Timothy Rosenkoetter

The Logical Home of Kant’s Table

of Functions
Abstract. In order to evaluate Kant’s Table of the Functions of the Understanding
in Judging, we need to know to which logic it belongs. Considerable textual evi-
dence speaks for each of two incompatible candidates: pure general logic and
transcendental logic. This paper solves this puzzle by showing that Kant uses
“pure general logic” to refer to two genuinely distinct logics. Each is general rather
than transcendental by virtue of treating all uses of the understanding. They are
distinguished most fundamentally by the different perspectives that they take on
that capacity, with the logic of the Table specifying the basic operations that the
understanding uses in cognizing objects by means of intuitions.

Um Kants Tafel der Funktionen des Verstandes in Urteilen zu evaluieren, müssen wir
wissen, zu welcher Logik sie gehört. Textbelege gibt es für zwei inkompatible Kan-
didaten, die reine allgemeine Logik und die transzendentale Logik. Dieser Aufsatz
löst dieses Rätsel, indem gezeigt wird, dass Kant mit „reine allgemeine Logik“ auf
zwei genuin zu unterscheidende Logiken Bezug nimmt. Jede ist allgemein, aber nicht
transzendental, weil sie alle Fälle von Verstandesgebrauch behandelt. Sie sind
fundamental unterschieden durch die verschiedenen Perspektiven, die sie zu diesem
Vermögen einnehmen; die Logik der Tafel spezifiziert die basalen Operationen, die
der Verstand vornimmt, wenn er Objekte mittels Anschauungen erkennt.

1 The Question
The argument of the Metaphysical Deduction begins with one table and ends
with another. More specifically, it begins with a list of the basic “functions”
that the understanding uses in judgment and moves to the Table of Categories,
whose centrality to Kant’s overall project can hardly be overstated.¹ According to
a nearly universal consensus, the starting point of this argument is both (i) entire-

 I will be using “Metaphysical Deduction” as a rough label and do not wish to take a position
on whether it actually begins with argumentation supporting the first Table. Throughout the
paper italics will be used when introducing terms, as well as for standard reasons of emphasis.
I use translations from The Cambridge Edition of the Works of Immanuel Kant as my starting
point, though I modify them when necessary without noting this in each instance.

DOI 10.1515/9783110521047-002

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
30 Timothy Rosenkoetter

ly formal, and (ii) distinctively logical. Kant seems to provide straightforward

confirmation of the former in this short Preamble to the Table of Functions:

If we abstract from all content of a judgment in general, and attend only to the mere form of
the understanding in it, we find that the function of thinking in the same [in judgment,
T. R.] can be brought under four titles, each of which contains under itself three moments.
(A 70/B 95)

As to (ii), the distinctively logical nature of the Table of Functions follows, first,
from its formality, since logic is the only body of a priori cognitions that is purely
formal.² In addition, Kant’s texts confirm the logical nature of this starting point
in any number of ways, beginning with the fact that he calls the functions “log-
ical functions” and labels their table the “Logical Table of Judgments”.³ Kant
does not call them “epistemic functions”, and they are not gathered together
in a “Table of Knowledge” or “Table of Inquiry”. This is one reason why critical
discussion surrounding the Table has often centered on which logical constants
are necessary – or at least necessary for anything that Kant would have recog-
nized as a logic.⁴
For anyone familiar with the multiplicity of Kantian logics, the crucial next
question is “Which logic?”. Since the Table of Functions is pure and so cannot
belong to an applied logic, Kant’s taxonomy of logics would seem to leave just
two candidates: pure general logic or transcendental logic.⁵ But since transcen-
dental logic (TL) is partially constituted by the fact that it does not “abstract from
all content of cognition” (A 55/B 80), the home of the purely formal Table⁶ seems
obvious. It must belong to pure general logic (PGL). This PGL-interpretation has
long been the nearly universal consensus among commentators.⁷ My contention

 Cf. GMS 387.

 A 70/B 95; Prol 302.
 Critics typically assume that the Table has been superseded by the Fregean revolution in
logic. M. Wolff’s (1995) important defense of Kant against that charge shares the critics’ assump-
tion that “for Kant the Table of Judgments is a ‘logical’ matter” (p. 3, my translation), an assump-
tion that I will challenge below. Wolff’s defense instead challenges their assumption that Kant
would have made Fregean logic part of pure general logic, had he been aware of it (p. 202).
 Cf. A 52–56/B 76–78.
 I will normally refer to the Table of Functions simply as the “Table”, adding its full title only
when there is some risk of confusion. In contrast, all references to the Table of Categories will be
spelled out.
 Adherents include: Allison 2004, p. 137; Beck 1992, p. xiv; Bennett 1966, pp. 79–83; Bird 2006,
pp. 259–268; Forster 2008, pp. 71–75; Guyer 2010, p. 127; Hanna 2001, pp. 76–78; Kemp Smith
1992, p. 178; Paton 1936, p. 245; Prichard 1909, pp. 152–156; Strawson 1966, pp. 74–82; Tolley
2012, p. 425; Waxman 2014, pp. 266–277; M. Wolff 1995, p. 3; and R. Wolff 1963, p. 62.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 31

is that we are bound to misunderstand the Metaphysical Deduction so long as we

locate the Table within PGL. The next section will survey evidence for the PGL-
interpretation and sketch two other natural readings that appear in the litera-
ture. Then I will present an alternative that has remained unconsidered, most
likely because it at first appears to require a radical departure from Kant. My
goal in this paper is to show that this initially improbable alternative has the
best claim to be Kant’s considered position.

2 The Problematic Texts and an Alternative

The main reason for the PGL-interpretation’s long-standing dominance is the
broad support that can be found for it in centrally located texts. It will be useful
to distinguish three ways in which these texts provide support:
(α) “General”. First, there are passages that characterize the Table’s func-
tions as “general,” thereby suggesting that they belong to general logic, and
thus to pure general logic. Here is one relatively discrete example:

In the metaphysical deduction the origin of the a priori categories in general was establish-
ed [dargetan] through their complete coincidence with the general logical functions of
thinking […]. (§ 26, B 159)

The most significant (α)-passage spans the entire first half of § 10, just after the
Critique has presented and explained the Table of Functions in § 9. For now, what
is important about this extended text is just that it contrasts TL as the home of
the Table of Categories with a logic that it associates with the Table of Functions
and repeatedly calls “general logic”.⁸
(β) “Form”/“Content”. Second, there are a number of passages in which we
are told that the Table “abstracts” from “the content of judgments” (or some-
times “the content of cognition”) and, correlatively, attends merely to “form.”⁹
These passages are important because they are strikingly similar to descriptions
that Kant provides of PGL. Most importantly, the Critique introduces TL by point-
ing to the possibility of “a logic in which one did not abstract from all content of
cognition”, offering this in explicit contrast to PGL, the logic that abstracts from

 A 76–79/B 102–105. To the list of (α)-passages we can add the opening of the Analytic of Prin-
ciples (A 130–136/B 169–175), which in relevant respects is similar to § 10.
 Aside from the Preamble (A 70/B 95), the most important (β)-passages are: A 76/B 102, A 79/
B 105, A 299/B 355–356, A 321/B 377–378, and MAN 475.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
32 Timothy Rosenkoetter

“all content of cognition” and considers only “the form of thinking in general”
(A 55/B 79).
(χ) “Logical”/“Transcendental”. Third, there are passages in which Kant
contrasts the Table of Functions and the Table of Categories as “logical” versus
“transcendental”. This occurs already in the Critique, whose section containing
the first Table is labeled “On the logical function of the understanding in judg-
ments”, while § 10’s preparation for the Table of Categories makes clear that
with it we have entered TL.¹⁰ While the Critique makes this contrast easy to over-
look, it is impossible to miss in the Prolegomena, whose two tables are presented
in immediate succession and labeled: “Logical Table of Judgments” and “Tran-
scendental Table of Concepts of Understanding” (4:302). This is important in the
first place because Kant often uses “logic” and “logical” as shorthand for PGL.¹¹
More significant, though, is that if Kant in fact considers the Table of Functions
to be a part of TL – apparently the only other possibility – his exposition is pos-
itively irresponsible, and consistently so. After all, he uses the term “transcen-
dental” to characterize the Table of Categories. If the Table of Functions likewise
belongs to TL, why not simply call it the “Transcendental Logical Table of Judg-
ments” at least once?¹²
Beyond this textual support, several substantive considerations likely ex-
plain the PGL-interpretation’s dominance in the literature. Prime among these
is the sense that it would somehow be circular for Kant to use TL itself as a start-
ing point to derive the categories.¹³ Though a full explication of this worry has (to
the best of my knowledge) yet to be given, in some versions it is buttressed by the
belief that Kant’s compilation of the Table is a distinctively logical undertaking.
Since TL is properly concerned with epistemological questions that extend be-
yond the merely logical, the news that the Table belongs to TL would rob the Met-
aphysical Deduction of what is taken to be one of its notable features, viz., that it
uses facts about logic to reach a conclusion that has epistemological and onto-
logical significance.¹⁴

 A 70/B 95, italics added; cf. A 78–79/B 104–105.

 E. g., B vii; GMS 387.
 For further examples of (χ), see A 299/B 355–356.
 Allison explains that “any appeal to transcendental logic in order to derive the moments of
judgment in general logic would be viciously circular, since the whole purpose of the Metaphys-
ical Deduction is to derive the categories of the former from the logical functions of judgment
specified (supposedly on independent grounds) in the latter” (2004, p. 140). I can see no reason
why the grounds supporting the Table, were it to belong to TL, could not nonetheless be inde-
pendent in the sense relevant to circularity.
 For instance, Kemp Smith tells us that “formal logic […] can supply a criterion for the clas-
sification of the ultimate forms of judgment just because its task is relatively simple, and is in-

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 33

Even if one doubts that straying from this argument strategy would suffice
to render the Metaphysical Deduction circular, it is nonetheless easy to see, sec-
ond, how the PGL-interpretation brings with it the promise that a candidate list
of categories might be justified to a level of credence that would be otherwise
elusive. The thought might take this shape: whereas proposals within TL regard-
ing the identity of the categories are likely to be just as controversial as most
philosophical theses, it should be possible at least in principle to obtain wide-
spread agreement concerning a logic that is “finished and complete” in all es-
sential respects and allows “the different co-workers to achieve unanimity”
(B vii – viii). The person making this point would likely go on to concede that
Kant’s subsequent realization of the Table does not in fact live up to this prom-
ise. Yet the fact that an epistemologically sound footing for the project of iden-
tifying the categories is in principle available by recourse to PGL would at
least appear to explain Kant’s argument strategy in the Metaphysical Deduction.
For all of the evidence which points toward the PGL-interpretation, several
recalcitrant texts are difficult to square with anything but the competing TL-in-
terpretation. ¹⁵ They center on what we can call the Appendices to the Table,
which Kant includes so as to avert misunderstandings, since the Table “departs
from the customary technique of the logicians” (A 70–71/B 96). Though I will
argue below that there are other, less obvious departures, Kant makes two
easy to spot. Namely, two functions that the Table recognizes as separate and
irreducible are said to enjoy no special status within PGL: the singular function,
which is actualized when judging “This S is P”, and the infinite function (“S is
not-P”).¹⁶ I will sometimes call them the extraordinary functions, thereby con-
trasting them with the ten remaining ordinary functions.
The relevant passages present modest interpretative challenges. Appendix 1
does not explicitly tie the Table to TL. Moreover, it attributes disregard of the sin-
gular form’s special status to “that logic that is limited only to the use of judg-
ments with respect to each other” (A 71/B 96–97). Allison has argued that this

dependent of all epistemological views as to the nature, scope, and conditions of the thought
process” (1992, p. 185).
 Passages (α) – (χ) make outright support for the TL-interpretation – as opposed to holding
that there is “something transcendental” about the Table – relatively rare. Greenberg 2001 is
an exception.
 Though this paper cannot focus on the function-form distinction, I disagree with interpreters
such as M. Wolff who suggest that it is the key to resolving significant apparent contradictions
(1995, pp. 28–32). I take functions to be act-types of synthesis. When these act-types take con-
cepts as inputs, they yield judgments. In some cases the resulting judgment instantiates the cor-
responding logical form. However, for reasons that will become clear in § 5, we should not as-
sume that which function was used can always be determined simply by consulting logical form.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
34 Timothy Rosenkoetter

is not a reference to PGL but rather to “syllogistics”, which he takes to be no

more than a part of PGL (2004, p. 141). However, the beginning of Appendix 2
would seem to provide retroactive confirmation that it is indeed TL that recogniz-
es the singular function and none other than PGL that disregards that function’s
special contribution:

Likewise, in a transcendental logic infinite judgments must also be distinguished from af-
firmative ones, even though in general logic they are rightly included with the latter and
do not constitute a special member of the classification. General logic abstracts from all
content of the predicate (even if it is negative), and considers only whether it is attributed
to the subject or opposed to it. Transcendental logic, however, also considers the value or
content of the logical affirmation made in a judgment by means of a merely negative pred-
icate, and what sort of gain this yields for the whole of cognition. (A 71–72/B 97)

Shortly thereafter we are told that the distinction between affirmative and infin-
ite judgments “may not be passed over in a transcendental table of all the mo-
ments of thought” (A 73/B 98). In sum, if we could just ignore the textual evi-
dence provided by (α) – (χ) and restrict our focus to these two Appendices, it
would be wholly uncontroversial that the Table belongs to TL. Moreover, the
clear message of the first two Appendices is that if the Table had belonged to
PGL, there would have been just ten categories instead of twelve. One could
scarcely wish for a more perspicuous refutation of the PGL-interpretation.
All of this leaves the charitable interpreter in a difficult position. First, the
alert reader will note that Appendix 2 also stands in apparent contradiction to
(β)-passages such as the Preamble, which had told us that the Table abstracts
“from all content of a judgment in general [überhaupt]” (A 70/B 95). Now just
two pages later, and in an avowed effort to avoid misunderstandings, Kant ex-
plains that the distinction between affirmative and infinite judgments is part
of the Table precisely because it does consider “content”, viz., the content of
the predicate-concept (A 72/B 97). Perhaps it can be shown that the content of
the predicate-concept is not an instance of the “content of a judgment in gener-
al” (A 70/B 95).¹⁷ But even if this is successful, it seems like a merely local fix to a
more pervasive problem. It would appear, namely, that Kant simply avoids
choosing between PGL and TL and locates the Table in both.

 This is in fact the position that will emerge. To anticipate, we will see that the categories are
the “content of a judgment in general [überhaupt]”, since they are the concepts whose content
has its origin in the very nature of judgment. Of course, the Table abstracts from these concepts,
for it is nothing more than a table of functions (§ 3). Yet Kant can maintain this and still hold that
which function a subject is using is sometimes partially determined by the content of a predi-
cate-term (§ 6).

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 35

Some interpreters have made this their final take on the matter. Kemp Smith,
for one, presents Kant as admitting “in the frankest and most explicit manner” in
the Appendices that the Table is not, in fact, exclusively formal, though that is
what he has announced as well as what his project demands (1992, p. 192). Ac-
cording to a more charitable proposal, which I will call inclusivism, Kant believes
that no simple answer to the question “To which logic does the Table belong?” is
available. It might, for instance, be that though the Table does in some sense le-
gitimately belong to PGL, the distinctive contributions of the two extraordinary
functions are recognizable only from the perspective of TL.¹⁸ If this is the case,
then there is something wrong with the very expectation that Kant’s answer
will take the simple form: “The Table belongs to logic x.”
My aim in this paper is to argue that an answer of just this simple form is
both available and well supported by Kant’s texts. The root cause of the confu-
sion regarding the Table’s home, I claim, is that Kant uses a single term, “general
logic” (as well as the fully explicit “pure general logic” and his standard short-
hand “logic”), for both PGL and a separate logic that he never explicitly distin-
guishes from PGL. I will call the unnamed logic the logic of cognition (LC).
A logic, in Kant’s broad sense of the term, is a theory that considers the use
of a particular capacity, the understanding.¹⁹ Yet no logic considers all aspects
of thought. Hence, any logic is constituted, for Kant, by a perspective that deter-
mines which aspects are relevant and which must be ignored, if one is to remain
within that logic. This point will be used below to argue against inclusivism,
which in effect helps itself to the idea that one and the same logic can be regard-
ed from different perspectives, all the while maintaining its identity. My contrary
suggestion will be that each perspective, so long as it succeeds in capturing gen-
uine features of thought, defines its own logic. This sets us up to notice that there
are two distinct perspectives, each of which is sometimes intended (to the exclu-
sion of the other) when Kant uses the term “general logic”. The constitutive per-
spective of PGL dictates that it focuses on the relations that thoughts bear to
other thoughts, while abstracting from how these thoughts relate to intuitions.
It is this focus that makes PGL distinctively logical. In contrast, the constitutive
perspective of LC dictates that it focuses on the relations that thoughts bear to
objects. LC builds upon an analysis of our capacity to cognize objects. It will

 This is normally taken to render the Appendices fully consistent with the PGL-interpretation.
Proponents include: Allison 2004, p. 141, Krüger 1968, pp. 347–348, and Lu-Adler 2014. Cf. Brandt
1991, pp. 72–74.
 Cf. A 51–52/B 75–76. See KU 176 for a particularly clear instance of Kant distinguishing the
study of capacities (in this passage, “critique”) from the study of objects (“doctrine”). Logic is
one form that critique takes. Their relation is complex, but this will suffice for present purposes.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
36 Timothy Rosenkoetter

prove crucial that this capacity for cognition cannot be actualized without the
contribution of intuition.²⁰
Though LC takes thought’s dependence on intuition into account, it still
qualifies as a “logic” in Kant’s broad sense of the term, since it attends specially
to the contributions that thought makes to cognition. Yet this is, understandably
enough, not what commentators typically have in mind when they take the Met-
aphysical Deduction to argue from a “logical” starting point.²¹ And it is of course
worlds away from PGL, whose constitutive perspective requires that it abstracts
from thought’s relation to anything except thought, and thus from thought’s ul-
timate reliance on intuition. Below we will see that Kant positively intends for
the Table to take “content” that is provided by intuition into account when clas-
sifying acts of judgment. So not only is the Table not (ii) distinctively logical in
the sense usually assumed. Neither is it (i) purely formal.
Here’s the plan for the rest of the paper. The important task for § 3 below is to
show that the passages that are almost single-handedly responsible for the dom-
inance of the PGL-interpretation are in fact making a subtly different point. Kant
is instead telling us that the Table is, in a sense to be worked out, an object-free
logic. As we will also see in § 3, it follows that the Table does not belong to TL,
since as a particular logic TL is precisely not object-free. Now, PGL is an object-
free logic and thereby a candidate to be the home of the Table. Yet PGL need not
be the sole member of that genus. § 4 argues that the genus also includes LC,
while providing a diagnosis of why Kant makes do with a genus-term (“pure gen-
eral logic”) that fails to distinguish between PGL and LC. § 5 takes up two pro-
posals that promise all of the advantages of the LC-interpretation without requir-
ing that we recognize a logic distinct from PGL and TL. Here we will see that
though Kant attaches great importance to his claim that the Table is a system,
at key points he correctly treats PGL as though it is not a system. This gives us
a compelling reason to expect that Kant himself would have agreed to disambig-
uate the genus-term “pure general logic”, if only he had witnessed the confusion
that his more relaxed nomenclature has invited. My short final section (§ 6)
makes a start at using the LC-interpretation to explain the Table’s inclusion of

 Hence, these objects cannot simply be other thoughts (cf. A 108). Nor does the “x of judg-
ment” that is implied by any thought that relates concepts to one another (apart from their con-
nections to intuition) qualify (cf. Longuenesse 1998a, p. 88). Unless otherwise noted I will use
“cognition” in the robust sense, in which it contrasts with mere thought (cf. B xxvi).
 The most significant exception is Longuenesse: “Kant asked himself which logical forms of
judgment should be considered primitive if the original function of judgment is […] to relate our
representations to objects” (1998a, p. 78; cf. 2006, p. 144). This paper can be read as a working
out of this important thought.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 37

the infinite function. We will see that, contrary to what is nearly always assumed,
there are differences between PGL and the Table not only with respect to the in-
finite function but also with respect to the affirmative and negative functions.
This suggests that future work would do well to scrutinize the very natural as-
sumption that the ten apparently ordinary functions can be adequately under-
stood from the perspective of PGL. It might just be that only LC equips us to un-
derstand all of the Table’s functions.

3 General versus Transcendental Logic

It is a familiar point that what Kant means by “logic” is different from what
we mean by the same term.²² Rather than the study of inferences that are
valid by virtue of their form, logic for Kant is a study of the subject – in partic-
ular, the study of the use of a certain capacity of the subject. This conception
leads Kant to recognize logics that are difficult to distinguish from sciences
that study particular types of objects. All of the examples of this phenomenon
that matter for our purposes are particular [besondere] logics, which contain
“the rules for thinking correctly about a certain kind of object” (A 52/B 76).²³
Let’s begin with an example of a commonplace particular logic such as juridical
logic, which is defined by its attention to legal objects. It qualifies as a logic be-
cause it contains rules for our thought about contracts and the like.²⁴ Yet it is evi-
dent that at least some of these rules will amount to truths about the objects
themselves (contracts), even if they are couched as truths about our thought
about those objects. My claim is that Kant takes logicians who are pursuing par-
ticular logics to have objects in view. Moreover, it is not surprising that Kant
would adopt this understanding, given the absence of a clear boundary between
(e. g.) juridical logic and the study of legal objects.
Why is this relevant? First, we will see that Kant conceives of TL, no less than
juridical logic, as a particular logic with a certain kind of object in view. Second,

 Cf. Brandt 1991, p. 53, Longuenesse 1998a, p. 74, and M. Wolff 1995, p. 19.
 A different example is applied logic, which is simply (a branch of) empirical psychology, as
Kant acknowledges (A 53/B 77). Our capacity of thought, when studied empirically, is itself one
of the many types of objects that doctrines can study.
 I use juridical logic as my stock example because it may well be a particular logic that Kant
actually envisaged (cf. M. Wolff 1995, p. 210). No importance should be attached to the fact that
legal objects are different in kind from ordinary objects such as pomegranates. The particular
logic of pomology would serve equally well as an example, save for the fact that there is no
sign that Kant took that particular logic seriously.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
38 Timothy Rosenkoetter

Kant does not treat general logics (the contrast concept to particular logics) as
having objects in view in the same way. General logics are object-free. This sec-
tion will work at specifying this claim and showing how it enables us to make
sense of passages (α) – (χ). I will approach the contrast between general and par-
ticular logics from two directions, each of which will help us to understand a dif-
ferent set of passages.
I. Scope-difference. General logics do not pick out a set of privileged ob-
jects and focus on them, while particular logics do precisely that. The logic of
the Table is “general” in the sense that its distinctions apply to the thought of
anything whatsoever. For instance, the distinction between the affirmative “S
is P” and the infinite “S is not-P” is not dependent on the substitution of any par-
ticular concepts for “S” or “P”. This is the primary reason that Kant calls the
Table and its functions “general” in (α)-passages.²⁵
The very same message can instead be conveyed using the term “content”,
as we see in this description of PGL: “[A]s general logic it abstracts from all con-
tent of cognition of the understanding and the difference in its objects” (A 54/
B 78). To understand this passage we need to know that here Kant is using “con-
tent” much as we now use the term “intensional content”.²⁶ PGL abstracts from
the meanings of the concepts that it treats, such that its contributions are most
perspicuously represented if variables are used. Yet if a logic is to treat juridical
objects differently from all others, it cannot simply use variables. Juridical logic
is constituted by the fact that it privileges certain concepts such as “contract”,
while not privileging other concepts such as “pomegranate”.²⁷ In contrast, be-
cause PGL abstracts from all contents, it cannot pay heed to the “difference in
its objects”, e. g., to the difference between contracts and pomegranates. PGL at-
tends instead to formal differences between thoughts (whatever their intensional
content), prime among them the difference between contradictory and non-con-
tradictory thought. Notice, though, that the mere fact that PGL abstracts from in-
tensional contents, and therefore fails to distinguish any particular sphere of ob-
jects, does not itself determine which forms are privileged by PGL. This point will
be central to § 4.

 Most straightforwardly, § 26 (B 159). The same point is sometime made without using the
word “general”: the Table is “undetermined with respect to every object” (Prol 323).
 As he does also at Prol 266 (cf. A 65/B 90). More often, Kant uses the term “content” in re-
lated senses that will be covered below.
 Kant says little about particular logics other than TL, so it is unclear precisely how such a
logic would go about privileging these concepts. I presume that it would at least include
some rules for how to think about contracts, but no similar rules for how to think about pom-
egranates.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 39

Above it was simply asserted that TL is a particular logic. This classification

is not made evident by the Critique, so it requires defense. We must also under-
stand why TL is particular in order to comprehend Kant’s reasons for locating the
Table outside of TL. Finally, this discussion will help us to understand the sense
in which TL has objects in view and the Table does not. This is essential if we are
to make sense of (β)-passages.
If TL is indeed a particular logic, the preceding discussion suggests that
there will be concept(s) that pick out its privileged objects, thereby distinguish-
ing TL from other particular logics. These concepts are the categories. So one rea-
son that TL qualifies as particular is simply that this set of twelve concepts is not
equivalent to the set of all possible intensions.²⁸ We can note immediately that
this fits one of the broad patterns to be found in the (α)-aspects of § 10, the sec-
tion which effects a transition from “general logic” with its Table of Functions to
TL with its Table of Categories. The end point of this transition is a logic defined
by a set of concepts, which are concepts (as opposed to “functions” or “forms”)
only because they have “content” (A 79/B 105). This stands in contrast to “gen-
eral logic”, which, as we are repeatedly reminded, abstracts from all content.
This is inter alia a way of making the by now familiar point that the Table of
Functions is not constituted by its attention to particular concepts.
The most important objection to the classification of TL as a particular logic
is that the categories have all objects whatsoever as their objects. How then, it is
natural to wonder, could TL be anything but a general logic?²⁹ I propose that
Kant in fact recognizes two possible grounds for the particularity of a logic.
First, a logic can be either general or particular with reference to the set of ob-
jects of our thoughts. This is the rationale for particularity that is cited in Kant’s
initial definition quoted above (A 52/B 76), and it suffices to explain the partic-
ularity of juridical logic. Second, a logic can be either general or particular with
reference to the set of uses of the capacity for thought, rather than the objects of
these thoughts. This second rationale for particularity leaves Kant free to con-
cede that we use the categories to think about anything whatsoever, while clas-
sifying TL as particular because it treats only a particular subset of thoughts, viz.,
pure thoughts. Even if all objects whatsoever fall under the categories, thoughts
of pomegranates are not pure thoughts.

 The next paragraph will suggest that this is not the full story, as can be seen if we ask how
the present rationale can explain TL’s inclusion of the Transcendental Dialectic, which is itself
defined by three different concepts. I begin with the simpler story because it helps to clarify
some (α)-passages.
 Cf. Tolley 2014.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
40 Timothy Rosenkoetter

If we look at how Kant introduces his new logic, it reflects this second ration-
ale for particularity: “But now since there are pure as well as empirical intu-
itions […], a distinction between pure and empirical thinking of objects could
also well be found”.³⁰ This is different from the sense in which PGL is pure:
“[A]s pure logic it has no empirical principles” (A 54/B 78). Though its principles
are pure, PGL treats all thought, whether empirical or pure. We find Kant empha-
sizing just this point about PGL when preparing his introduction of TL: “A gener-
al, but pure logic therefore has to do with strictly a priori principles, and is a
canon of the understanding and reason, but only in regard to what is formal
in their use, be the content what it may (empirical or transcendental)”.³¹ TL’s pu-
rity, in contrast, is defined by its exclusive focus on pure content: “[T]hat logic
that contained merely the rules of the pure thinking of an object would exclude
all those cognitions that were of empirical content” (A 55/B 80).
II. Object-free? We have seen in outline why TL is a particular logic and why
this makes it natural for Kant to contrast the Table of Functions with TL in terms
of scope, as general versus particular. We have also seen how “content” can be
used to express the same contrast. However, the same term typically appears
within the Metaphysical Deduction in the service of a second (albeit related) con-
trast: concepts represent objects, while functions and forms do not. ³² In Kant’s ter-

 A 55/B 79–80, italics added; for parallel terminology applied specifically to reason, cf.
A 303–309/B 359–366. I believe that it is no accident that Kant often makes this point using be-
sondere (A 842/B 870, A 845/B 873, and GMS 390), including: “[…] general logic is not limited to
any particular [besondere] kind of cognition of the understanding (e. g., not to the pure cognition
of the understanding) nor to certain objects […]” (A 708/B 736, italics added). Note that the two
italicized phrases offer a choice between the two alternative rationales for particularity identi-
fied above.
 A 53/B 77, italics elided; cf. A 796/B 824. “Empirical or pure” would have been an extension-
ally equivalent formulation. I would suggest that Kant chose “transcendental” in order to flag
the particularity of TL.
 A remarkable confirmation of this claim can be found when the Paralogism-chapter argues
that pure self-consciousness does not consist in cognition of oneself as an object:
I do not cognize any object merely by the fact that I think, but rather I can cognize any ob-
ject only by determining a given intuition with regard to the unity of consciousness, in
which all thinking consists. Thus I cognize myself not by being conscious of myself as think-
ing, but only if I am conscious to myself of the intuition of myself as determined in regard to
the function of thought. All modes of self-consciousness in thinking are therefore not yet
themselves concepts of the understanding of objects (categories), but rather mere logical
functions, which provide thought with no object at all, and hence also do not present myself
as an object to be cognized. (B 406–407, italics added)
The functions that we use in all thought do not by themselves already represent an object. When
I self-consciously think the contentful concept “pomegranate”, it represents an object. Yet the

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 41

minology, concepts “have content”, which functions and forms lack.³³ Since the
Table attends merely to functions, it will be object-free. TL, in contrast, also has
the objects of its particular concepts in view.³⁴
I propose that it is this contrast, rather than any point that specially con-
cerns PGL, which is the primary message of the (β)-passages. This holds, first,
for the all-important Preamble, which announces that the Table will “abstract
from all content of a judgment in general [überhaupt]” (A 70/B 95).³⁵ Kant adopts
the overall plan of beginning with the functions because they provide an ac-
count of the “origin” of the categories (A 57/B 81). This particular reason for pro-
viding the Table is not cited in its Preamble, but it assumes centrality after the
presentation of the functions. Consequently, not only does § 10 contrast “general
logic” and TL in terms of whether they contain representations of objects, the
discussion also explains how object-free functions yield concepts.
The fact that object-free functions can be used to explain concepts (of ob-
jects) figures prominently in an exceptionally illuminating (χ)-passage. This pas-
sage provides a summary of § 10 in the course of preparing readers for a second
metaphysical deduction: “As in the case of the understanding, there is in the
case of reason a merely formal, i. e., logical use, where reason abstracts from

same is not true of “the mere logical functions” that unify the concept with my “representa-
tion ‘I’, which for itself is completely empty of content, and of which one cannot even say
that it is a concept […]” (A 345–346/B 404). Functions operate on concepts but are not them-
selves concepts.
 Famously, A 55/B 79 paraphrases “all content of cognition” as “all relation of [cognition] to
the object”. Once we are clear that Kant takes concepts, in contrast to forms, to represent objects,
it is less mysterious why he uses Inhalt to mean both “intensional content” and “object”, de-
pending on context.
 A full reconstruction of TL as a particular logic would require additional attention to these
object(s), which Kant calls Gegenstände überhaupt (cf. A 11). For current purposes it suffices to
note that (α) – (χ) treat them much like ordinary objects. To make complete sense of Kant’s po-
sition, though, we would need to see why he thinks that to study them really is to study the sub-
ject (“critique”), not objects (“doctrine”) (see ft. 19). For instance, Kant does not conceive of the
four species of nothing as types of objects, but rather as four ways that our cognition can fail (cf.
A 290–292/B 346–349). Full development of this and related points would allow us to marry the
two rationales for TL’s particularity. This must be attempted elsewhere.
 The target of this abstraction is the content “of a judgment überhaupt” (A 70/B 95), which I
suggest we render as “as such”. This conveys the Preamble’s point that the Table of Functions
abstracts from the categories (which represent the properties that any object has simply by virtue
of being the object of a judgment) better than “in general”, which falsely suggests that the Table
brackets any and all consideration of content. As noted in § 2, the Table must attend to the “con-
tent” of the predicate-term if it is to include the infinite function (A 72–73/B 97–98).

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
42 Timothy Rosenkoetter

all content of cognition, but there is also a real use, since reason itself contains
the origin of certain concepts […]”.³⁶ After referring, somewhat inaccurately, to
each use as a “capacity” and calling the second a “capacity, which itself gener-
ates concepts”, we read:

Now since a division of reason into a logical and a transcendental capacity occurs here, a
higher concept of this source of cognition must be sought that comprehends both concepts
under itself, while from the analogy with concepts of the understanding, we can expect
both that the logical concept will put in our hands the key to the transcendental one and
that the table of functions of the former will give us the family tree of the concepts of rea-
son.³⁷

This passage illustrates nicely why we should not simply assume that (χ)-style
contrasts of the “logical” with the “transcendental” are referring to PGL in par-
ticular. In each of the (χ)-passages Kant is instead using “logical”, just as he does
here, to designate how we treat thought when we abstract from concepts and
their implied objects and focus instead on forms or functions, which by them-
selves are not concepts. Though Kant’s terminology makes it natural to assume
that there is just one such object-free perspective on thought, we have not en-
countered any substantive support for this assumption. The next section will dis-
tinguish the two object-free perspectives that constitute PGL and LC and sketch
how it comes to be that Kant calls both logics by a single name.

4 General Logics
Kant’s use of a single term, “pure general logic”, to designate two distinct logics
can be traced to an ambiguity in the sense in which general logics can be said to
contain “[…] the absolutely necessary rules of thinking, without which no use of
the understanding takes place […]”.³⁸ Taken in one natural sense, the absolutely
necessary rules of thinking are just those with which every thought must be con-
sonant, since otherwise it would not be a thought. This sine qua non construal of
the quoted definition will turn out to yield PGL.

 A 299/B 355, italics added; cf. A 333/B 390.

 A 299/B 355, italics added. The inaccuracy noted above is that there are really just two uses (a
logical use and a real use) of a single capacity. The section containing this passage, when viewed
as a whole, makes this tolerably clear.
 A 52/B 76, italics added. Nominally, this is a definition of general logic. Yet since it directly
contradicts what Kant goes on to say about the only other named species of general logic (ap-
plied general logic), it is clear that Kant intends it as a definition of pure general logic.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 43

One hint that there is an alternative is that within the genus of general logic
Kant treats applied general logic as the sole alternative to pure general logic.
Since the applied version treats the use of the understanding “under the contin-
gent conditions of the subject, which can hinder or promote this use”, there is
space for a general logic that is not applied – and so “pure” by the terms of
Kant’s taxonomy – simply because it brackets the variable ways in which the un-
derstanding is realized “in concreto” (A 54/B 78–79). The rules of this general
logic – or the distinctions that it makes³⁹ – are “absolutely necessary” only in
the weaker, capacity-relative sense that they reflect facts about the capacity it-
self, as opposed to the variable contexts in which it operates. This is relevant be-
cause Kant clearly takes the distinction between the general and singular func-
tions and the distinction between the affirmative and infinite functions to belong
to the nature of the understanding itself, rather than to any contingent circum-
stances of its operation. Yet the Appendices are unambiguous that these distinc-
tions, unlike the principle of contradiction, are not sine qua non conditions on all
thought whatsoever. So it makes sense that they belong to a general logic whose
rules and distinctions are “absolutely necessary” in the weaker, capacity-relative
sense.
If Kant had only been consistent in maintaining this weaker construal of the
relevant necessity, there would have been nothing misleading about locating the
Table within “pure general logic”. Yet in some contexts, as we will see shortly,
the strong construal is precisely what Kant is looking for. The result is that the
Transcendental Analytic moves back and forth between these two conceptions
of general logic without notice, making use of whichever construal meets the de-
mands of the moment. In order to track this movement and ask whether it can
help us to identify the logic of the Table, we need first to sketch how the logic
that arises from the sine qua non construal becomes PGL as interpreted above
in § 2, namely: the logic that “[…] considers only the logical form in the relation
of cognitions to one another […]” (A 55/B 79, italics added).
Kant takes for granted that there is only one property that thoughts must
possess in order to qualify as such: all thoughts whatsoever are non-contradic-
tory. Yet it is not immediately clear what this has to do with PGL’s constitutive

 Though Kant introduces the very notion of a logic by speaking of “rules” (A 52/B 76), we
must be careful about concluding from this that general logics consist exclusively in rules.
PGL contains both the requirement to avoid contradiction and permissive inference rules. Yet
we should not simply assume that a general logic can only recognize the special status of
(e. g.) the infinite function if it also includes a corresponding requirement or permission. Logics
are distinct from doctrines because they study a capacity, and some versions of this study may
bottom out in a list and classification of the capacity’s fundamental actions.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
44 Timothy Rosenkoetter

focus on the relations that thoughts bear to one another. After all, the use of the
principle of contradiction as a “conditio sine qua non” (A 59/B 84) can proceed
one representation at a time. So where is the relation? The answer is that contra-
diction is not a monadic property. Even when the principle of contradiction is
used merely as a filter to exclude non-thoughts singly, it is the relation of contra-
diction between two (or more) of their constituents that justifies this exclusion.
Of course, Kant takes PGL to do more than just (i) partition representations into
two classes, thoughts and non-thoughts. He includes within the same logic the
“positive use” use of the principle of contradiction in cognizing (ii) analytic
truths and (iii) rules of valid inference (A 151/B 190). What ties (i) – (iii) together
as all belonging to PGL is not merely that they share the same supreme principle.
It is also that each is composed of the rules that we discover when we ask which
pure principles govern the relations that thoughts bear to other thoughts, all the
while bracketing consideration of anything other than thought, including the in-
tuitions by means of which thoughts refer to objects.
Now that we have some insight into the connection between the sine qua non
construal of necessity and the constitutive perspective of PGL, let’s compare it to
the perspective that gives rise to a second logic that is both general and pure. We
can do this by examining three sections of the Critique in which Kant moves back
and forth between these perspectives: (1) § III of the Introduction to Transcen-
dental Logic (A 57–62/B 82–86); (2) the opening of the Transcendental Analytic,
stretching from A 64/B 89 until § 10’s transition into TL; and (3) On the Supreme
Principle of All Analytic Judgments (A 150–153/B 189–193).
The first of these sections is largely taken up with PGL’s role as a filter ex-
cluding non-thoughts and so is a pure expression of the sine qua non construal
of pure general logic. At first (3) appears to be different, since recognition of an
analytic truth counts as “cognition” of an “object [Objekt]” (A 151/B 190). But
these terms are polysemous. Because PGL brackets thought’s relation to intu-
ition, the objects of analytic cognition are no more than posits from within
thought.⁴⁰ Though it may be the case that the object of this or that analytic cog-
nition is real-possible, analytic cognition by itself does not require this.⁴¹ Corre-

 We find hints of this in Kant’s careful wording: “For the contrary of that which as a concept
already lies in, and is thought in, the cognition of the object [Objekt] is always correctly denied
[…]” (A 151/B 190, italics added).
 One indication of this is that throughout the section Kant uses Ding and Objekt for objects
known entirely through the positive use of the principle of contradiction. The sole occurrence
of Gegenstand comes when Kant wishes to convey a more robust conception of objects: “But
even if there is no contradiction within our judgment, it can nevertheless combine concepts
in a manner not borne out by the object” (A 150/B 189–190).

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 45

latively, Kant is using “cognition” here in a weak sense that contrasts with the
robust sense that figures in LC’s perspective. Thus, (1) and (3) are both concerned
exclusively with standards of consistency internal to thought.
What about (2)? Though these sections follow almost immediately upon (1),
the species of general logic that figures in these opening sections of the Tran-
scendental Analytic is engaged in a much more ambitious undertaking than
PGL. The section “On the Logical Use of the Understanding in General [über-
haupt]” opens with a sentence that flags the fact that general logic, as the reader
has encountered it thus far (namely, as PGL), has been considering the under-
standing in abstraction from intuition: “Above the understanding was explained
merely negatively: as a non-sensible capacity of cognition” (A 67/B 92). The re-
mainder of that section offers, by contrast, a positive explanation of the capacity,
which requires that Kant take into account that concepts refer to objects via in-
tuitions.⁴² Consequently, its logic regards judgment as “the mediate cognition of
an object”, with “cognition” now intended in the robust sense (A 68/B 93, italics
added). This logic, which is both pure and general, is LC. There is much more to
be said about the connections between LC’s perspective and Kant’s choices in
“On the Logical Use”. However, the crucial point for present purposes is that
there is absolutely no indication of a sudden shift in the logic under considera-
tion in the transition from that section to § 9, which contains the Table. Accord-
ingly, even prior to considerations that hinge on the Table’s particular list of
functions, we should expect that it reflects the perspective of LC.
Having marked off PGL and LC from one another, we can now formulate a
minimal claim concerning the Table: the logic that distinguishes three functions
(and their corresponding forms) under quantity and quality is LC, while PGL rec-
ognizes only two logical forms in each case. Though I believe this minimal claim
to be correct, it stops short of what the identification and isolation of LC can do
to further our understanding of the Table. Our working hypothesis should be that
all of the Table’s functions reflect the perspective of LC. To be clear, this is conso-
nant with recognizing that PGL makes distinctions in logical form that corre-
spond to some of the Table’s functions. This is obviously true of the general, par-
ticular, affirmative, and negative forms, which must be distinguished from one
another if PGL is to contain rules of valid syllogistic inference for A-, E-, I-,
and O-judgments. However, this surface-level coincidence does not show that
these are precisely the same distinctions.⁴³ Each function will require its own in-

 For related observations, cf. Longuenesse 2006, p. 139.

 Even correcting for the fact that the Table contains functions, while PGL concerns only
forms. For a prima facie case that there is a difference between the two logics with respect to
the universal-particular distinction, cf. Prol 302.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
46 Timothy Rosenkoetter

vestigation. This paper can do no more than begin this project for the functions
of quality, thereby at least providing an example of the kinds of questions that
we need to ask about all of the remaining functions (§ 6). First, however, it is im-
portant to gain a more secure foundation for that project by considering two
challenges to the separateness of PGL and LC.

5 Individuating Logics
The interpretation on offer requires that we posit a logic that is distinct from both
PGL and TL. It is obviously important to consider whether the same work can be
done more economically. The most popular way to resolve perceived tensions in
Kant’s statements is inclusivism, the position that denies that there is a simple
answer to the question of the Table’s home. Inclusivism almost always takes
the particular form of holding that though the Table belongs to PGL, the two ex-
traordinary functions reflect the perspective of TL. For this to be true there
must be one sense in which PGL recognizes singular and infinite judgments,
though in another sense it does not and TL does. Inclusivists have specified
the second sense by suggesting that only TL recognizes the singular and infinite
forms of judgment as “basic”, “relevant”, or “useful”.⁴⁴ What they seem to have
in mind for the first sense is that PGL recognizes these forms insofar as it treats
them as well-formed judgments, thereby enabling us to see that “This AB is B” is
true and to construct syllogisms that include negated predicate-concepts. This
claim is true but irrelevant. There is no meaningful sense in which a list of pu-
tatively fundamental forms (or functions) of judgment belongs to PGL simply be-
cause PGL treats those forms as syntactically well-formed and distinct from one
another. After all, the same argument might be used to show that a table that
includes indeterminately many forms (including, e. g., disjunctive judgments
with seven disjuncts) belongs to PGL.⁴⁵ It belongs to PGL only in a weak sense
that is irrelevant to identifying the logic whose analysis of the understanding
finds expression in the Table. In the relevant sense of belonging, the Table be-

 Lu-Adler 2014, pp. 372 & 369; Allison 2004, p. 141.

 An alternative form of inclusivism (perhaps closer to Lu-Adler 2014) would hold that PGL at-
taches importance to the distinction between universal and singular judgments (and likewise for
quality), as opposed to simply declaring both forms syntactically acceptable and distinct (along
with indeterminately many other forms). What PGL does not do, but TL does, is to grant them a
further status, which we might call robust importance. I find it mysterious what these two grades
of importance could amount to.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 47

longs to a logic only if it reflects the perspective that this logic takes on our ca-
pacity.⁴⁶
Other than inclusivism, the most promising alternative to recognizing LC as a
separate logic is the Robust PGL-interpretation. This assigns the Table to what
Kant calls “pure general logic”, while interpreting that logic robustly, so that it
includes all uses of the principle of contradiction, as well as the positive explan-
ation of the understanding in “On the Logical Use” (A 67–69/B 92–94). PGL and
LC are thereby combined into a single logic. This is a serious position that merits
close consideration. Yet it faces three problems that speak in favor of what I take
to be its sole remaining competitor, the LC-interpretation.
First, the Robust PGL-interpretation has difficulties explaining Kant’s expos-
itory choices in the Appendices. The central point is that when Kant is contrast-
ing the logic that does not recognize the special contributions of the extraordi-
nary functions with the logic that does, he treats them as two distinct logics.
This is a strange choice if Kant really believes that the logic responsible for rec-
ognizing analytic truths and valid inferences is the same as the logic that recog-
nizes the special contributions of those two functions. The LC-interpretation ob-
viously fares better in this respect. Yet this may not matter if it falters on the fact
that Kant calls the second logic TL. Somewhat surprisingly, though, so long as
the TL-interpretation is indeed untenable (§ 3), Kant’s mention of TL in Appen-
dix 2 actually speaks in favor of the LC-interpretation. First, though LC does
not meet the technical definition of TL, it is still recognizably transcendental
in a relaxed sense, since it concerns “our manner of cognition of objects”
(B 25). Second, Kant finds himself in Appendix 2 in the unusual situation of
needing to contrast PGL and LC, so he cannot follow his usual practice of simply
letting “general logic” refer to whichever one fits the present context (§ 4). Under
these circumstances it is not surprising that Kant would call LC “transcenden-
tal”, opting for the closest logic that will not leave him, absurdly, contrasting
“general logic” with “general logic”.
The final two problems share the same form: if there is only one logic, Ro-
bust PGL, then Kant assigns inconsistent properties to it. We know, for instance,
that “pure general logic” is defined by its abstraction from all content, including
the intensional content of the predicate-terms in judgments that it treats.⁴⁷ Yet

 A radical version of inclusivism would hold that the reason why there is no simple answer to
the question of the Table’s home is that in the final analysis the Table belongs to two different
logics, with the singular and infinite forms (or functions) assigned to TL and the remainder as-
signed to PGL. In contrast to the versions discussed above, there is nothing confused about this
version of inclusivism. However, it conflicts with Kant’s belief that the Table is a (single) system.
 Cf. A 72/B 97 and A 598/B 626.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
48 Timothy Rosenkoetter

Appendix 2 makes clear that attention to that very content provides the sole ra-
tionale for distinguishing the affirmative and infinite functions. This second
problem is neatly resolved if we posit LC.
The final problem arises because Kant is convinced that the Table is a sys-
tem.⁴⁸ This means that it instantiates a species of unity that can only be under-
stood when its manifold is interpreted in light of an idea that is also an end.⁴⁹
Kant’s claim raises many important and difficult questions that cannot be treat-
ed here. What is essential for our purposes is a series of broad contrasts that we
find when we compare PGL with the Table and its logic. First and primarily,
though the Table and PGL are both unities, we can understand PGL’s unity with-
out appealing to the elaborate theory that Kant uses in his attempt to clarify sys-
tematic unity. As sketched above, PGL’s unity can be traced to the fact that a sin-
gle principle – one whose application is determinate in a way that we simply do
not find with ideas – can be used negatively as well as positively. The bounds of
PGL are traced by what can be accomplished when this rule concerning contra-
diction is applied to individual cases. In contrast, the Table’s unity is ultimately
explained by the fact that the understanding is a capacity for judgment. This uni-
fies the parts of the Table into the “form of a whole” (A 832/B 860), rather than
by serving as a rule that can be mechanically applied to individual cases. These
primary differences just sketched should not be surprising, second, given that
the Table arises from an account of a capacity and its end. In order to grasp
the principle of contradiction, in contrast, one need only recognize that contra-
diction annuls thought.
A third difference between the Table and PGL hinges on the fact that the
techniques of PGL require no special mental acuity among their practitioners,
which is not true of philosophy. This is not to say that every competent subject
can employ all valid rules of inference, no more than every competent subject
knows the truths of mathematics. Yet Kant is convinced that mathematics and
logic can be successfully taught, so that a competent subject who is paying at-
tention will gain these skills. He argues that this is not true of philosophy.⁵⁰ In
the context of making this point he takes the important step of denying the
title “philosopher” to the “mathematician” and the “logician”, insisting that
they are only “artists of reason [Vernunftkünstler]” (A 839/B 867). Of course,
Kant would never think of denying that the Table is philosophical. Achieving
philosophical insight into the Table’s unity is not a “trade” that can be taught,

 Cf. A 64–67/B 89–92, A 80–83/B 106–109, and Prol 322–326.

 For Kant’s theory of systematic unity, cf. esp. A 832–851/B 860–879. A 67/B 92 speaks of the
role of an “idea” in unifying “transcendental-philosophy” (by means of the Table).
 Cf. A 836/B 864.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 49

in which incremental progress is guaranteed to the diligent pupil.⁵¹ So the logic

that Kant considers a mere art must be distinct from the Table’s. It seems clear
that it is PGL. This fits Kant’s belief that PGL was completed long ago, save for a
few inessential subtleties, whereas he was the first to grasp the Table in its sys-
tematic unity.⁵² Of course, this is not to say that logicians prior to Kant were un-
familiar with the various logical forms whose corresponding functions make up
the Table. Yet even if some of Kant’s predecessors had happened upon an exten-
sionally equivalent list, they will not have attained the perspective of the Critique
so long as they did not understand each function in light of its relation to the
idea that explains the Table’s systematic unity. Now, one insight that Kant
takes himself to have been the first to achieve in full clarity is that both concepts
and intuitions are required for cognition of an object. Accordingly, if the above
reading is correct in taking this insight to be essential to the logic of the Table, it
should be no surprise that previous logicians were without this key to the foun-
dation of philosophy.

6 Quality
The Table is naturally understood as split into two parts: a larger portion, which
can be fully understood from the perspective of PGL, and the two extraordinary
functions, which require something more. At first glance, the functions of quality
fit this expectation perfectly. What makes a judgment affirmative or negative is a
particular relation between its subject- and predicate-concepts: “Logical nega-
tion, which is indicated solely by the little word “not”, is never properly attached
to a concept, but rather only to its relation to another concept in a judgment
[…]”.⁵³ However, the distinguishing feature of infinite judgment is not similarly
relational. An affirmative judgment, “The soul is mortal”, is converted into an
infinite judgment simply by negating the predicate-concept, while leaving the
copula unchanged.
At this point we can pause and notice that Kant could have stopped with this
characterization of infinite judgments, which would have sufficed to distinguish
them from affirmative judgments. Instead, we are also told that the logic of the

 Here I assume that Kant would extend the following claim to logics as an “art”: “For math-
ematics a completely different mind is required than for philosophy […]. Philosophy is more a
science of genius, mathematics in contrast more an art[;] one can learn it as a trade [Handwerck]
[…]” (Ak. 25, p. 164, my translation).
 Cf. B viii; A 767/B 795, on how Hume did not see the whole.
 A 572/B 600; similarly, NG 172.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
50 Timothy Rosenkoetter

Table considers whether these negative predicates augment “the whole of cogni-
tion” (A 72/B 97). It is by virtue of the fact that infinite judgments are “merely
limiting with respect to the content of cognition in general” that the correspond-
ing function deserves a special place in the Table (A 73/B 98). Now, clearly, the
logic that distinguishes the infinite function regards judgments inter alia as cog-
nitions. Yet why should the shift to this perspective yield just one new function?
Why doesn’t quality include a fourth function, which is distinguished by the fact
that it amplifies the content of cognition? The answer, I suggest, is that this miss-
ing fourth function is already to be found in the existing Table in the guise of its
affirmative function. Otherwise, the fact that infinite judgments do not augment
the content of cognition, despite having an affirmative “logical form”, would sim-
ply not count as noteworthy (A 72/B 97).
If this line of reasoning is correct, then PGL’s exclusively relational under-
standing of affirmative judgment is foreign to the logic of the Table. This is
not to deny that the affirmative function enables us to relate subject- and pred-
icate-concepts using a copula. Kant’s position is instead that it accomplishes this
as part of an act whose purpose is to increase what is known about objects.
Kant’s parallel position regarding the negative function is that its purpose is
to avoid error. Appendix 2 itself contains tolerably clear evidence of this (cf. A 72/
B 97), but it is explained at greater length later in the Critique:

Logically one can, to be sure, express any proposition [Sätze] that one pleases negatively,
but in regard to the content of our cognition in general, whether it is expanded or limited
by a judgment, negative judgments have the peculiar job, solely of preventing error. ⁵⁴

It would of course be misguided to hold within PGL that the purpose of negative
judgments is strictly to prevent error. Kant recognizes this and treats affirmative
and negative judgments symmetrically within the context of inference. He also
treats affirmative and negative analytic truths symmetrically.⁵⁵
Though this leaves several important questions regarding functions of qual-
ity untouched⁵⁶, this brief coda has at least made the case that it is not merely

 A 709/B 737. Kant uses almost exactly the same phrase in Appendix 2, when explicating the
perspective which distinguishes the infinite function: “in regard to the content of cognition in
general” (A 73/B 98).
 Regarding the former, the minor premise of a disjunctive syllogism is always a negative judg-
ment. Yet when used in this context its purpose is not to avoid error. Regarding the latter, the
treatment of the positive use of the principle of contradiction leaves no doubt that affirmative
and negative judgments are on equal footing (A 151/B 190).
 Prime among them is Kant’s case for holding that the subject-concept in an infinite judgment
does not “grow” and is not “affirmatively determined”, simply because the predicate-concept is

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
 The Logical Home of Kant’s Table of Functions 51

the infinite function, but all three functions of quality, that reflect the perspective
of LC. This shows that it is dangerous to assume that there are two extraordinary
functions, while the remaining ten are taken over from PGL. It behooves us to
examine each of the remaining functions closely with the constitutive perspec-
tive of LC as our interpretative key.

References
Ak. | Kant, Immanuel (1900 ff.): Gesammelte Schriften. Königlich Preußische Akademie der
Wissenschaften (ed.). 29 vols. Berlin: De Gruyter.
GMS | “Grundlegung zur Metaphysik der Sitten”. In: Gesammelte Schriften, Vol. 4.
KU | “Kritik der Urteilskraft”. In: Gesammelte Schriften, Vol. 5.
MAN | “Metaphysische Anfangsgründe der Naturwissenschaft”. In: Gesammelte Schriften,
Vol. 4.
NG | “Versuch, den Begriff der negativen Größen in die Weltweisheit einzuführen”. In:
Gesammelte Schriften, Vol. 2.
Prol | “Prolegomena zu einer jeden künftigen Metaphysik, die als Wissenschaft wird auftreten
können”. In: Gesammelte Schriften, Vol. 4.

Allison, Henry (22004): Kant’s Transcendental Idealism: An Interpretation and Defense. New
Haven, London: Yale University Press.
Bennett, Jonathan (1966): Kant’s Analytic. Cambridge: Cambridge University Press.
Beck, Lewis White (1992): “Foreword”. Reich, Klaus: The Completeness of Kant’s Table of
Judgments. Kneller, Jane; Losonsky, Michael (trans.). Stanford: Stanford University Press,
xi–xx.
Bird, Graham (2006): The Revolutionary Kant: A Commentary on the Critique of Pure Reason.
Chicago, La Salle (Ill.): Open Court.
Brandt, Reinhard (1991): Die Urteilstafel. Kritik der reinen Vernunft A 67–76; B 92–101.
Hamburg: Felix Meiner.
Forster, Michael (2008): Kant and Skepticism. Princeton, Oxford: Princeton University Press.
Greenberg, Robert (2000): Kant’s Theory of A Priori Knowledge. University Park, PA:
Pennsylvania State University Press.
Hanna, Robert (2001): Kant and the Foundations of Analytic Philosophy. Oxford: Clarendon.
Kant, Immanuel (1995 ff.): The Cambridge Edition of the Works of Immanuel Kant. Guyer, Paul;
Wood, Allen (eds.). Cambridge: Cambridge University Press.
Kemp Smith, Norman (21992): A Commentary to Kant’s “Critique of Pure Reason”. Atlantic
Highlands, NJ: Humanities Press.
Longuenesse, Béatrice (1998a): Kant and the Capacity to Judge. Wolfe, Charles T. (trans.).
Princeton: Princeton University Press.

negated (A 72–73/B 98). From today’s perspective this asymmetry is puzzling. My conjecture is
that Kant’s case for the asymmetry makes specific use of LC’s attention to the dependence of
conceptual cognition on intuitions.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM
52 Timothy Rosenkoetter

Longuenesse, Béatrice (2006): “Kant on A Priori Concepts: the Metaphysical Deduction of the
Categories”. In: Guyer, Paul (ed.): The Cambridge Companion to Kant and Modern
Philosophy. Cambridge: Cambridge University Press, pp. 129–168.
Lu-Adler, Huaping (2014): “Kant on the Logical Form of Singular Judgements”. In: Kantian
Review 19, pp. 367–392.
Paton, H. J. (1936): Kant’s Metaphysic of Experience, 2 vols. London: George Allen & Unwin.
Prichard, H. A. (1909): Kant’s Theory of Knowledge. Oxford: Clarendon.
Reich, Klaus (1992): The Completeness of Kant’s Table of Judgments. Kneller, Jane; Losonsky,
Michael (trans.). Stanford: Stanford University Press.
Strawson, Peter F. (1966): The Bounds of Sense. London: Methuen.
Tolley, Clinton (2012): “The Generality of Kant’s Transcendental Logic”. In: Journal of the
History of Philosophy 50, pp. 417–446.
Waxman, Wayne (2014): Kant’s Anatomy of the Intelligent Mind. Oxford, New York: Oxford
University Press.
Wolff, Michael (1995): Die Vollständigkeit der kantischen Urteilstafel. Frankfurt a. M.: Vittorio
Klostermann.
Wolff, Robert Paul (1963): Kant’s Theory of Mental Activity. Cambridge: Harvard University
Press.

Brought to you by | University of Sydney Library

Authenticated
Download Date | 6/3/17 11:02 PM