but
is rather coordinated with the other simultaneously and reciprocally
as cause with respect to the determination of the other things (e.g., in
a body, the parts of which reciprocally attract and also repel each other).
(B 112, translation modified)14

What is surprising here is that Kant appears to assimilate a logical
relation between concepts and a material relation between things: the
mutual exclusion and complementarity of spheres or extensions of con-
cepts is assimilated to the mutual determination, by attraction and repul-
sion, of material bodies or parts of material bodies.
But this cannot possibly be right. Assimilating in this way the relation
of mutually exclusive concepts in a disjunctive judgment and the
relation of things belonging to one world-whole, or of parts making up
one material thing, is prima facie precisely the kind of move Kant
rejects throughout the Critique. As he insists in the appendix to the
Transcendental Analytic, On the Amphiboly of Concepts of Reflection,
this rejection is the core of his opposition to Leibnizian rationalist meta-
physics. Leibniz™s major metaphysical mistake, according to Kant, is to
have thought that things could be distinguished and determined by
concepts alone, specified all the way down to individuals, so that the

When Kant talks about ˜˜the multitude . . . contained under a judgment™™ he presumably
means: the multitude thought under each sub-species of the divided concept (for instance,
the multitude thought under AB, and the multitude thought under AC).

latter are completely determined as infimae species, lowest specifications
of concepts. Against this view Kant maintains, in the Amphiboly, that two
drops of water, for instance, may be identical as to their concepts, namely
as to the discursive representation of their internal determinations of
shape, size, and quality, and nevertheless be numerically distinct, solely
by virtue of their position in space (A264/B320).15 Similarly, any two
surfaces may be identical to one another as to their concept, namely their
internal determinations of size and shape, and nevertheless be numeric-
ally distinguished by their position in space as a whole. Now, it seems that
the parallel Kant draws, in the Metaphysical Deduction of the
Categories, between the logical relation of mutually exclusive and com-
plementary concepts in disjunctive judgment on the one hand, and the
relation of things expressed in the category of community on the other
hand, is just the Leibnizian error Kant denounced in the Amphiboly
chapter. This impression is only enhanced by the fact that in the text
quoted earlier, Kant describes the reciprocal action between parts of
things in terms of attraction and repulsion, namely in terms of precisely
the kind of external relation that he insists is quite distinct from the
relation of internal determinations expressed in a logical disjunction of
completely determined concepts, as Leibniz would have it (cf. A265“6/
B321, A274/B330).16 This being so, the skepticism or even derision
frequently directed at Kant™s claim concerning the parallel between the
logical form of disjunctive judgment and the category of community
seems to be a very healthy one indeed by the terms of Kant™s own
doctrine. For if this parallel displays the very confusion Kant himself

Cf. Leibniz, Nouveaux essais sur l™entendement humain, ii, ch. 27, x3. Engl. trans. and ed. Peter
Remnant and Jonathan Bennett New Essays on Human Understanding (New York:
Cambridge University Press, 1981).
One may argue on Kant™s behalf that he explains the form of disjunctive judgment in terms
of the division of the sphere or extension of a concept into its subspheres, which is the
division of a whole into its parts and thus grounds the parallelism with the division of a
whole of physical things into its parts, or even the division of one physical thing into its
parts (category of community). This is correct as far as it goes, but it is not sufficient to
alleviate the charge of amphiboly. First, it remains that if things are represented as the
ultimate parts of the sphere of a concept, then they are individualized as ultimae species,
lowest specification of a concept, instead of being, as Kant claims they should be, individ-
uated (represented as numerically distinct) by virtue of their position in space and time as
forms of sensible intuition. Second, Kant invariably presents the category of community as
a concept of the universal interaction of empirical things. We need more than a consid-
eration of concepts according to their extension to explain how such an interaction might
relate to the community of concepts under a higher concept, and thus clear Kant of the
suspicion of amphiboly. And indeed, Kant does provide us with more justification than
this, as I show below. See also KCJ, pp. 436“53.

denounces in the Amphiboly, there is every reason for discounting this
particular correspondence between logical form and category.
However, I want to argue that this suspicion, despite its seeming
plausibility, is unwarranted. Kant™s point is not that relations of things
in space (the a priori form of external sense) are essentially the same as
relations of concepts in logical space. If we follow the general thrust of his
metaphysical deduction of the categories, we should understand his
point as being, rather, that the same act of the mind which, by means
of analysis, generates the form of disjunctive judgment and eventually,
the form of a unified system of such judgments, also generates, by means
of the synthesis of spatiotemporal manifolds, the representation of a
community of interacting things or parts of things “ ˜˜for instance™™
(B112 quoted above) the relations of reciprocal attraction and repulsion
of parts in a material body. And indeed, this is what Kant writes:
The same procedure of the understanding when it represents to itself the
sphere of a divided concept, it also observes in thinking of a thing as
divisible; and just as in the first case the members of the division exclude
each other, and yet are connected in one sphere, so in the latter case the
understanding represents to itself the parts of the latter as being such
that existence pertains to each of them (as substances) exclusively of the
others, even while they are combined together in one whole. (B113,
emphasis mine; translation modified)

Note here how systematic the correspondence is. Just as the under-
standing represents to itself the subspheres (the extensions of the sub-
concepts) of a divided concept as excluding one another (if one of the
specifications is asserted of the divided concept, the others are
excluded), so it represents to itself the existence of an individual sub-
stance as excluding the existence of all others (where one exists, no other
can exist at the same time). Just as the subconcepts are represented as
combined together in one whole, so the things or parts of things are
represented as constituting one material whole.17 However, this simi-
larity in the relations represented by the understanding should not lead
us to forget “ on pain of amphiboly “ the dissimilarity between the two
cases: the individuation of things in space cannot be represented by way
of the specification of concepts. What we want to know, then, is how this
individuation is represented by the understanding. Kant™s answer,

I am grateful to Steve Engstrom for pressing me on this point and bringing to my attention
the full measure of the structural similarities Kant underlines here.

according to the Metaphysical Deduction of the Categories cited in the
first section of this chapter, is that individuation of things in space is
represented by way of the acts of synthesis that are necessary if any
analysis of the sensible given into concepts is to be possible.
I intend to show that Kant™s argument in the Third Analogy is meant
to lay out just those acts of synthesis by way of which things are indivi-
duated in space and time. According to Kant, those acts of synthesis are
acts by means of which things are represented as being in relations of
universal causal interaction. Only insofar as they are so individuated can
they also be thought under concepts of natural kinds (namely, under a
universal scale of genera and species) ordered according to the form of
disjunctive judgment and a system of such judgments.
If this is correct, one can perhaps complete Kant™s elliptic statement in
the passages just cited by saying the following. For a Leibnizian, the
similarity between the understanding™s representation of the mutual
relation of disjunctive spheres of a divided concept on the one hand,
and the mutual relation of things or parts of things in space on the other
hand, goes all the way down: individual things just are ultimate specifi-
cation of concepts. For Kant, by contrast, although there is indeed the
systematic similarity described above between the understanding™s
representation of the two relations (between concepts, between empiri-
cally given things in space), one of them (the relation of concepts)
is thought by way of analysis (of the sensible given into concepts; and
of concepts into higher concepts); the other (the relation of things)
is represented by way of synthesis of manifolds in space and time, a
synthesis that results in presenting things as individuated in space by
their relations of universal interaction.
Here again the Amphiboly chapter may help us clarify Kant™s view, no
longer as a warning against possible amphibological interpretations of
his point, but rather as a confirmation of the positive account I just gave
of the correspondence between the logical disjunction of concepts and
the category of community. Kant explains, in the Amphiboly, his oppos-
ition to Leibniz™s view according to which substances are individuated
by their intrinsic determinations (determinations they have on their
own, independently of any external relation to other substances).
According to Kant, on the contrary, substances, i.e. material things
whose essential properties persist while their accidental properties
change,18 are recognized under concepts of external relations (mutual

On Kant™s concept of substance, see ch. 2 in this volume, pp. 53“4.

causal determination). This means, then, that the move from recogniz-
ing things as individuated in space and time, to thinking them under
concepts of natural kinds, is a move from representing them in relations
of universal mutual interaction, to thinking them under concepts of
relational properties (cf. A274“5/B330“1, A283“4/B339“40).
Let me summarize my argument so far: it might seem that in relating
the category of community, or universal interaction, to the logical form
of disjunctive judgment, Kant is guilty of the very amphiboly that he
denounces in Leibniz (confusion between the relation of mutual deter-
mination between spheres of concepts, and the relation of mutual causal
determination between things). However, I argue that Kant is not guilty
of this confusion. Rather, Kant™s point is that the concepts of natural
kinds under which we know material things in nature (and thus, classify
them under hierarchies of genera and species according to the form of
disjunctive judgment) are concepts of relational properties “ univer-
sal causal interaction. This being so, the category of community
(Gemeinschaft), by virtue of which things are thought as belonging
under one logical space of concepts, is also a category of universal causal
interaction (durchgangige Wechselwirkung), by way of which they are
thought as universally related in one empirical space (and time).
To examine Kant™s argument for this point, I now turn to the Third
Analogy of Experience.

Kant™s proof of the Third Analogy: simultaneity
and universal interaction
In the Third Analogy of Experience, Kant argues that our experiencing
the simultaneous existence of appearances is sufficient to attest that
these appearances are in relations of thoroughgoing community
(Gemeinschaft) or interaction (Wechselwirkung). This is because, Kant
argues, representing the simultaneous existence of appearances is our
doing, and this representation is possible only if we represent appear-
ances as being, with respect to one another, in relations of universal
interaction. Thus the statement of the Analogy: ˜˜All substances, insofar
as they can be perceived in space as simultaneous, are in thoroughgoing
interaction [in durchgangiger Wechselwirkung]™™ (B256).19

In the first edition, the Analogy is stated as follows: ˜˜Principle of community. All sub-
stances, insofar as they are simultaneous, stand in thoroughgoing community (i.e. inter-
action with one another).™™ The formulation in B is superior in that it makes clearer that
˜˜simultaneous™™ means: ˜˜something we represent, or perceive, as simultaneous.™™ Similarly,

As any careful reader of Kant™s Analogies of Experience knows, the
three Analogies should be read together as one argument, which concerns
the conditions of our representation of objective time-determinations.
Kant™s question is: how do we come to have any representation at all of
objective temporal determinations of appearances, since our apprehen-
sion of them is always successive, and since we have no given temporal
framework that might allow us to locate events and states of affairs in
time? In the Second Analogy, Kant explains how the subjective succession
of perceptions in apprehension can be the experience of an objective
succession of states of things; in the Third Analogy, he explains how
the subjective succession of perceptions in apprehension can be the
experience of an objective simultaneity of things in particular states.
Prior to this, in the First Analogy he has argued that any representation
of objective temporal order (succession or simultaneity) rests on the
presupposition of something permanent, as the substrate of objective
temporal determinations. I do not propose here to evaluate Kant™s over-
all argument in the Analogies of Experience.20 What I am mainly
concerned with is how discursive forms (forms of analysis or reflection)
and forms of sensible synthesis relate, according to Kant, in the particular
case of the Third Analogy.
Kant™s reasoning proceeds, roughly, according to the following

the argument in B is more clearly laid out as an argument about the conditions for our
experiencing things as simultaneous. One may wonder how such conditions put any
constraint at all on how things actually are. But the Transcendental Deduction is supposed
to have established just this point: the conditions of possibility of experience are the
conditions of possibility of the objects of experience. Evaluating the argument of the
Deduction is of course beyond the scope of this chapter. One should at least remember
one essential aspect of its conclusion: the objects we are talking about here are objects as
appearances “ as individuated in space and in time, the forms of our sensible intuition.
For an analysis and evaluation of Kant™s Analogies of Experience, see KCJ, ch. 11. On the
Second Analogy, see ch. 6 in this volume.
There are two expositions of the argument in the Third Analogy. The first in A, remain-
ing unchanged in B: A211/B258“A213/B260. The second added in B: B256“8. In my
view, the exposition in B is the clearer of the two, for reasons similar to the ones
I advocated in the previous footnote: the argument in B, just as the formulation of the
Analogy itself, makes it clearer that what Kant is talking about are the conditions for our
experience of objective simultaneity (which is also the only context in which the very
notion of simultaneity has any meaning at all). In my reconstruction of the argument
I will thus follow the order of the B edition. In an effort to limit the length of the
footnotes, I shall indicate the textual support for each step simply by the reference in B
(i.e. the 1787 version) and A/B (when the 1781 version provides useful additional textual
support). I shall not quote the texts themselves.

1 The synthesis of our apprehension in imagination is always
2 We nevertheless experience a subjective succession in apprehension
as an objective simultaneity of things in particular states if, and only
if, we experience the subjective succession as being order-indifferent.
For example, we are conscious that we could direct our gaze indif-
ferently from the moon to the earth or from the earth to the moon;
it is in this way that, even though we might never perceive at the
very same time the moon at its zenith and the surface of the earth,
we do experience these objects as simultaneously existing (B257;
3 We have no perception of time itself that would allow us to derive from
the simultaneity of objective states of things the order-indifference of
the subjective succession in our apprehension of these states
4 Nor would the mere subjective succession of perceptions in our
apprehension suffice to generate either the representation of its own
order-indifference or the interpretation (experience) of this order-
indifference as objective simultaneity. Subjective succession in appre-
hension would, by itself, give us only: one perception, then the other,
and reciprocally, the latter, then the former. It would give us no access
to the simultaneity of things as the necessary condition for the order-
indifference of the perceptions (B257).
5 We are conscious of the subjective succession as order-indifferent, and
thus as a representation of objective simultaneity if, and only if, in
relating the subjective succession of perceptions in apprehension to
objects, we form judgments such as: if object X (recognized under
concept A) exists at time t at point p1, then object Y (recognized under
concept B) exists at that same time at point p2, and reciprocally, if the
latter exists, then the former exists at the same time. We thus think X
and Y as being in themselves determined with respect to the logical
form of a hypothetical judgment whose reciprocal converse is also
thought to be true (if X, recognized under A, exists at p1 at t, then Y,
recognized under B, exists at p2 at t; and conversely if Y recognized
under B at p2 at t, then X under A at p1 at t). Thus a pure concept of

This premise is not explicitly stated in the argument of the B edition, but it is common to all
three Analogies, and explicitly stated in the first and second: see A182/B225 (First
Analogy), A189/B234, A198/B242 (Second Analogy); in the Third Analogy, this premise
is implicit at B257.

the understanding is applied whenever we experience objective
simultaneity (B257).23
6 This concept is that of mutual conditioning, i.e. interaction. Thus the
coexistence of things in space can be experienced only under the
presupposition that they are in relations of universal interaction or
community (B257“58; also A212“13/B259“60).
7 So, all appearances, insofar as we perceive (experience) them as
coexisting, exist in relations of thoroughgoing reciprocal influence
(B258; also A213/B259“60).24

In what I present as step (5), I am making explicit that the ˜˜pure concept of the under-
standing™™ needed to represent the reciprocal sequence as objective is the ˜˜concept of an
object, by means of which its intuition is regarded as determined with regard to one of the
logical functions for judgment™™ (cf. B128), in this case the function of a hypothetical
judgment together with its reciprocal converse. Here again, as in the case of the Second
Analogy, I hope to show why it is helpful to stress this relation between the pure concept of
the understanding and the corresponding logical function of judgment. Note already that
the logical function at work here is not that of a disjunctive judgment, but that of
a hypothetical judgment (and its reciprocal converse). This is quite explicit in Kant™s
presentation of his example, that of perceiving the earth and the moon to exist
˜The synthesis of imagination in apprehension would only present each of these
perceptions as one that is present in the subject when the other is not, and conversely,
but not that the objects are simultaneous, i.e., that if the one is then the other is also at the
same time, and that this is necessary in order for the perceptions to be able to succeed
each other reciprocally (B257, emphasis mine).
Of course a disjunctive judgment might itself be translated into hypothetical judg-
ments, such as: ˜˜if the one is at a given point, then the other is not,™™ where each of the
two simultaneously existing things excludes the other from the point in space which it,
itself, occupies, just as each of the two concepts B and C dividing a higher concept A in
the disjunctive judgment: ˜˜A is either B or C™™ each exclude the other™s extension from
their own. But it is important to note that it is not this negative form that Kant
mentions in expounding his argument: what he says is that ˜˜If the one is then the
other is also at the same time.™™ This, it seems to me, expresses the relation of mutual
conditioning that would be captured by two reciprocal hypothetical judgments. Each of
the two coexisting things is thus individuated as to its existence in space by its relation
to the other (and in fact, each of the indefinitely many coexisting things is thus
individuated by its relation to all the others) and eventually reflected under concepts
that can be combined according to the form of disjunctive judgments, say for instance:
˜˜this is either outside the solar system or a satellite of the sun or a satellite of another
body within the solar system.™™
Note that here we find the same move as in the Second Analogy, from what we presuppose,
to what is true of objects (as appearances). I suggest that the move is (implicitly) justified
here in the same way as it was there: by referring back to the argument of the
Transcendental Deduction to the effect that ˜˜the conditions of possibility of experience
are the conditions of possibility of the object of experience™™ (cf. above, ch. 6, p. 159).
I will not dwell on this point. What interests me about the third Analogy is more specifically

Now, this conclusion is prima facie completely implausible.It is simply
not true, one might object, that I perceive my desk and my chair as
simultaneously existing only if I suppose a relation of interaction
between them, and it is also not true that I perceive the earth and the
moon as coexisting only if I suppose reciprocal influence between them.
The objection seems only too obvious. However, it might be overcome if
we remember that there is originally nothing more to the pure concept
of cause than ˜˜the concept of an object, by means of which its intuition is
regarded as determined with regard to . . . the logical function of a
hypothetical judgment™™ (B128). Thus by ˜˜interaction™™ (namely recipro-
cal causal action), Kant means nothing other than the relation between
the states of one (relatively permanent) substance and the states
of another, such that they can be regarded as determined with regard
to the logical function of a hypothetical judgment whose reciprocal
converse (the consequent taking the place of the antecedent, and conver-
sely) is also taken to be true. What Kant is saying is that interpreting two
successively apprehended states, say A and B, as simultaneously existing
states of objects, is thinking something like this: ˜˜If X (recognized under
concept A) is part of the present whole of my experience, then Y (recog-
nized under concept B) is part of the same whole. And if Y (recognized


. 6
( 10)