publisher colophon

What is the nature of Kant's new 'transcendental' logic, and how is it to differ from the traditional logic? Many have argued that transcendental logic is distinguished by the domain of understanding it investigates - whether by focusing on a domain excluded from traditional logic, or by focusing a more specific domain. Here I argue that transcendental logic should not be characterized by a difference in domain at all, but rather in terms of the aspect of understanding at issue: while traditional logic investigates the form of understanding, transcendental logic studies its content. This interpretation fits better with Kant's claim that the transcendental-logical categories and the traditional-logical forms "completely coincide." It also clarifies Kant's doctrine of the spontaneity of our understanding, by highlighting what our understanding is capable of achieving on its own.

§ 1. introduction

In the first Critique, Kant introduces a new logic, one that he calls "transcendental logic." The importance that Kant attributes to this new logic would seem to be evident from the fact that he uses 'Transcendental Logic' as the title for what is by far the largest part of the first Critique itself. What is more, the core elements of this logic—what Kant calls the "pure concepts" or "categories" of understanding— are seen by Kant to play an absolutely crucial role, not just in his account of the limits of theoretical cognition in the first Critique, but throughout the rest of his Critical philosophy as well.1 Even so, since the publication of the first Critique, there has been considerable disagreement about what exactly Kant takes the nature of his new logic to be, and disagreement, in particular, about how Kant means to distinguish his new transcendental logic from what had traditionally gone under the name of 'logic.' My goal here is to make headway on this topic by showing the inadequacies of two of the most prevalent interpretations of the status of transcendental logic, and then arguing for a more promising alternative.

As I describe below in §2, most interpreters have construed the difference between the traditional logic and Kant's new logic in terms of a difference in scope or domain. One version of this interpretation takes transcendental logic to have in view a domain that is excluded from the traditional logic. Another takes the distinction to lie instead in the fact that, while the traditional logic gives principles that govern all thinking and judging, and hence is universal in its scope, transcendental logic, by contrast, focuses only on particular species of thinking and judging, and so is more restrictive. [End Page 417]

In §3 I present reasons for worrying about the viability of each of these interpretations. In §§4-5 I then argue that if we take a closer look at the sections in the first Critique in which Kant first introduces the idea of his new logic, an alternative understanding of this discipline emerges. I argue, more specifically, that these sections make it clear that Kant means to distinguish transcendental logic from the traditional logic, not in terms of domain at all, but rather in terms of the aspect of understanding that is in view. Unlike the traditional logic, which focuses only on the form of thinking and judging, Kant intends his new transcendental logic to focus on the content of thinking and judging, albeit in a very abstract manner.

In §§6-7 I supplement this argument from textual analysis with a more systematic argument against such "domain-sensitive" interpretations, as I will call them. This argument draws on premises from Kant's doctrine of conceptual content in general. Here I will show that Kant's conception of the nature of the content at issue in transcendental logic—what Kant calls the "transcendental content" of the pure ("unschematized") concepts of understanding (B105)—shows it to be as universally present throughout our thinking and judging as are the traditional-logical forms. For this reason, I will conclude that, rather than standing in a relation of exclusion or subordination, the two logics are, in fact, domain-coincident, with transcendental logic being every bit as unrestrictedly general or universal as the traditional logic.

In the final section (§8), I return to reconsider one of the points that will have been initially presented in §2 as offering, prima facie, key support for the domain-sensitive accounts, in order to show how it can be accommodated by the interpretation of the two logics I develop below. I conclude with a brief summary.

Revisiting the nature of Kant's transcendental logic promises to have several significant payoffs for our understanding of Kant's project in the first Critique and beyond. For one thing, it will allow us to clarify Kant's strategy in the so-called "Metaphysical Deduction" of the categories from the elements of the traditional logic, both by making clearer what Kant could mean by his claim that the categories and these elements "completely coincide" (B159), but also by more precisely delineating what Kant means by claiming that the traditional logic is a "merely formal" logic and why he thinks a non-formal logic is necessary.2 More generally, it will let us carve out the proper place within Kant's system for what have come to be known as the "unschematized" categories, one that highlights the extent to which they are products of the understanding alone, via its spontaneity, and [End Page 418] one that allows them to have both the significance and even the "use" outside of theoretical contexts that Kant clearly accords to them.

§ 2. A Special Domain for Transcendental Logic?

As I have indicated above, the most common interpretations of the significance of transcendental logic take Kant's key motivation in introducing the new logic to be a new sensitivity to the variety of domains of understanding. The first version is one that I will call the 'domain-exclusive' interpretation. This version was especially common for much of the later nineteenth and early twentieth century, being embraced by Friedrich Ueberweg, Hermann Cohen, and Norman Kemp Smith, among others.3 Its proponents maintain that Kant found it necessary to introduce transcendental logic in order to deal with a new kind of judgment that he had discovered—namely, synthetic judgment. Whereas the logic of the Leibnizians such as Wolff, Baumgarten, and Meier had been based on the assumption that all judgments were, in effect, what Kant himself would now call analytic judgments, the fact that Kant had shown this assumption to be false required that he introduce a parallel logic to cover the new kind of judgment, and in this way supplement the old logic. Yet because the two species of judgments (i.e. analytic and synthetic) are exclusive of one another, it was presumed that their corresponding logics should also be distinguished by their exclusive domains—hence the label 'domain-exclusive interpretation.'

As evidence for this interpretation, such readers pointed to the fact that in the first Critique, Kant clearly means to associate the traditional logic with analytic judgments and transcendental logic with synthetic judgments. For one thing, although Kant claims explicitly that the law of contradiction—clearly a law of the traditional logic—is the "supreme principle" of analytic judgments (B189), the principle that Kant identifies for synthetic judgments is one that does not seem to be logical in any familiar sense, insofar as it makes reference to inner sense and time (B194). What is more, Kant claims explicitly that although "the explanation of the possibility of synthetic judgments" is "the most important business of all" for transcendental logic, this explanation "is a problem with which [traditional] logic has nothing to do, indeed whose name it need not even know" (B193).

Despite such textual grounding, the domain-exclusive interpretation came in for heavy criticism shortly into the twentieth century. Leading the way were Klaus [End Page 419] Reich and especially H. J. Paton.4 Both Reich and Paton argued that this interpretation fundamentally misunderstood the unrestricted universality that Kant means to ascribe to the traditional logic—an objection that itself enjoys considerable textual support, as we will see in a moment (see §3). For his part, however, Paton accepted that there was something correct about the earlier interpretation's characterization of transcendental logic in terms of its peculiar domain. This is its claim that transcendental logic is focused on only one species of judgment— namely, synthetic judgments. In Paton's words, whereas the traditional logic "is concerned with the necessary rules, or the necessary form, of all thinking," Kant's new logic, by contrast, "studies, and studies only [my emphasis], the rules of synthetic a priori thinking" (Metaphysic, 222); "while [traditional] Logic deals with all thought," transcendental logic "deals with a particular kind of thought (synthetic a priori thinking)" (Metaphysic, 223; my emphasis). Hence, rather than standing in an exclusion-relation, Paton argued that Kant instead takes the domains of the two logics to stand in a relation of subordination, with the traditional logic being distinguished by having a wider scope than transcendental logic.5

Among more recent commentators, this 'domain-subordinative' interpretation, as I will call it, has enjoyed increasingly widespread acceptance. Giorgio Tonelli, Michael Wolff, John MacFarlane, and Jay Rosenberg, among several others, have all followed Paton in claiming that the main contrast between transcendental logic and the traditional logic lies in the fact that transcendental logic is concerned only with a particular subset of thinking and judging.6 Such readers have claimed, moreover, that this means that transcendental logic should be classified as a logic of what Kant calls the "special or particular [besondere] use of our understanding," one that "contains the rules for correctly thinking about a certain kind [eine gewisse Art] of objects," rather than all objects (B76; my emphasis). For this reason, they conclude that transcendental logic should be viewed as what Kant would call a "special or particular" logic, rather than a general or universal one that would be on par in this respect with the traditional logic.

To be sure, not all of those drawn to the domain-subordinative interpretation have agreed as to which subset of thinking and judging it is that is in view in transcendental logic. Tonelli and Wolff follow Paton in claiming that transcendental logic is a special logic because it is devoted exclusively to the synthetic a priori thinking and judging about the special objects of metaphysics.7 MacFarlane, by [End Page 420] contrast, has argued that transcendental logic is a special logic because it involves a "restriction to objects capable of being given in human sensibility," and hence a restriction to synthetic a priori judgments about objects of possible experience.8 This, in turn, would make transcendental logic even more restricted in its scope than Paton's original interpretation, since it would entail that thoughts and judgments about the traditional objects of special metaphysics (God, the world-whole, and the soul) would be excluded from transcendental logic's domain.9

Perhaps even more so than the domain-exclusive interpretation, the domain-subordinative interpretations would seem to have a relatively straightforward textual basis for several of their claims. Though its adherents will give it a different gloss, they too can appeal to Kant's claim (cited above) that while "the explanation of the possibility of synthetic judgments" is "a task" that the traditional logic "has nothing to do with" and indeed "does not even need to know its name," this is by contrast "the most important business of all in a transcendental logic" (B193). For it will now be urged that the traditional logic does not need to know its name because it is aimed at a higher level of abstraction than is the new logic. What is more, in Jäsche's edition of Kant's lecture notes on logic, we find transcendental logic contrasted with the traditional logic on just these grounds: in transcendental logic, "the object itself is represented as an object of the mere understanding," whereas the traditional logic "deals with all objects as such" (auf alle Gegenstände überhaupt geht, Jäsche Logik §I, 9:15; my emphasis).

For their part, those following Paton can point as well to the fact that the portion of the first Critique entitled "Transcendental Logic" includes the Transcendental Dialectic, in which Kant submits the traditional metaphysics to some of his most trenchant criticism and which does in fact treat of thinking about three special objects (the immortal soul, the world-whole, and God). It also includes the Transcendental Analytic, which culminates in Kant's well-known claim that this inquiry itself must supplant traditional ontology: "[T]he proud name of an ontology . . . must give way to modest one of a mere Analytic of the pure understanding" (B303). MacFarlane and others, by contrast, can point to the inclusion of second main part of the Transcendental Analytic within the section entitled "Transcendental Logic," a part that Kant calls "the Analytic of Principles," because it is undeniable that Kant's analysis is guided throughout, from the Schematism [End Page 421] onwards, by reflection on the basic features of space and especially time as our particular forms of sensing objects (see B177-78).

§3. Problems with Domain-Sensitive Interpretations

What I want to show now is that despite such prima facie textual support for each view, a closer look at Kant's actual descriptions of the subject-matter of transcendental logic makes all of the foregoing versions of domain-sensitive interpretations look less and less plausible.

As both Reich and Paton insist, the domain-exclusive interpretation stands in direct conflict with Kant's persistent description of the traditional logic as a genuinely "general or universal [allgemeine]" science, whose principles were to be valid of all kinds of thinking and judgings, and hence valid of both analytic and synthetic judgments alike. In Kant's words, this logic "contains the absolutely necessary rules of thinking without which no use of understanding takes place" (B76; my emphasis). Indeed, Kant seems to say exactly this in the Prolegomena's discussion of the analytic/synthetic distinction, insofar as Kant claims there that this distinction is not one that pertains to the "logical form" of judgments at all, but instead pertains to their "content [Inhalt]."10 For these reasons, any interpretation that takes Kant to exclude a species of judgment from the jurisdiction of the traditional logic must be viewed with considerable suspicion.

There are, however, considerable problems facing the two versions of the domain-subordinative interpretations as well. Against the idea that transcendental logic will be focused on thought of sensible objects, we should note that Kant claims explicitly that "in a transcendental logic, we isolate the understanding" (B87; my emphasis), such that its subject-matter, "the pure understanding," is something that "separates itself completely not only from everything empirical, but even from all sensibility" (B89; my emphasis). This is further confirmed by remarks throughout in the first Critique,11 where Kant makes it clear that the core elements of the subject-matter of transcendental logic, the pure concepts or categories of understanding, are such as to "relate to objects generally without any conditions of sensibility" (sich auf Gegenstände ohne alle Bedingungen der Sinnlichkeit allgemein beziehen, B120; my emphasis). Far from being restricted, Kant thinks that "the categories extend themselves farther than sensible intuition, because they think objects in general [überhaupt], without seeing to the particular manner [Art] (of [End Page 422] sensibility) in which they might be given" (B309; my emphasis).12 All of this makes it hard to see how a science concerned with such elements should nevertheless have its focus restricted by conditions imposed by sensibility, as MacFarlane and others would have it, given that sensibility and its conditions are said explicitly not to be in view at all.

Against the idea that transcendental logic is concerned only with thought about the non-sensible objects of the traditional metaphysics, we should note that Kant links his new logic not just with metaphysics, but with ontology in particular. This is significant because, for Kant as for his predecessors, ontology itself does not focus on any "particular or specific" kind of object at all, but is instead itself a "general or universal" science, as metaphysica generalis, i.e. as a science whose domain is simply that of "things in general" (Dinge überhaupt), without qualification.13 Indeed, throughout his work Kant continues to describe the basic elements of his successor-science to traditional ontology as "categories" or "concepts of an object in general [Gegenstand überhaupt]" (B128; my emphasis). This makes it hard to see how such a science could nevertheless go on to focus on only one species of objects, whether this species consists in the sensible objects in space-time or the non-sensible objects at issue in special metaphysics, as Tonelli and Wolff would appear to have us think. Rather, Kant's use of 'überhaupt' suggests instead that transcendental logic will operate at a level that abstracts from these sorts of differences between kinds of objects, focusing instead on what is common to all species of objects, by considering what pertains to thinking an object "as such."

There is reason to think, moreover, that it is precisely the generality of its subject-matter that motivated Kant to choose the label 'transcendental logic' in the first place, given the significance that many metaphysicians, from the Scholastics onwards, had accorded to the term 'transcendental' itself. In Part II of Spinoza's Ethics, for example, Spinoza identifies "entity [ens], thing [res], something [aliquid]" as "transcendental terms" (termini transcendentales), because they are thought to apply to everything "without any distinction" (sine ulla distinctione, cf. IIp40s1). Similarly, in Alexander Baumgarten's 1757 Metaphysica, the textbook Kant himself used in his metaphysics lectures, the author identifies "one" (unum, §73), "true" (verum, §90), "perfect" (perfectum, §99), and "good" (bonum, §100) as predicates which are true of "every entity" (omne ens) in a "transcendental" sense (transcendentaliter); what is more, Baumgarten claims that these predicates are "absolutely necessary" of every entity (§116 et seq.).14 In fact, Kant himself connects just these [End Page 423] traditionally "transcendental" predicates with the pure concepts or categories of his own transcendental logic in §12 of the first Critique—and so, only a few pages after first introducing the categories—claiming that his new discipline will itself be able to better accommodate the insight behind "the proposition, so famous among the scholastics: quodlibet ens est unum, verum, bonum" (B113).15

This conclusion is further encouraged by the unique standing that Kant ultimately attributes to the categories in the early sections of the Transcendental Logic. These pure concepts are what, for Kant, make up "the elements of the pure cognition of the understanding and the principles without which no object can be thought at all" (B87; my emphasis). This is surely what lies behind Kant's decision to label the pure concepts as "categories" in the first place, and so in this way take over Aristotle's label for the most basic kinds of entities, since it is "by these concepts alone can our understanding understand something in the manifold of intuition, i.e. think an object of it" (B106; my emphasis). As Kant claims later in the Transcendental Analytic, "[W]e cannot think any object except through categories" (B165; my emphasis; cf. A97). This standing of the categories within our thoughts, therefore, would seem to parallel quite directly the standing that we have just seen Kant ascribing to the principles of the traditional logic. But then, because transcendental logic, no less than the traditional logic, would thereby provide a necessary condition for any case of thinking and understanding, the new logic should also be unrestrictedly general in its scope.16

If, finally, we add to this the simple fact that Kant himself does not ever directly classify transcendental logic as a special or particular logic in any of his published writings,17 the list of reasons for suspicion of the domain-subordinative interpretation grows quite long indeed.

Even deeper grounds for worry arise for domain-sensitive interpretations more generally once we turn to Kant's frequent identification of the elements of his new [End Page 424] logic (the pure concepts) with the basic elements of the traditional logic, what Kant calls the forms of judgment (see B95).18 In fact, in the B-deduction Kant goes so far as to claim that there is a "complete coincidence [Zusammentreffung]" between the categories of transcendental logic and the traditional logic's forms (B159). Yet if the unrestricted generality of the traditional logic entails that these forms are such as to figure in every act of understanding (thought, judgment), and if the categories really are in some sense identical ("coincident") with these forms, then the very idea of a species of thinking and judging that involves the forms but somehow does not involve the categories as well would seem to be a non-starter.19

§4. From a Difference of Domain to One of Aspect: The "Introduction"

In the previous section we saw that both the domain-exclusive and the domain-subordinative interpretations face significant obstacles. Fortunately, neither interpretation is forced on us by Kant's texts. Rather, as I will now show, if we revisit the sections of the first Critique in which Kant first introduces and defines transcendental logic—in particular, to §§I-IV of the Introduction to the section of the first Critique entitled "Transcendental Logic," as well as the sections that immediately follow, sections which contain what Kant calls the "Metaphysical Deduction" (i.e. those numbered §§9-12 in the B-edition)—we will find an entirely different line of thought emerge, one that shifts our focus away from questions of sensitivity to domain or scope altogether. Let us begin by looking to the Introduction in this section, and turn to the Metaphysical Deduction in the following.

In §I of the Introduction, Kant claims that logic is "the science of the rules for our capacity for understanding in general" (B76). Our understanding is a capacity for thinking, which Kant defines as "cognition through concepts" (B94).20 Kant thinks that the paradigmatic case of thinking is the cognition through concepts in judgment. He famously claims that judgment is that which we can "trace back all acts of understanding" (B94).21 Logic itself, then, could equally be described as the science of thinking and judging "in general." [End Page 425]

It is shortly after these introductory remarks that Kant introduces the two further divisions within logic that are crucial for our purposes here: first, the division between a "general or universal [allgemeine]" logic and a "special or particular [besondere]" logic, and then, secondly, a separate division between a logic that is formal and one that is not. We have already met with the general/special distinction above, so our treatment of it here can be brief. Kant recognizes that we can make different kinds of judgments, that we can put our understanding to use in many different ways in thinking, and that there are many different kinds of objects we can represent in our thoughts and judgments. For this reason, Kant thinks that logic, as the study of our capacity of understanding, is itself something that "can be undertaken with two different aims": either as an account of what he calls "the general or universal [allgemeine] use" of understanding, or as an account of one or another "special or particular [besondere] use" of the understanding (B76). The former account of understanding is called "general logic," and "contains the absolutely necessary rules of thinking, without which no use of the understanding takes place" (B76). Kant takes this to imply that general logic "therefore concerns these rules without regard to the difference of the objects to which it may be directed" (B76). The logic of a special use of the understanding would thus be called a "special logic," one which, by contrast, "contains the rules for correctly thinking about a certain kind [Art] of objects" (B76; my emphasis), as we have already noted above.

Now, domain-subordinative interpreters are surely right to note that Kant makes it clear in §I of the Introduction that he takes the core of the traditional logic to consist in the provision of a general logic in this sense: it concerns itself with just those conditions that are necessary for any thinking at all to take place, just those features that remain constant no matter what particular kind of object is being thought about. For this reason, Kant also thinks that this core could be called "elementary logic" (B76), because it gives the most basic characterization of any act of understanding. A special logic, by contrast, would be "the organon of this or that science," and so not elementary for all uses of understanding (B76). An example of a special logic would seem to be provided by mathematics, insofar as Kant consistently identifies it as an "organon" and takes it to "contain the ground for the expansion of our cognition with respect to a certain use of reason," but not necessarily others (see Jäsche Logik §I, 9:13; my emphasis).22

Yet though Kant begins his discussion in §I by highlighting the generality of the traditional logic, before he goes on to actually introduce his new logic, Kant first points out a further peculiarity of the treatment of our understanding by traditional logic. This begins to emerge toward the end of §I of the Introduction [End Page 426] and more emphatically at the outset of §II. As Kant sees it, rather than simply abstracting from the different kinds of objects to which our thinking can be directed, the traditional logic goes further and "abstracts from all relation [Beziehung] of cognition to the object" altogether (B79; my emphasis). Since Kant takes a cognition's "relation [Beziehung] to its object" to be what makes up its "content" (Inhalt, B83; cf. B79), the result is that, in Kant's view, the traditional logic has actually been treating thinking in complete abstraction from anything having to do with its content, and instead "treats only the form of thinking in general" (B79). For this reason, Kant later describes the traditional logic itself as a "merely formal logic" (B170; my emphasis).23

Although the need to distinguish between general and special logic is something that already had some currency within the textbooks of Kant's time,24 the characterization of the traditional logic as formal was much less common in Kant's day. Some have argued, therefore, that it is Kant's own emphasis on this feature that is responsible for the characterization of logic as a "formal" discipline even to this day.25 Even so, at least prima facie reasons for such a characterization are [End Page 427] not hard to find among Kant's predecessors. In Christian Wolff's 1728 Philosophia rationalis sive logica, for example, we find judgments being depicted in the familiar schematic way: Omne A est B, Quoddam C est A, and so on (see §§380 et seq.), with 'A,' 'B,' and 'C' indeterminately indicating places to be filled by some subject-concept and some predicate-concept, but not singling out any in particular. Leibniz gives a similarly schematic depiction of the syllogistic figures in his Nouveaux Essais (see G IV.17.4, G V.461f), and even refers to them as "formal arguments [arguments en forme]" because they are arguments that "conclude by force of the form" alone (G V.460-61). This way of capturing the logical treatment of judgment would seem to be implicit as well in the second edition of Baumgarten's 1773 Acroasis logica, where Baumgarten refers to the concepts in a judgment as its "material parts" (§207), with the copula functioning as its "formal part" (§208).

This last use of "form" and "matter" is, in fact, repeated by Kant himself in the first Critique (see B312). And because the copula itself is something that was traditionally identified throughout the early modern period (e.g. by Locke and by the authors of the Port Royal Logique) with an activity of combination on the part of the mind,26 Kant's own alignment of logical forms like the copula with the "logical functions" that guide the activity of understanding in judging (see B95) would not have seemed especially unnatural either. Hence, even if describing it as a "merely formal" discipline might have struck Kant's readers as a novel way of construing the nature of the traditional logic, it should not have seemed unwarranted.

In any case, it is only at this point in his discussion of the traditional logic—i.e. in §II, the second section of the Introduction, immediately after drawing our attention to the implicit restriction of its focus to the form of acts of thinking and its abstraction from the content of thoughts (see B79f)—that Kant introduces the possibility of a new kind of logic, the one that he will ultimately call "transcendental" logic. What is more, the key feature that Kant here picks out as distinguishing the new logic from the traditional logic is not that the new logic will be concerned with a "particular" use of understanding, but instead precisely that the new one will be "a logic in which one did not abstract from all content of cognition" (B80; my emphasis)—i.e. will not be "formal" in the relevant sense.

Hence, not only does Kant choose to not even mention the idea of the new logic in the section in which he does explicitly draw our attention to the (at the time more familiar) general logic/special logic distinction (i.e. in §I), Kant introduces this idea instead only after highlighting the (relatively unfamiliar) distinction he wants to make between treating thinking as to its form rather than as to its content. And in the section in which Kant does introduce transcendental logic (i.e. in §II of the Introduction), the general/special distinction does not make any explicit reappearance whatsoever. The form/content distinction, by contrast, is front and center from the outset. [End Page 428]

Now, the distinction between the form of A and its content is neither a distinction between a genus and a species of A (as the domain-subordinative interpretation would require), nor is it a distinction between two kinds of A (as the domain-exclusive interpretation would have it). Rather, it is a distinction between two abstract parts or aspects of one and the same thing—namely, A itself—such that, whenever there is an A, both aspects (form and content) are present. But if this is so, then disciplines that are distinguished by whether they focus on the form or the content of A would not also be distinguished by the domain of A to which they are related, since they are each focused on something present in every A. Hence, Kant's own clear focus on the form/content distinction in the Introduction as the principle for the distinction between the two logics strongly suggests that any attempt to distinguish the two logics in terms of domain has gotten off on the wrong foot.

§5. The Distinction Between Aspect Sin the "Metaphysical Deduction"

It is the form/content contrast that is also clearly in focus when Kant moves on to tell us how the elements of the new logic are to be discovered in the course of §§9-12, i.e. the sections that include what Kant will later call the "Metaphysical Deduction" of the pure concepts (B159). The key thought underwriting this process is captured in one of the most frequently cited passages in the first Critique:

The same understanding, therefore, and indeed by means of the very same actions [Handlungen] through which it brings the logical form of a judgment into concepts . . . also brings a transcendental content [Inhalt] into its representations by means of the synthetic unity of the manifold in intuition in general [überhaupt] . . . .

(B105; my emphasis)

As we have seen, Kant thinks that what the traditional logicians had focused on is the capacity of our understanding to combine representations in activity that takes certain familiar forms and, in Kant's terms, is guided by certain "functions." Yet with their focus on the form of combinatory activity, Kant thinks that logicians have systematically neglected the fact that "in one and the same act," our understanding itself introduces a "content," a "relation to an object" (see B83), into our representations. It does this by taking a particular way of combining representations to be necessary because of the way things stand with what is thereby being represented, i.e. with an object that could be given in some intuition. For example, by combining representations S and P according to the categorical form of subject and predicate in "S is P," we are at the same time representing S itself as a substance that bears the property or accident P. Kant's key thought here is that both the form of our representing in judgment and part of the content of our judgment itself (part of the distinct relation that our representing bears to its object) have one and the same ground.

In fact, as we have already anticipated above (in §3), Kant goes further than this. In the sentence prior to this quote, Kant claims that both tasks are accomplished by one and the same function (see B104-5). This is what underwrites Kant's claim that the traditional logical forms of acts of combination ultimately stand in [End Page 429] "complete coincidence" with this "transcendental" content of understanding, as he puts it later in the B-Deduction (B159; my emphasis). For each form of combining identified by the traditional logic, then, Kant thinks there will be a distinctive relation to an object that is thereby achieved. When viewed as to the object-relatedness they institute, these basic logical forms of combination (e.g. categorical judging) can be seen to introduce basic contents, basic ways of representing objects (e.g. substance-accident). These are what Kant calls the "categories" of understanding, or the "pure concepts of an object überhaupt."27 This is how, and why, Kant thinks that the forms of the traditional logic will provide us with a key ("clue" [Leitfaden]) to the "discovery" of the categories (see B91f), and point the way to the possibility of the new logic.28

Such a "complete coincidence" of the elements of the two logics is an awkward one for either of the domain-sensitive interpretations to accommodate, because it again strongly suggests that wherever one element is present, so too is the other. If, by contrast, we reorient our conception of the difference between the two logics by seeing it as a question of the difference in aspect that is in focus, the "coincidence" that lies at the heart of the Metaphysical Deduction can be taken perfectly in stride, since the aspects in question (form/content) are such that the one is present wherever the other is found.

At this point, however, a worry might arise, one that might seem to push us back toward certain versions of domain-subordinative interpretations. This is the worry that in order to talk about the content of an act of understanding at all, transcendental logic will need to incorporate a more determinate reference to our sensibility than our discussion in the previous section seemed to allow. This is because the very idea of our understanding itself "introducing" content into its own representations might seem to be ruled out, since at times Kant appears to claim that our intuitions are the only thing that can supply "content" to our thoughts (see B175). Indeed, immediately after Kant insists that transcendental logic effect the "complete isolation and separation" of the understanding from sensibility that we noted above, he goes on to say that "without intuition all of our cognition would lack objects, and therefore remain completely empty" (B87; my emphasis). What is more, just before the remark at B105 cited above, Kant reminds us that at this stage in the Critique, "transcendental logic has a manifold of sensibility that lies before it a priori, which the Transcendental Aesthetic has offered to it" (B102). Does not this imply that transcendental logic does, after all, have as its domain only a subset of thinking in general—namely, thinking that has "sensible" content?

In fact, such a conclusion is not forced on us. We can see why it is not if we recall Kant's official demarcation of the subject-matter of the new logic: [End Page 430]

In a transcendental logic we isolate the understanding (as we did above with sensibility in the Transcendental Aesthetic), and bring to the fore, out of our cognition, merely the part of our thinking that has its origin [Ursprung] solely in the understanding.

(B87; my emphasis)

Here Kant claims explicitly that the logic will not consider the "parts" or aspects of our thinking whose "origin" lies outside of our understanding. Because Kant takes our understanding to be distinct from our capacity for intuition (sensibility), this entails that the subject-matter of transcendental logic must "separate itself completely . . . from all sensibility," as Kant puts it shortly thereafter (B89; my emphasis). This is so, even though at this point in the progression of the first Critique, Kant thinks that we have discovered that there is a certain a priori "content" that is available to the mind by being "contained" in sensibility itself and representable in "pure" intuition—namely, a "pure" spatial and temporal manifold (B102). Despite this, Kant is clear that transcendental logic itself must abstract away from this content as well, because it "belongs to the conditions of receptivity of our minds" and does not have its origin in the understanding itself (B102). The genuine "elements" of transcendental logic, Kant tells us, must "belong not to intuition and to sensibility, but to thinking and understanding" (B89).

With this in mind, we can now see that Kant does, in fact, mean to attribute the introduction of the particular "transcendental content" that transcendental logic will investigate to the understanding alone. This is not, however, because Kant now means to say that our understanding is capable of presenting any actual object "immediately" before the mind. Kant is quite clear that he thinks our understanding is not capable of intuiting anything (B93). Kant alludes to this fact in the very passage at B105 itself by acknowledging that the thing that is to be unified by the synthetic activity of understanding is whatever manifold it is that is ultimately given "in intuition." Yet because transcendental logic views our understanding in "complete separation and isolation" from sensibility, it cannot determine which kind of manifold it is that will, in fact, be given in our intuition. This is not just because we cannot anticipate this manifold in all of its concrete empirical determinateness, but rather because from the point of view of transcendental logic, we cannot even anticipate which of many forms of sensibility that Kant admits may be possible will be the one that our own (human) species of sensibility will take.29 For this reason, transcendental logic itself can only make indeterminate reference to the connection of the understanding's activity to intuition as such. This is what lies behind Kant's description in B105 of the manifold at issue as a manifold of "intuition überhaupt."30

Kant's main thesis here, then, is that when we take up the transcendental-logical point of view and view the traditional-logical forms as basic ways of unifying whatever manifold is ultimately given, we are already able to identify, in anticipation as it were, a part of the relation to an object—and hence, the content—that such [End Page 431] combination will represent, no matter what the species of sensibility will be. This feature of the content, in turn, is attributable of our understanding alone. Hence, despite our mind's dependence on our sensibility to relate us immediately to objects, and so, in this more demanding sense, "have" content, our understanding is capable of relating us (representationally, intentionally) to objects "mediately" all by itself.31

It is this ability that lies behind Kant's description of our understanding as a "spontaneity of concepts" (B74; my emphasis): understanding is "the capacity to bring forth representations itself [Vorstellungen selbst hervorzubringen]" (B75; my emphasis). It is also this capacity that makes transcendental logic itself the "science of pure understanding . . . by means of which we think objects completely a priori," a science at the center of which are "concepts that may be related [sich beziehen] to objects a priori . . . as acts of pure thinking," concepts "of neither empirical nor aesthetic origin [Ursprung]" (B81). The "origin" of these concepts is purely intellectual, as is their content.

Now, as I noted above (cf. §3), it is hard to see how Kant's emphasis on the complete isolation and separation of the subject-matter of transcendental logic from sensibility is consistent with versions of the domain-subordinative interpretation that would have transcendental logic nevertheless restrict its focus to just those objects that can be given through our sensibility.32 Indeed, far from keying us into that domain of objects, Kant's claims here point rather explicitly in just the opposite direction: in transcendental logic, the particular nature of our sensibility must be put into brackets, with the focus instead being on content that arises from the understanding itself.33

These considerations, finally, help to flesh out why the rules of transcendental logic will parallel those of the traditional logic, as was claimed above in §3. Recall [End Page 432] that Kant takes the traditional logic to contain the "absolutely necessary rules for thinking, without which no use of understanding would take place" (B76). With its focus on the categories, transcendental logic is also tasked with specifying equally necessary conditions for thinking—albeit as to its content. For in specifying the basic concepts through which it is able to "understand" anything at all, transcendental logic specifies the basic categories of being an object of our understanding as such.34

§6. Generality in Kant's Doctrine of Conceptual Content

In the previous two sections, I have presented a series of textual considerations from the early parts of the section entitled "Transcendental Logic" (i.e. the Introduction and the Metaphysical Deduction) that push strongly against construing the difference between the two logics in terms of a difference in domains. Rather than focusing on either the contrast between general and special uses of understanding, or the contrast between analytic and synthetic judgments, these texts suggest instead that Kant means for the contrast between form and content to be the axis along which transcendental logic will be distinguished from the traditional logic. Because this contrast is one best understood in terms of abstract parts or aspects of the same thing, rather than in terms of the partitioning of domains, these texts imply that the two logics ultimately do not differ in their domain at all, but rather "coincide."

In these next two sections, I want to turn to a second, more systematic argument toward the same conclusion. More specifically, I want to show that it follows from Kant's doctrine of conceptual content that the pure content of the categories is, in fact, universally and necessarily involved in every act of understanding. In order to see why this content must enjoy such universality, however, we will first need to understand a bit more about Kant's views on the nature of conceptual content in general.

As we have noted above, Kant characterizes the "content" (Inhalt) of cognitions in general as their particular "relation to an object" (B83). The particular kind of relation to objects that pertain to conceptual cognitions is "mediate," one achieved by means of a relation to a "mark" that several objects might have in common.35 Now, as Kant makes clear in Prolegomena, the content of a concept can also be characterized as what is "thought" (gedacht) in the concept, and therefore as what is brought to light or "clarified or elucidated" in the analysis of concepts and set [End Page 433] forth in analytical judgments.36 In his logic lectures, Kant describes the content of a concept as what is "contained [enthalten] in" the concept, something which he identifies with a set of simpler representations, all of whose contents combine to make up the initial concept's content.37 To take a familiar example, the content of the concept <human> is traditionally taken to be composed out of the contents of the concepts <rational> and <animal>. For this reason, Kant would say that the concept <human> "contains" the concepts <rational> and <animal> "in" itself, and also that the concepts <rational> and <animal> are already "thought" as well in any thought that involves the concept <human>.38

Ultimately, Kant thinks that these containment relations among concepts form a hierarchy, along the lines of what has come to be known as the Tree of Porphyry.39 Concepts that have a given concept <B> in their content are said to be contained under <B>, making <B> itself "higher" than those concepts that contain <B> in them, and making these concepts "lower" than <B>, which then contains them under itself.40 In our above example, then, because <rational> and <animal> are both higher relative to <human>, <human> would therefore be contained under each of them, in each concept's "extension" (Umfang) or "sphere" (Sphäre), while nevertheless itself containing both <rational> and <animal> in itself, as parts of its own content.41 [End Page 434]

Now, because a concept is defined as a "general [allgemeine] representation," in the sense of representing a "mark" that can be "common" (gemein) to several things (see Jäsche Logik §1n1, 9:91 and B377), Kant thinks that every concept is higher relative to at least some further representation. Like many before him, however, Kant thinks that there is an upper limit to the progression to higher concepts. As Kant makes clear in his logic lectures, if we recursively pursue the analysis of concepts, looking for what concepts each of the concepts contained in a given concept <B> themselves contain, and so on, eventually we will come to an end. That is, Kant thinks that we will come to a conceptual content that cannot itself be further analyzed, a concept "which is contained under no other, is not a partial concept, i.e. has no further part [Teil]" (Logik Dohna-Wundlacken, 24:755; my emphasis). When viewed in relation to the rest of the hierarchical structure, this will be the conceptus summus.

Now, because this concept will be the highest concept, it will not only be contained under no other, it will also contain all other concepts under itself, in its extension. But then, given what we have just seen about the nature of the containment hierarchy, it follows that this highest, simplest, concept will itself be contained in every other concept, as one of its parts. And since the content of any given concept is constituted by those higher concepts that are contained in it, it follows as well that this highest concept is a part of the content of every other concept. But this is just to say that there could be no concept whose content did not include this highest concept. In this sense, then, the highest concept stands as a conditio sine qua non for conceptual content as such: something simply is not a concept if it is not something whose content includes the highest concept.

What, then, is the highest concept? Kant's thesis is that it is nothing other than "the concept of an object in general" (Begriff von einem Gegenstande überhaupt; my emphasis), as he claims in both the first Critique as well as in his logic lectures.42 As we shall see in a moment, Kant takes this concept to be higher than the abstract metaphysical categories <substance>, <reality>, <existence>, and higher even than the concepts <something> and <nothing>. This points up the fact that Kant means 'object' here in a very abstract, generic sense indeed. In fact, in Kant's metaphysics lectures from the 1790s, he identifies the "highest concept," "object of thinking" (Object des Denkens), with the concept of "something in the logical sense" (aliquid in logico sensu, Metaphysics L2, 28:552; my emphasis).43 But this is simply because Kant means for it to be generic enough to apply to anything at all which can be [End Page 435] the subject-matter of a thought. 'Object,' therefore, is here meant to function as the label for the genus summum within which are grouped all things that can be a topic (a logical "object") of thinking—which would seem to be absolutely everything whatsoever.

To be sure, the content that this highest concept itself has is exceedingly thin, as it is the most abstract thing we can think. What is more, being the highest, this concept itself cannot "contain" any further concept "in" itself, for this would imply that it itself would be contained "under" something else. But this implies that, strictly speaking, <object> does in fact not have any content in this specific sense. Nevertheless, in thinking the concept <object> we are thinking something and not nothing; we are thinking something rather than not thinking at all. We are thinking the primitive positive "mark" that is borne by every object.44

Once Kant's identification of <object> as the highest concept is conjoined with the just-rehearsed line of reasoning, however, it follows at once that Kant is committed to a view according to which the content of the concept <object>—however minimal it may be—is "contained" in the content of all other concepts, that it is already "thought" in all other concepts.45 But then with the concept <object> we have found something that fulfills Kant's description of the pure content that pertains to the categories—i.e. a concept without which we cannot "understand" anything at all. Yet this means that the content of the concept <object> therefore possesses a universality that is as unrestricted with respect to thinking and understanding as the logical forms identified by the traditional logic. The concept <object>, therefore, provides us with the beginnings of a subject-matter for a parallel "general or universal" science of the content of understanding—in effect, a general logic that would not be "merely formal," i.e. a transcendental logic.

§7. The Generality of the Content of the Pure Concepts

With this we can now finally pick back up our original line of thought concerning the content of the pure concepts themselves. Or rather, more precisely put: we are now in a position to see that the initial line of thought concerning the pure concepts and the more recent line of thought concerning the concept <object> are in fact following out one and the same thread in Kant's system. For, as the attentive reader will have already noticed, it is precisely the pure concepts (categories) of understanding themselves that are described as "concepts of an object überhaupt" (B128; my emphasis). This fact alone strongly suggests that Kant takes the pure concepts themselves to be, or at least to be among, the highest concepts that there are. In light of the conclusions just reached, it is now evident that this would, in turn, imply that Kant takes the pure concepts to contribute to the content of all other, "lower" concepts—which is to say, all other concepts whatsoever.

In fact, Kant is quite straightforward about his intention to link his description of the categories as "concepts of an object überhaupt" with the claim that they are among the "highest" concepts. At the end of the Amphiboly in the first Critique, [End Page 436] Kant argues explicitly that ontologists have been wrong in thinking that the highest concepts are the concepts of "the possible" and "the impossible," because these concepts represents two sides of the "division" (Einteilung) of a still-higher concept—namely, "the concept of an object überhaupt (taken problematically, leaving undecided whether it is something [Etwas] or nothing [Nichts])" (B346). Instead, Kant claims here again that it is this concept (i.e. <object>) that is the absolutely highest concept. As Kant makes clear in this passage, both everything that is "something" and everything that is "nothing" is nevertheless an "object" in the relevant sense.46

Now, in this passage Kant admits that the division into <something> and <nothing> represents one of the highest or first divisions of this highest concept. The determination of this division, however, is something that Kant takes himself to have already laid the groundwork for, if not already accomplished, in his own analysis of the categories in the earlier parts of the Transcendental Analytic: "[B]ecause the categories are the only concepts that relate to objects in general [die einzigen Begriffe sind, die sich auf Gegenstände überhaupt beziehen], the distinction of whether an object is something or nothing must proceed in accordance with the order and guidance of the categories" (B346; my emphasis). Yet for the categories to provide order and guidance for the further division between the concepts <something> and <nothing>, the categories would need to already be in place on the hierarchy. This, in turn, would imply that Kant takes the division of the absolutely highest concept <object> into the various categories of objects— i.e. into <substance>, <accident>, <cause>, <effect>, etc. (see B106)—to be itself a higher division than that into <something> and <nothing>. This is confirmed by the fact that Kant takes the "division" of the extensions of both <something> and <nothing> to be determined according to the very same categories (see B348), since this implies that both <something> and <nothing> themselves fall under the categories, with the categories themselves therefore being "common" to all of the things that fall under both concepts.

The division of the sphere of the concept <object> into the plurality of pure concepts would thus seem to be among the very highest divisions possible. To be sure, given that there are twelve such pure concepts, difficult questions arise here about what the intermediate steps must be, and also about the order in which these steps should transpire, for the division of the concept <object> to yield just the structure that Kant identifies on the famous Table in §10 of the first Critique.47 It [End Page 437] is also possible, first, that the twelve categories themselves cannot be (uniquely) "ordered" under one another, but are instead each representative of equally basic "marks,"48 or, secondly, that the categories are not exclusive of one another in the way that may seem requisite of a proper conceptual division.49 There are, finally, the perennial questions about why just these twelve categories rather than others, questions that Kant purports to address in the so-called Metaphysical Deduction of the categories from a privileged set of twelve traditional-logical functions of judgment, though hardly anyone has been pleased with this part of Kant's account.50

Yet however these questions are addressed—and, indeed, whether or not Kant says enough to develop a definitive, let alone coherent, answer on his behalf—the more important point for our purposes here is this: in Kant's system, every other concept besides the categories is not only (i) lower than the concept <object>, and (ii) lower than whatever concepts may lie hidden behind the higher divisions encapsulated by the Table, but, ultimately, is (iii) lower than at least one (if not more) of the twelve categories or pure "concepts of an object überhaupt" themselves. In short, though it is true that no single one of the pure concepts is itself the highest concept, they are higher than all other concepts not recorded by divisions on the famous Table itself.51

If this is correct, then we can put to work the reasoning just outlined in §6 to yield the following series of conclusions. It should follow, first, that all other concepts are "contained under" the pure concepts, in their extensions. Secondly, it should follow that the pure concepts must be "contained in" all other concepts without restriction. It should follow as well, therefore, that some of the pure content presented on the Table itself must belong as a part of the content of every other concept. But then, since there can be no case of thinking that does not involve at least some concept—because, as we have already seen, thinking is defined as "cognition through concepts" (see B94)—the universality of pure content with respect to all thought is secured.52 [End Page 438]

Let me conclude this section by showing how the generality of the categories is connected to one of his other more widely recognized doctrines. As several commentators have rightly insisted, any successful interpretation of Kant's views will need to leave room in his system for what have come to be called the "pure" or "unschematized" categories, or the categories considered and used independently of their connection or schematization to our particular form of sensibility.53 A simple motivation behind carving out such a content for the categories that is not dependent on sensibility comes from the fact that Kant explicitly ascribes such content to them.54 A second reason stems from Kant's recognition (noted above) that a discursive understanding like ours, operating with the same categories, could be conjoined with forms of sensibility different from our own. A third, also simple, motivation lies in the fact that some such content would need to accrue to the categories "in themselves," as it were, simply in order to be distinguished from one another, for it is hard to see how else the categories, e.g. of reality and negation, could be distinguished if not at least in terms of what is "thought in" each concept.55 [End Page 439]

A fourth, more systematic motivation lies in the essential role that Kant takes such content to play outside of the context of theoretical sensibility altogether. As the second Critique especially makes clear, Kant needs the categories to have a content that can be involved in our practical judgments and reasoning—indeed, to make sense of practical philosophy as a whole—insofar as this involves representing things which are intrinsically non-sensible (e.g. a good will). Without this pure content, Kant's use of the very same "categories of understanding" in the second Critique's derivation of the pure concepts of an object of pure practical reason could not get off the ground (see 5:65). There Kant insists explicitly that though they are subordinate to the categories (in particular, they are "modes" [modi] of the category of causality), these pure practical concepts are not meant to conform to "a theoretical use of the understanding" in which we "bring a priori the manifold of (sensible) intuition under one consciousness," but instead are meant to consist in the basic ways in which we are "to subject a priori the manifold of desires to the unity of consciousness of a practical reason commanding in the moral law, or of a pure will" (5:65). In fact, in this last function, Kant appears to attribute to the "categories of understanding" a content that, if not entirely disconnected from a relation to intuition in particular, is at the very least orthogonal to it.

This line of thinking actually suggests that Kant means for one and the same "categories of understanding" to function as the ground for the derivation of both the basic concepts of nature (concepts of an object of intuition) and the basic concepts of freedom (concepts of an object of desire). For this to be possible, these categories could not be intrinsically restricted only to either a "use" in relation to objects that can be given in intuition or even to a "theoretical" or "speculative" use. Rather, their pure ("transcendental") content—and with it, the subject-matter of transcendental logic itself—must be intrinsically even more generic than this as well, in order to be present in both of the basic kinds of uses of our understanding.

On the domain-subordinative interpretations of transcendental logic—especially those that restrict the subject-matter to thought about objects in space and time—the possibility of deployments of the same transcendental, pure contents of our understanding in contexts beyond that of theoretical knowledge is greatly obscured, if not eliminated altogether. By contrast, the unrestricted interpretation of the subject-matter of transcendental logic I have developed here handles these cases with ease. By recognizing a genuinely universal science of the categories as such, we are therefore better placed to accommodate the fact that Kant thinks that these same categories have not just an unschematized significance, but have an altogether distinct practical—and non-theoretical—"use."

§8. Concluding Remarks: The Transcendental Logic in "Transcendental Logic"

I have argued that the content of the pure concepts is universally present throughout every act of thinking. Since transcendental logic is tasked with specifying the basic concepts through which we are able to "understand" anything at all (the basic categories of being an object of our understanding), transcendental logic should be seen as providing conditions for thinking that are every bit as universal [End Page 440] and necessary as those of the traditional logic. Far from providing the rules for a certain kind of judgment (as the domain-exclusive interpretation would have it) or for thinking about "a certain sort of object" (as would be required if the domain-subordinative interpretation were correct and if transcendental logic were to be a "special or particular logic"), transcendental logic instead has been shown to deal with principles that govern all kinds of thinking and judging, no matter what sort of object is being thought about. This is because transcendental logic specifies a condition without which thinking would have absolutely no content whatsoever, because there simply is no other kind of content that is possible for thinking. To think at all is to cognize, to consciously represent an object, through concepts;56 to think at all is to think about an object. Yet because the generic concept of an object of thought just is the subject-matter of transcendental logic, transcendental logic, no less than traditional logic, provides a conditio sine qua non for any instance of thinking and understanding. Both logics, therefore, are equally and unrestrictedly general in their scope, which implies that their domains must be viewed instead as perfectly coincident. The contrast between the logics is not to be understood in terms of the difference between kinds, or the difference between genus and species, but rather in terms of the difference between aspects of thinking or judgment that are at issue—namely, the difference between the form and the content of understanding.

In closing, let me now return to a feature of the first Critique that I identified above (in §2) as providing particularly compelling motivation for the domain-sensitive interpretations in the first place, in order to show how it can be accommodated on the present interpretation. This is the focus that Kant clearly accords to the principles that govern our thought about sensible (i.e. spatio-temporal) objects within the section entitled "Transcendental Logic." Especially by the beginning of the subsection of the "Logic" entitled the "Analytic of Principles," and especially in the Schematism, the nature of our sensibility and its distinctive forms (space and time) plays an altogether ineliminable role in Kant's investigation. Does this not entail that the discipline of transcendental logic itself is focused on the objects of our sensibility, on what Paton has called the "metaphysic of experience"?

Earlier I sought to counter this point by reminding us of Kant's explicit specification of the nature of transcendental logic: "[I]n a transcendental logic we isolate the understanding . . . and elevate from our cognition merely the part of our thought that has its origin solely in the understanding" (B87; my emphasis); that, in a transcendental logic, "the pure understanding separates itself completely not only from everything empirical, but even from all sensibility," such that "it is therefore a unity that subsists on its own, which is sufficient by itself, and which is not to be supplemented by any external additions" (B89-90; my emphasis). This, I think, makes quite clear that Kant means for this discipline to have a much more sharply delimited subject-matter. And in a discipline with this nature, reference to [End Page 441] the specific findings of an inquiry into our sensibility would seem to be decidedly out of place. Must we therefore ascribe to Kant a kind of confusion?

I think this apparent conflict can be resolved if we allow Kant to be working with a distinction between the discipline of transcendental logic per se and the section entitled "Transcendental Logic." What is contained in the latter is ultimately subservient to the more general project of the "critique of pure reason"—that is, "a critique of the capacity of reason in general, in respect of all the cognitions after which reason might strive independently of all experience" (Axii), i.e. a priori. To be sure, for the project of the critique of our ability to achieve a priori "knowledge" (Wissen), Kant undoubtedly thinks that more than the results of an analysis of our understanding in isolation from sensibility will be necessary. This follows from the fact that intuition is necessary for knowledge, and our understanding is incapable of intuiting anything. This, moreover, is behind Kant's belief that only a cumulative synthesis of the results of a transcendental logic with those of a transcendental aesthetic will be sufficient for a rigorous critique of this sort.

Yet the fact that Kant ultimately undertakes such a synthesis on pages with the heading "Transcendental Logic" should not obscure the possibility that the task of transcendental logic sensu stricto is a more preliminary one—namely, the provision of the analysis of only one of these "elements" of our cognition, in complete abstraction and isolation from the other. For it might very well be that Kant ends up including in the section entitled "Transcendental Logic" things that would not belong to the discipline per se, but are included in light of the overarching goal of the first Critique itself.

Is there any evidence that Kant himself had such a distinction in mind? In fact, Kant appears to draw our attention to just this distinction at the end of the B-edition version of the Transcendental Deduction (and so, at the end of the Analytic of Concepts).57 Kant ends the B-Deduction by remarking that he will here stop his practice of dividing the text into numbered paragraphs, claiming that, while up to this point "we have been dealing with the elementary concepts" (Elementarbegriffe), in what follows—i.e. in the Schematism, the Principles, and the Dialectic—"we will represent their use [Gebrauch]" (B169; my emphasis). What follows, then, is something Kant clearly thinks is of a different order than what has come before.

What is more, Kant's use of numbered paragraphs extends backwards, not just through the early parts of the Transcendental Logic, but all the way back to the [End Page 442] beginning of the Transcendental Aesthetic. On the present interpretation, this grouping makes perfect sense. In the Aesthetic itself, Kant had set out to treat our sensibility in complete separation from our understanding: "In the Transcendental Aesthetic, we will first isolate [isolieren] sensibility by separating out [absondern] all that the understanding thinks through its concepts" (B36). This language, of course, parallels quite precisely how Kant introduces transcendental logic, as Kant himself remarks parenthetically: "[I]n a Transcendental Logic we isolate the understanding (just as we did with sensibility in the Transcendental Aesthetic)" (B87). The continuous numbering through the Aesthetic and the first part of the Logic highlights the fact that each of these two stretches of text is to provide a separate treatment of one of "elements" in abstraction from their relation to the other.

The full-fledged integration of the doctrine of the pure intellectual contents ("elements") with that of the pure sensible contents does not commence until after B169, in the Schematism at the outset of the Analytic of Principles, where the categories are finally provided with distinctly sensible significance (via "time-determinations"; see B177f). This implies, however, that these earlier numbered sections of the Logic have not yet restricted their focus to those objects that can be given to us, or to the "use" of these categories in the achievement of knowledge via our sensible intuition. Rather, these sections are focused solely on what have come to be called the "pure" or "unschematized" categories, which bear an intrinsic connection only to "intuition überhaupt."

Of course, even after the provision of the Schematism, Kant does not limit himself to presenting rules for correctly thinking about only one "certain kind of object"—say, objects that we can sense. For though the rest of the Analytic is devoted to thought of objects that are sensible for us, the Dialectic, by contrast, is concerned with the complement kind of object—namely, objects that we can think, but are, for us, non-sensible: the objects of the pure concepts or "ideas" of reason (B367f). Hence, if we are going to classify any of what transpires under the heading 'Transcendental Logic' as a special logic, it would be better to say that there are, in fact, two special logics on display: the Analytic of Principles provides the special logic for thought about objects of experience, while the Dialectic is a special logic for thought about objects beyond experience.58

What precedes the Schematism, by contrast, is free from the inextricable tie to the particular forms of our intuitions that is present in the later sections. For this reason, it can count as the genuine Elementarlehre for a doctrine of the content of [End Page 443] understanding überhaupt.59 This is because it presents only what pertains to the pure understanding in itself, regardless of the "use" to which it is to be put, whether conjoined to our form of sensibility or another—indeed, whether it is put to a theoretical or a non-theoretical (because practical) use altogether.60

On both textual and systematic grounds, then, I have shown the domain-sensitive interpretations of the nature of transcendental logic to be inadequate in important respects. I have shown, first, that both the natural flow of the Introduction to the Transcendental Logic as well as Kant's strategy in the Metaphysical Deduction strongly suggest that Kant intends the main contrast between the traditional logic and his new transcendental logic to be drawn in terms of the form/content distinction, rather than in terms of exclusive domains or in terms of genus and species. I have shown, secondly, that Kant's account of the conceptual contents at issue in transcendental logic implies that these contents, too, will be universally present in thinking, and so in this respect on par with the forms of traditional logic. Finally, and most recently, I have shown that this account of the nature of the categories fits better with the fact that Kant explicitly takes them—at least in their unschematized form—to be involved even in acts of judgment and reasoning that extend well beyond those acts in which we have theoretical knowledge. Indeed, by emphasizing that the pure content that transcendental logic is concerned with is content that is present in all possible uses of understanding—whether directed at experience or beyond, whether true or false, whether achievements of "knowing" (Wissen) or of mere "believing" (Glauben, Bxxx), whether acts of "determining" (bestimmen) which concept a given object falls under in theoretical judgment, or acts of "making actual" (wirklich zu machen) the object so represented by a practical [End Page 444] judgment (Bix-x)—the present interpretation restores the categories—and with them, our understanding itself—to their rightful place at "the highest point" within Kant's transcendental philosophy as a whole (see B134n).61

Clinton Tolley

Clinton Tolley is Assistant Professor of Philosophy at the University of California, San Diego.

Bibliography and Abbreviations

Adickes, Erich, ed. Kritik der reinen Vernunft. Berlin: Meyer and Müller, 1889.
Adamson, Robert. A Short History of Logic. Edinburgh: Blackwood and Sons, 1911.
Allison, Henry. Kant's Transcendental Idealism. 2nd ed. New Haven, CT: Yale, 2004.
Ameriks, Karl. Interpreting Kant's Critiques. Oxford: Oxford University Press, 2003.
———. Kant's Theory of Mind. 2nd ed. Oxford: Oxford University Press, 2000.
Anderson, R. Lanier. "The Wolffian Paradigm and its Discontent: Kant's Containment Definition of Analyticity in Historical Context." Archiv für Geschichte der Philosophie 87 (2005): 22-74. ["Paradigm"]
Baumgarten, Alexander. Metaphysica. 4th ed. Halle: Hemmerde, 1757.
Beck, L. W. "Did the Sage of Königsberg Have no Dreams?" In Essays on Kant and Hume, 38-60. New Haven, CT: Yale University Press, 1978.
Boswell, Terry. "On the Textual Authenticity of Kant's Logic." History and Philosophy of Logic 9 (1988): 193-203.
Buroker, Jill. Kant's Critique of Pure Reason: An Introduction. Cambridge: Cambridge University Press, 2006. [Introduction]
Cohen, Hermann. Kants Theorie der Erfahrung. 2nd ed. Berlin: Dümmler, 1885.
de Jong, Willem. "Kant's Analytic Judgments and the Traditional Theory of Concepts." Journal of the History of Philosophy 33 (1995): 613-41.
Eisler, Rudolf. Wörterbuch der philosophischen Begriffe. 2nd ed. Berlin: Mittler, 1904.
Friedman, Michael. The Parting of the Ways. Chicago: Open Court, 2000.
Hacking, Ian. "Trees of Logic, Trees of Porphyry." In Advancements of Learning, edited by J. Heilbron, 219-61. Florence: Olschki, 2007.
Heidegger, Martin. Phänomenologische Interpretation von Kants Kritik der reinen Vernunft. Frankfurt: Klostermann, 1977. [Phänomenologische Interpretation]
Heis, Jeremy. "'Critical Philosophy Begins at the Very Point where Logistic Leaves Off': Cassirer's Response to Frege and Russell." Perspectives on Science 18 (2010): 383-408.
Holger, Katharina, et al., eds. Personenindex zu Kants gesammelten Schriften. Berlin: de Gruyter, 1969.
Jungius, Joachim. Logica Hamburgensis. Hamburg: Offerman, 1638. Edited by Rudolf Meyer. Hamburg: Augustin, 1962.
Kant, Immanuel. The Cambridge Edition of the Works of Immanuel Kant. Edited by Paul Guyer and Allen Wood. Cambridge: Cambridge University Press, 1991-.
———. Kants gesammelte Schriften. Edited by the Königlich Preussische Akademie der Wissenschaften. Vols. 1-29. Berlin: de Gruyter, 1902-.
Kemp Smith, Norman. Commentary on Kant's Critique of Pure Reason. London: Macmillan, 1918. [Commentary]
Kinkel, Walter. "Introduction." In Kants Logik, edited by Walter Kinkel. Leipzig: Durr'sche Buchhandlung, 1904.
Körner, Stephen. Kant. London: Penguin, 1955.
Leibniz, Gottfried. Philosophische Schriften. Edited by C. I. Gerhardt. Berlin, 1882. [G]
Longuenesse, Béatrice. "The Division of the Transcendental Logic and the Leading Thread." In Kritik der reinen Vernunft, edited by G. Mohr and M. Willaschek, 131-58. Berlin: Akademie, 1998. ["Division"]
———. "Kant's Theory of Judgment and Judgments of Taste." Inquiry 46 (2003): 143-63.
MacFarlane, John. "Frege, Kant, and the Logic of Logicism." Philosophical Review 111 (2002): 25-65. ["Logicism"]
———. "What Does it Mean to Say That Logic is Formal?" PhD diss., University of Pittsburgh, 2000.
Meier, Georg. Auszug aus der Vernunftlehre. Halle: Gebauer, 1752. [Auszug] [End Page 445]
Parsons, Charles. "Kant's Philosophy of Arithmetic." In Mathematics in Philosophy: Selected Essays, 110-42. Ithaca, NY: Cornell University Press, 1983.
Paton, H. J. "Formal and Transcendental Logic." Kant-Studien 49 (1958): 245-63.
———. Kant's Metaphysic of Experience. London: Macmillan, 1936. [Metaphysic]
———. "The Key to Kant's Deduction of the Categories." Mind 40 (1931): 310-29.
Pozzo, Ricardo. "Kant within the Tradition of Modern Logic." Review of Metaphysics 52 (1998): 295-310.
Prauss, Gerold. Erscheinung bei Kant. Berlin: de Gruyter, 1971.
Reich, Klaus. Die Vollständigkeit der kantischen Urteilstafel. 2nd ed. Berlin: Rostock, 1948. Translated as The Completeness of Kant's Table of Judgments, by J. Kneller and M. Losonsky. Stanford, CA: Stanford University Press, 1992. [Urteilstafel]
Rosenberg, Jay. Accessing Kant. Oxford: Clarendon, 2005.
Ryckman, Thomas, and Michael Friedman. "Analytic and Continental Traditions." In The History of Continental Philosophy, vol. 3, edited by A. Schrift, 111-48. London: Acumen, 2010.
Strawson, Peter. The Bounds of Sense. London: Methuen, 1966.
Tonelli, Giorgio. Kant's Critique of Pure Reason within the Tradition of Modern Logic. Zürich: Georg Olms, 1994. [Tradition]
Ueberweg, Friedrich. System der Logik. 4th ed. Bonn: Adolph Marcus, 1874.
Walsh, W.H. Kant's Criticism of Metaphysics. Edinburgh: Edinburgh, 1975.
Wolff, Christian. Philosophia prima sive rationalis. 2nd ed. Frankfurt, 1737.
Wolff, Michael. Vollständigkeit der Kantischen Urteilstafel. Frankfurt: Klostermann, 1995. [Vollständigkeit]
Zabarella, Jacopo. "De Natura Logicae." In Opera Logica, vol. 2. Cologne, 1597. Reprinted in Jacobi Zabarellae Opera logica, edited by Wilhelm Risse. Hildesheim: Olms, 1966. [End Page 446]


1. Their most visible, and perhaps most notorious, role is the function Kant assigns them to play in the systematic organization ("architectonic") of both the metaphysics of nature and that of morals; see Metaphysical Foundations of Natural Science, 4:474; and Critique of Practical Reason, 5:65. Throughout I will cite Kant's works according to the Akademie Ausgabe volume number and pagination (Kants gesammelte Schriften), with the exception of the first Critique, which I will cite according to the B-edition pagination, save for places where the passage is only to be found in the A-edition. All decisions on translations throughout are my own, though I have consulted and usually followed the Cambridge Edition translations (Guyer and Wood, eds.) when available.

2. This, in turn, will help provide necessary background for the historical analysis of developments in the philosophy of logic in the nineteenth and early twentieth centuries that were heavily influenced by Kant's insistence that logic cannot rest content with being "merely formal." Kant's influence here is felt not just within the tradition of German Idealism (e.g. in Fichte) and in later neo-Kantian philosophy (e.g. in Natorp and Cassirer) but also in the phenomenological tradition as well (e.g. in Husserl), and especially in the neo-Kantian and phenomenological criticisms of the now-classical logic initiated by Frege and Russell, insofar as these criticisms were posed precisely in terms of the failure of the new logic to take the full measure of "transcendental"-logical considerations. On this see Friedman, The Parting of the Ways, as well as Ryckman and Friedman, "Analytic and Continental Traditions," and also Jeremy Heis, "'Critical philosophy Begins at the Very Point where Logistic Leaves Off': Cassirer's Response to Frege and Russell."

3. In his System der Logik, Ueberweg takes Kant to align the "analytical formation of judgments" with "formal logic, in the sense of that which presents the norms of analytical cognition," whereas the "synthetic formation of judgments" is aligned instead with "the critique of pure reason, which asks after the possibility of universally valid synthetic cognition" (§2, 4). Similarly, in his Kants Theorie der Erfahrung, Cohen claims explicitly that "the species of judgment that is in view in formal or general logic" is that of "analytic judgments" (242). Kemp Smith follows the same line in his Commentary, where he also draws the distinction between the two logics in terms of the difference between "analytic" and "synthetic thinking" (see 176, 182f). Kemp Smith actually provides a corresponding chart depicting the exclusivity of the domains of the logics (see 176), a chart that he takes from an editorial note in Erich Adickes's 1889 edition of the first Critique, where Adickes places transcendental logic and the traditional (formal, pure general) logic on distinct sides of a family tree of logics; see Adickes, Kritik der reinen Vernunft, 100n. For another interpretation along similar lines, see Walter Kinkel's "Introduction" to his edition of Jäsche's Logik, viii-xi.

4. See Klaus Reich, Urteilstafel; and Paton, "The Key to Kant's Deduction of the Categories," Metaphysic, and "Formal and Transcendental Logic."

5. In his "Formal and Transcendental Logic," Paton claims explicitly that traditional logic is "more general" and "more abstract than transcendental logic" (247).

6. See Tonelli, Tradition; Wolff, Vollständigkeit; MacFarlane, "Logicism"; and Jay Rosenberg, Accessing Kant. See as well Jill Buroker, Introduction. In his essay "Kant within the Tradition of Modern Logic," Ricardo Pozzo begins by endorsing Tonelli's reading, though he adds that transcendental logic must be "the first of the special logics for the generality of its object" (301), which makes transcendental logic "the most general" among special logics (307). An even more qualified adherent to this interpretation would seem to be Béatrice Longuenesse; compare her claim in "Division": "Transcendental logic might count as a case of 'logic of the special use of the understanding,' although it is itself extremely general: it is concerned with the a priori content of thought in any science" (136; my emphasis).

7. Compare Tonelli: "It is my contention that transcendental logic, and the critique of pure reason in general, belong to the class of special logic. More precisely, they are a special logic for metaphysics" (Tradition, 81; cf. 85). Wolff repeats this classification, even providing his own pictorial depiction of a family tree of logics to replace that of Adickes and Kemp Smith, one which places transcendental logic under 'besondere' rather than 'allgemeine' (Vollständigkeit, 204), claiming later that it is clear that Kant "is thinking of his own project of transcendental logic in connection with special logics" because "in Kant's opinion, this logic is to be nothing other than the logic of a particular science, namely, metaphysics" (210).

8. See MacFarlane, "Logicism," 48n35. On the same page, MacFarlane claims explicitly that transcendental logic is to be distinguished from the traditional logic "by its lack of generality," with the result being that "it is a special logic" (48).

9. In imposing this more severe restriction, MacFarlane is joined by Rosenberg, who claims that transcendental logic "is a species of pure specialized logic" because it is "concerned with the most general principles of our thinking about objects experienced as in space and time" (Accessing Kant, 90). Similarly, Buroker claims that "transcendental logic is a special logic falling under pure general logic, for it is the science of the necessary rules of thought about objects given in space and time" (Introduction, 79).

10. See Prolegomena §2: "Judgments . . . may be constituted in whatever manner according to their logical form, and yet there is nonetheless a distinction between them according to their content, by dint of which they are either merely explicative and add nothing to the content of cognition, or ampliative and augment the given cognition; the first may be called analytic judgments, the second synthetic" (4:266; my emphasis).

11. Hence, from a text that (unlike Jäsche's Logik) Kant himself actually prepared for publication. For a discussion of some of the worries surrounding the status of Jäsche's text within Kant's corpus, see Reich, Urteilstafel, 18 and 117n19, as well as Boswell, "On the Textual Authenticity of Kant's Logic." I have been unable to find a corresponding text in Kant's own Reflexionen that corresponds to the passage from Jäsche's Logik §I (9:15) cited above in §2 in support of the domain-subordinative interpretation.

12. To this, compare Kant's "reflection" on metaphysics from the late 1770s, in which Kant allows that the pure concepts can be "extended to all objects of thinking as such" (alle Gegenstände des Denkens überhaupt ausgedehnt) (R5552, 18:220; my emphasis). This is of a piece with Kant's general thesis that "thinking in itself [an sich] . . . is not limited [eingeschränkt] by [the senses]" (B343; my emphasis), that "I can think whatever I wish, just so long as I don't contradict myself" (Bxxvi-fn; my emphasis).

13. Compare, for example, the Reflexionen on metaphysics from 1776-78; see especially R4851 (18:9); see also R5644 from the 1780s (18:284). Compare as well the textbook Kant used for his metaphysics lectures, Alexander Baumgarten's Metaphysica: ontologia est scientia praedicatorum entis generaliorum (§4, 17:24). Wolff gives a similar description in his Philosophia prima sive rationalis: ontologia est scientia entis in genere (Prolegomena, §1).

14. For references to similar uses by the Scholastics, see Rudolf Eisler's Wörterbuch der philosophischen Begriffe, II.514-15.

15. See as well the Metaphysik L2 (Pölitz), 28:555-56. In light of the explanations of 'transcendental' that Kant himself gives here and earlier in the first Critique, the further significance of this label (i.e. over and above the universal and necessary applicability of the given content) in Kant's own hands is that (a) it is content that is not drawn a posteriori directly from our encounters with objects themselves, but is instead drawn a priori from reflection on the nature of our capacity for representing objects, but also that (b) it is content that he thinks we can see "can be applied [angewandt] entirely a priori" to objects. For (a), compare B25: "I call all cognition transcendental that is concerned not so much with objects, but rather with our mode [Art] of cognition of objects as such [überhaupt], insofar as this should be possible a priori"; for the necessity of (b), compare B80, where Kant claims explicitly that "not every cognition a priori must be called 'transcendental,' but only cognition through which we cognize that and how certain representations (intuitions or concepts) are applied entirely a priori or are possible (i.e. the possibility of cognition or of its use a priori)" (my emphasis).

16. Here I agree wholeheartedly with W. H. Walsh, Kant's Criticism of Metaphysics: "[T]he laws of transcendental logic are thus sine quibus non, just as those of [the traditional] logic are; in this respect the two disciplines run parallel rather than diverge" (36).

17. This is something MacFarlane, for one, admits; see MacFarlane, "Logicism," 48n35. Tonelli (Tradition, 85-86) attempts to provide textual grounding for his claim by pointing us to a passage in the student-transcripts of Kant's 1780s lectures on metaphysics (Metaphysik Volckmann) where Kant claims that the critique of the pure use of reason is something that requires a "special or particular" logic: "[W]ith respect to the pure use of reason, a particular or special [besondere] logic will be necessary, which is called transcendental philosophy" (28:363). Crucially, though, Kant does not here identify this special logic with transcendental logic, but rather with transcendental philosophy as a whole. I return to this point in the final section.

18. Kant makes this identification at several places; see B143, B187, B288, and especially B298-305. Cf. Henry Allison, Kant's Transcendental Idealism, 153f.

19. This way of putting things might seem to stand in direct conflict both with Kant's account of judgments of perception in Prolegomena §18 (et seq.) and also with his account of aesthetic judgments in the third Critique, as many readers take Kant to maintain that neither kind of judgment involves the categories at all. I show how the present interpretation is compatible with these doctrines below, in a note to §7.

20. For Kant, a 'cognition' in the broadest sense of this term is a representation that is accompanied "with consciousness" and is "objective" (B376), by which Kant means it "is related [sich beziehen] to an object" (B377). In this broad sense, cognitions can be false; cf. B83, B737, as well as Logik Pölitz, 24:548 and 554; Wiener Logik, 24:832; and R3707, 17:246. This contrasts with a narrower sense of 'cognition,' which draws the term closer to "knowledge" (Wissen) (discussed below) and so applies only to things that are objectively valid, and which Kant often places in opposition to mere thinking; see Bxxvi-n, B146, and B165. Another contrast is to be found in the fact that we have "cognition" of things in themselves in the broad, weak sense, simply in virtue of consciously representing them via the concept of a thing in itself, despite Kant's well-known thesis that we cannot have knowledge of them.

21. In fact, in his later unpublished essay written in response to the 1791 Preisfrage of the Royal Academy of Berlin, Kant claims that "understanding shows its capacity [Vermögen] solely in judgments" (20:271; my emphasis).

22. This claim is repeated in several of Kant's logic lectures; see Logik Busolt, 24: 610; Logik Pölitz, 24:565; and Logik Dohna-Wundlacken, 24:695-96; cf. as well Tonelli, Tradition, 87. In Kant's lectures, mathematics is often grouped together with "morals" (Moral) in this respect; see Logik Pölitz, 24:502; and Logik Dohna-Wundlacken, 24:695. In the latter passage, both are grouped together with physics. This would fit with Kant's description of physics and ethics in the Preface to the Groundwork as sciences that "have to do with determinate [bestimmten] objects and laws"—i.e. having to do with the specific realms of objects that Kant refers to there generically as "nature" and "freedom"—rather than with "the universal rules of thinking as such [überhaupt] without distinction of objects" (4:387-88; my emphasis).

23. Compare B78: "[A]s a general logic it abstracts from all content [Inhalt] of cognition of understanding and from the differences of its objects and has to do with nothing other than the mere form of thinking." For similar characterizations of logic, see Bix, along with the Preface to the Groundwork, 4:387-88; as well as Logik Dohna-Wundlacken, 24:695, 699-700.

24. Among Kant's more immediate predecessors, the general/special distinction can be found, for example, in the writings of the seventeenth-century German logician Joachim Jungius; in both his 1635 Disputationes noematicae as well as his 1638 Logica Hamburgensis, Jungius distinguishes between logica generalis and logica specialis; compare Disputationes noematicae, 23, and Logica Hambergensis, Prolegomena, §§18-22. It is not clear, however, whether Kant had any first-hand knowledge of Jungius's texts, and the Personenindex zu Kants gesammelten Schriften does not list an entry for Jungius. Kant did have access, however, to Leibniz's Nouveaux essais, in which Leibniz (himself a fan of Jungius) also alludes to the general/special distinction within logic, calling the "logic" of the geometers une extension ou promotion particulière de la Logique Generale (Nouveaux Essais IV.2.13; my emphasis; in G V.351). In any case, this distinction is also present in the writings of Martin Knutzen, one of the professors who taught Kant logic at university. Near the end of Kant's time as his student (1746/47), Knutzen published his two-volume Elementa philosophiae rationalis, which bears the subtitle: "With a general as well as a more special logic demonstrated according to the mathematical method" (logicae cum generalis tum specialioris mathematica methodo demonstrata). That Kant was aware of Knutzen's text can be seen from the fact that he singles it out for mention in his own lectures on logic (see Wiener Logik, 24:796-97).

Apparently the general/special distinction dates back at least as far as Averroes; see Zabarella, "De Natura Logicae": [t]ota logica duas habet praecipuas partes . . . quarum unam vocat Averroes universalem seu communem; alteram particularem sive propriam" (53). For these historical leads I am indebted to both Rudolf Meyer's editor's note on page 3 of his edition of Jungius's Logica Hamburgensis, as well Michael Wolff, Vollständigkeit, 202n14 and 207n44.

25. MacFarlane claims that "characterizing logic as formal" is simply "not part of the tradition to which Kant was reacting" ("Logicism," 44). A more substantial defense of this claim is given in MacFarlane's dissertation, "What does it mean to say that logic is formal?", ch. 4.

An earlier argument for the same claim to historical innovation can be found in Adamson, A Short History of Logic, ch. 7, where Adamson argues that "the modern doctrine of logical theory" according to which logic is "a purely formal science" is something that "springs directly from the Kantian philosophy" (110). For even earlier claims to this effect, see the references to nineteenth-century historians of logic given in MacFarlane's essay ("Logicism," 45n33).

In light of these assertions of Kant's novelty, however, it is worth noting that, in addition to Leibniz's and Baumgarten's uses of 'form' and 'formal' cited below, Jungius also had already made use of the language of formality, so much so that he actually already distinguished between logica formalis and logica materialis in both the works cited above; cf. Wolff, Vollständigkeit, 203n16. All of this suggests that this way of thinking about logic was in the air, even if MacFarlane's main point holds—namely, that the philosophical consequences of this feature of the traditional logic were not drawn in a systematic way before Kant.

26. Locke, for example, gives a similar account of the significance of 'is' and other "particles" in Book II, ch. VII of his Essay: such words "signify the connection that the mind gives to ideas" (§1), and thereby serve as "marks of some action, some intimation of the mind" (§4). In their Logique of Port-Royal, Arnauld and Nicole similarly describe the "principal usage" of words like 'is' as that of "signifying movements of our soul" (see Part II, ch. II).

27. Cf. here Allison, Kant's Transcendental Idealism, 153-56.

28. Cf. Kant's description of the process of the discovery of the categories in Prolegomena §39, 4:323-24: "I cast about for an act [Handlung] of the understanding that contains all the rest . . . and I found that this act of the understanding consists in judging. Here lay before me now . . . the work of the logicians, through which I was put in the position to present a complete table of pure functions of understanding, which were however undetermined [unbestimmt] with respect to every object. Finally, I related these functions of judging to objects überhaupt . . . and there arose the pure concepts of understanding."

29. For Kant's acknowledgement of the possibility of this sort of variation in forms of intuition, see e.g. B72 and B150.

30. Strawson makes this point quite clearly; see The Bounds of Sense, 77. See as well Allison, Kant's Transcendental Idealism, 154-55.

31. Kant takes it to be axiomatic that "to every object there must correspond some possible intuition or other," as he asserts in Prolegomena §34n, 4:317 (my emphasis). To be sure, as he also makes clear in this section (and elsewhere), the intuition in question need not be a kind of intuition that our human minds could ever have, even in principle. Perhaps it is only an intuition that some other "finite" intelligence could have, due to its being equipped with an alternative form of sensibility; perhaps it is only one that the infinite divine mind could have. Nevertheless, it seems that, for Kant, the very concept of being an object as such carries with it a reference to the possibility of it (the object) being apprehended in some possible corresponding intuition or other. (For a version of this condition, see B298: "For every concept there is requisite, first, the logical form of a concept [of thinking] in general, and then, second, the possibility of giving it an object to which it is to be related" [my emphasis].) This is why the pure concepts themselves, as concepts of objects in general, are eo ipso concepts of objects of intuition in general. This remains true, even when the objects we are thinking of through such concepts are objects that are evidently not possible objects of human intuition, such as God or the world-whole (see B383f).

32. In fact, the possibility arises that the findings of an inquiry that abstracts from sensibility altogether might well show that our understanding is capable of thinking about objects that we subsequently recognize cannot be given in our sensibility. This is precisely what happens in Kant's own inquiry into our understanding, as is brought out in the Transcendental Dialectic.

33. This does not, however, sever the above conceptual connection that we have just identified a few notes back, between the concept of being an object and concept of being an object of some possible intuition or other. Nor does it conflict with Kant's insistence that the use of these pure concepts for theoretical knowledge (cognition in the stronger sense) is something that cannot be specified independently of an analysis of the conditions on our actually being given the objects so represented: "The use of this pure cognition [from transcendental logic] depends on this as its condition: that objects are given to us in intuition, to which it can be applied [angewandt]" (B87). I return to the significance of this claim below in the concluding section.

34. This is, in effect, admitted by Paton, despite his persistent denial that transcendental logic is universal or general: "We must, however, remember . . . that though we can isolate pure concepts in philosophy and conceive of them 'in their purity' (B91), they are manifest throughout our ordinary experience in the very form of the judgments which we are continually making about empirical objects. As such [the pure concepts] may be said to be acts of pure thought, the formal acts which are present in every judgment and are the same whatever be the particular objects thought" (Metaphysic, 224-25; my emphasis). Compare as well once again Longuenesse: "[T]ranscendental logic might count as a case of 'logic of the special use of the understanding,' although it is extremely general: it is concerned with what might be an a priori content of thought in any science" ("Division," 136; my emphasis).

35. In Jäsche Logik a concept is defined as "a representation of what is common [gemein] to several objects" (§1n1, 9:91); compare R2836 from 1770s (16:538). That which can be "common" to many different objects is what Kant calls a discursive "mark" (Merkmal, see B377).

36. Compare again Prolegomena §2: "Judgments may have any origin whatsoever, or be constituted in whatever manner according to their logical form, and yet there is nonetheless a distinction between them according to their content [Inhalt], by dint of which they are either merely clarifying [erläuternd] and add nothing to the content [Inhalt] of the cognition, or ampliative and augment the given cognition; the first may be called analytic judgments, the second synthetic. Analytic judgments say nothing in the predicate except what was actually thought [gedacht] already in the concept of the subject" (4:266).

37. Compare the 1780s Wiener Logik: "We consider a concept according to content [Inhalt] when we look to the set of representations that are contained [enthalten] in the concept itself" (24:911).

38. We can connect the earlier conception of the content of a cognition as (i) the cognition's relation to an object with the present conception of content as (ii) what is contained or thought "in" the cognition in the following manner: when we relate to an object by thinking of it, say, as a human, we are implicitly relating to it as we would be by explicitly thinking of it as rational and as an animal. The converse is also true: thinking a set of concepts as related (e.g. <animal + rational>) just is representationally relating to an (indeterminate) object that bears the marks represented by those concepts (some thing that is animal and rational). This follows from Kant's persistent rejection of the analysis of the abstraction involved in concept-formation as the abstraction of a new something (abstrahere aliquid), a something that has a self-standing independent existence, rather than merely involving the abstraction from something else (abstrahere ab aliquo), a focus on a non-independent aspect of a thing; see Jäsche Logik §6n3, 9:95; and Logik Dohna-Wundlacken, 24:753-54.

39. As Willem de Jong and Lanier Anderson have emphasized; see de Jong, "Kant's Analytic Judgments and the Traditional Theory of Concepts"; and Anderson, "Paradigm." In fact, many different things might be meant by 'Tree of Porphyry,' especially as there is no drawing of a tree in Porphyry's own text (i.e. his Isagoge); for discussion of the history of this label, see Hacking, "Trees of Logic, Trees of Porphyry." I hope the general shape of the particular tree-like structure that Kant takes to constitute the conceptual hierarchy will become clear enough in what follows. (Thanks to Monte Johnson for discussion of this point.)

40. This is a common theme throughout his lectures on logic. For an encapsulation of the view, see Jäsche Logik, §§7-15, 9:95-99; see as well Wiener Logik, 24:910f; and Logik Dohna-Wundlacken, 24:753f.

41. Kant's conception of the extension or sphere of a concept is more intensional than our own, insofar as it is constituted out of the concepts that are lower than the given concept; see, e.g. Jäsche Logik §9, 9:96; as well as Logik Pölitz, 24:569. This conception of an extension was common among Kant's day; it is found, for example, in Georg Meier's Auszug aus der Vernunftlehre, the textbook Kant used in his lectures; see Auszug §262, 16:560. For some discussion, see Anderson, "Paradigm," 27.

42. See B346 and Logik Dohna-Wundlacken, 24:755; see as well Metaphysics of Morals, 6:218n; and Metaphysik Pölitz, 28:543 and 552. At times Kant seems to claim that this role is played by the concept of "something" (Etwas) (see Wiener Logik, 24:911), but at the end of the Amphiboly Kant is quite straightforward about the fact that the concept of something presupposes a higher concept, because it has its opposite, i.e. the concept of "nothing" (Nichts) (see B346). At other times Kant picks out the highest concept with words that presumably must be taken as synonyms for 'object,' such as "thing" (Ding) and "being" (Wesen); see Logik Dohna-Wundlacken, 24:754.

43. It must be very abstract given that Kant means for it to apply to all of the aforementioned cases (substances, realities, things which exist, "somethings," etc.) as well as to their opposites (accidents, idealities, non-existents, "nothings," etc.). In fact, in this very abstract sense, even a concept or a judgment— indeed, any representation—will count as an object; cf. Kant's remark in the A-deduction that "all representations can be objects [Gegenstände] of other representations" (A108).

44. I am indebted to an anonymous reviewer for pushing me to clarify this point.

45. It follows from this that judgments of the form 'A's are objects' will be analytic for any concept <A>.

46. Examples of ways of being nothing that Kant gives here are: being an "entity of reason" (ens rationis) (e.g. a noumenon), being a shadow or cold, being the form of intuition, being a two-sided rectilinear figure (B347-48). Though these are all "nothing," they are nevertheless all "objects" in the relevant (merely "logical") sense.

47. For one thing, it would seem that Kant's fourfold division into "moments" (quantity, quality, relation, modality) would represent a higher division than any of the threefold divisions of each of the four moments into the individual categories themselves. And it would seem that there must an even higher twofold division that guides and orders this fourfold division itself—something Kant himself seems to suggest in the section following the presentation of the Table of Categories; see §11 of the first Critique: "This table, which contains four classes [Classen] of concepts of the understanding, can first be split into two divisions, the first of which is concerned with objects of intuition (pure as well as empirical), the second of which, however, is directed at the existence of these objects (either in relation to each other or to the understanding). I will call the first class the mathematical categories, the second, the dynamical ones" (B110; cf. B199).

A further question pertains to the so-called "concepts of reflection" that Kant introduces in the Amphiboly, a set of eight also very abstract concepts (e.g. <identity>, <difference>, <agreement>, <opposition>, etc.), which is divided into a four pairs (see B317). Both the fourfold division, as well as each member of the pairs themselves, find their obvious correlate on the Table of Categories.

48. Kant describes the categories as "marks of pure understanding" in the second Critique, writing that the "explanation" (Erklärung) of the "faculty of desire" (Begehrungsvermögen) that is accomplished in that Critique is one that "is composed solely out of marks [Merkmalen] of the pure understanding, i.e. categories, that contain nothing empirical" (5:9n).

49. This question would become especially pressing if the categories under one title are not actually exclusive of one another—if, for example, two things' being related as substance-accident did not exclude the same two things' being related as cause-effect. (I am indebted again to an anonymous referee for bringing out the force of this worry.)

50. The classic defense of Kant's procedure is to be found in Reich, Urteilstafel; for a more recent attempt, see Wolff, Vollständigkeit.

51. In §10 of the first Critique, Kant himself points toward the first stages of the further division of the categories into "derivative though still pure" concepts that lie "under" them, pure concepts that Kant calls "predicables," with the goal in mind being that the "family tree [Stammbaum] of pure understanding" could be "fully illustrated" (B108; my emphasis).

52. Here we can return to the worry raised in a previous note concerning how the present interpretation will be able to accommodate either the account of "judgments of perception" that Kant gives in the Prolegomena (see §18 et seq.) or the account of aesthetic judgments that Kant gives in the third Critique. About the judgments of perception, Kant writes that they "do not require a pure concept" and do not yet have a "relation [Beziehung] to an object" (Prolegomena 4:298); about aesthetic judgments, Kant writes that they "precede all concepts of the object" (Prolegomena 20:243), and are "neither grounded on concepts nor aimed at them" (Critique of Judgment §5, 5:209). In both cases, therefore, Kant appears to say that such judgments do not involve the categories. In both cases, however, I think that this appearance is ultimately misleading. The difference Kant is marking, in each case, is not between category-involving and non-category-involving judgments, but rather a difference between ways in which judgments can involve categories. In judgments of experience, as with "determinative" judgments more generally, we intend our judgments to "determine" an object by a category; we take them to be objectively valid. In both judgments of perception and judgments of taste, by contrast, even if the content of our judgment still appears to involve categories in the ordinary sense (e.g. our judgment involves the representation of a physical body [wormwood, a rose]), we are not aiming to make a claim about ("determine") the object per se, but instead asserting how the representation of the object is related to ourselves. In both cases, what is being expressed is a relation between these representations and the subject, rather than whatever other objects may be prima facie involved: judgments of perception "express a relation of sensations to the same subject, namely myself" (Prolegomena §19, 4:299); aesthetic judgments express the "feeling" of pleasure or displeasure, "in which a subject feels itself as it is affected by the representation" (Critique of Judgment §1, 5:204). Yet, insofar as I am here representing myself as a substance, as the bearer of this feeling, I nevertheless mean to "determine" something about at least one object—namely, myself. For further development of this sort of interpretation of judgments of perception, see Prauss, Erscheinung bei Kant, 150ff; see as well Beck, "Did the Sage of Königsberg Have no Dreams?" For an intepretation of aesthetic judgments as directed at the subject, see Longuenesse, "Kant's Theory of Judgment and Judgments of Taste," esp. 149ff. (Thanks once again to a reviewer for pushing me to clarify this point.)

53. In addition to Strawson, Bounds of Sense, see Ameriks, Kant's Theory of Mind, 268, as well as 82n99. See also Ameriks, Interpreting Kant's Critiques, 28, 32; as well as Parsons, "Kant's Philosophy of Arithmetic," 117; and also Körner, Kant, 74-75. For an interpretation that would seem to leave no room for the unschematized categories, see Friedman, The Parting of the Ways, 27 and 91.

54. Compare Kant's remarks in the Schematism: "Even after abstraction from every sensible condition, significance [Bedeutung] is left to the pure concepts of the understanding, though only a logical significance of the mere unity of representations" (B186; my emphasis). Compare as well the "transcendental significance [Bedeutung]" that Kant accords to "the pure categories" even "without formal conditions of sensibility" (B305; my emphasis).

55. Along a similar line of thought, Kant also needs the pure concepts or ideas of reason to have an "unschematized" content sufficient to figure in our thoughts, "opinion" (Meinen), and (crucially) "belief" (Glauben) about objects that lie beyond possible intuition or experience. Though non-sensible, this content must be sufficiently substantive to allow for the differentiation of, e.g. thoughts about God from thoughts about our own immortal soul (see Bxxx and B850f).

56. I again mean 'cognize' here in the broader, weaker sense of this term, according to which thinking can simply be defined as "cognition through concepts" (B94), and not (of course) in the narrower, stronger sense of this term, according to which it stands in contrast to cognition, because thinking is not itself an intuiting (cf. Bxxvi and B165).

57. As Heidegger, for one, has noted; compare Heidegger's remarks in §13 of his 1927/28 lecture course, Phänomenologische Interpretation. Here Heidegger claims explicitly that "were the Transcendental Logic to be structured in a way corresponding entirely to the Transcendental Aesthetic, then it would have to end at B169" (165), pointing to Kant's own claim at the end of the Deduction, cited below. Heidegger takes this to mean that whereas "the preceding had to do with the concepts of the elements of cognition—sensibility and understanding—insofar as these elements were taken in isolation and analyzed in their structure," what follows will take up the main task of the Critique itself: "[T]his isolating analysis into elements—and this is the function of the Aesthetic and the Logik—is however a preliminary stage [Vorstufe] for investigating the whole of a cognition, i.e. the unification of sensibility and understanding, with respect to its possibility"(166). For this reason, Heidegger suggests that because "in what follows the employment of elements is made the theme," in effect, "it deals neither with transcendental aesthetic nor with transcendental logic, but either with both or with neither of the two" (166; my emphasis).

58. Hence, MacFarlane, Rosenberg, and Buroker would be on the right track if they limited their claims to the Analytic of Principles, rather than extending them to the discipline of transcendental logic per se. Similarly, though Tonelli's claim that Kant intends the first Critique as a whole to be a "special logic" for metaphysics (Tradition, 87f) surely goes too far, there is something right about the application of this label to these later parts of the Transcendental Logic. This would seem to be what Kant has in mind in the passage from his lectures on metaphysics that Tonelli directs us toward: "[W]ith respect to the pure use of reason, a particular or special [besondere] logic will be necessary, which is called transcendental philosophy" (28:363; my emphasis). Nevertheless, what I hope to have established is that it is equally crucial to recognize that there is a more fundamental core to transcendental logic per se, an Elementarlehre, which is not "special" at all, but is aimed at a genuinely universal doctrine of the content of understanding.

59. Hence MacFarlane is wrong to argue in his essay that "formality is for Kant merely a consequence of logic's generality" ("Logicism," 20; my emphasis), because with transcendental logic, as with the traditional ontology that it is meant to replace, we have a science that that is absolutely general but is decidedly non-formal.

60. The present interpretation would also seem to imply that what transpires in the second Critique's presentation of the doctrine of the concept of an object of pure practical reason—and hence, in the doctrine of the categories of freedom—should be counted as a special or particular logic, because it is concerned with a special use of these categories (in distinctly practical cognition of the good and the bad), though one that had the same transcendental-logical core as did the equally "special" logics for the distinctly theoretical uses of our understanding that occupy the Principles and the Dialectic of the first Critique. This actually fits quite well with Kant's own characterization of the sections at the end of the second Critique's "Analytic" as providing a "Logic" on par with that of the first Critique (5:90).

This also fits well with Kant's conception of the basic divisions within metaphysics itself, as he presents this in the chapter "Architectonic" at the end of the first Critique, as well as in the Preface to the Groundwork. In the Architectonic, Kant tells us that metaphysics as a whole divides into the "metaphysics of nature," which investigates "the speculative use [Gebrauch] of pure reason," and the "metaphysics of morals," which investigates reason's "practical use" (B869). In the Preface to the Groundwork, Kant describes both of these disciplines as dealing with "determinate objects and their laws" (4:387). Transcendental logic sensu stricto would stand in contrast to both of these sciences of special or particular uses of our understanding. While each of these focus on the use of our understanding in relation to a special class of objects (i.e. those that belong to nature or to freedom), transcendental logic in the strict sense would be focused on the basic concepts of an object überhaupt, concepts that leave it indeterminate whether the objects at issue belonged to nature or to freedom. In this way it would therefore have a more "elementary" subject-matter.

61. I would like to thank Michael Hardimon, Jeremy Heis, Karolina Huebner, Monte Johnson, Samantha Matherne, Don Rutherford, Gila Sher, and Eric Watkins—along with the extremely helpful referees from this journal—for many thoughtful comments and suggestions on earlier drafts of this essay. Versions of this material were presented at UCSD's History of Philosophy Roundtable and at the Pacific Study Group meeting of the North American Kant Society at the Claremont Colleges, and I would like to thank those audiences as well for very constructive discussion and feedback.

Additional Information

Print ISSN
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.