Analytic Cognition in Kant

Abstract

Kant refers to analytic cognition in several prominent places. The prevailing wisdom, however, denies the possibility of analytic cognition within his theory of cognition. I shall argue that this is mistaken. I show that we can account for analytic cognition’s possibility by appealing to variants of the more familiar conditions on the cognition of objects. I also highlight analytic cognition’s connection to insight and analytic knowledge. In the process, I provide a fuller account of Kant’s view of our mental lives than has been typically acknowledged.

To be sure, one has analytic cognitions a priori, if the concept of object is given, whether it be an empirical or a rational concept. (Refl 6050, 18: 437 [1780s?]) ¹

1. Introduction

In the Critique of Pure Reason, Kant claims that cognitions are representations that require both intuitions and concepts. This claim features in some of his most distinctive views, including, for instance, in the idea that we cannot cognise things-in-themselves and that the cognitions of mathematics and philosophy are synthetic and a priori. Accordingly, it comes as a surprise when Kant says that the mere analysis of concepts absent intuitions ‘affords us a multitude of cognitions’ (Menge von Erkenntnissen) (A5/B9) and that one can ‘cognize the concept of body analytically’ (kann den Begriff des Körpers vorher analytisch … erkennen) (A8/B12). References to the possibility of ‘analytic cognition’ (analytische Erkenntnis) are found in various other of Kant’s works (A12/B25, A151/B190–1; P, 4: 267; Log-W, 24: 845; Log-Bl, 24: 131). Indeed, Kant not only alludes to the possibility of analytic cognition but also claims that the analysis of concepts constitutes a ‘great part, perhaps the greatest part, of the business of our reason’ (A5–6/B9–10). Given the significance of cognition to Kant’s philosophy, one might wonder how analytic cognition can, for him, count as cognition.

Commentators, however, have typically treated Kant’s references to analytic cognition as instances of loose talk or have otherwise simply ignored them. ² The prevailing wisdom has therefore denied the possibility of analytic cognition within Kant’s theories, arguing that analytic cognition is ruled out by his definitions. Those who acknowledge his references to analytic cognition have assumed that analytic cognition counts as cognition merely in the broad sense of representing something or that analytic cognitions are the refinements of empirical concepts of objects given in intuitions. ³ These assumptions, I shall argue, are unwarranted.

This article aims to shed light on the sense in which analytic cognition can be accommodated within Kant’s theory of cognition and, in the process, to provide a fuller account of his view of our mental lives. I shall argue that a subject’s representation counts as an analytic cognition if and only if it satisfies three conditions: that it renders a mark, or feature of something as representable by a concept, available to reason; that it combines that mark with a concept; and that the subject is conscious of the formal standard of correctly combining mark and concept. ⁴ That standard, I shall argue, is the principle of contradiction.

I shall proceed as follows. In Section 2, I describe what I take to be three conditions on the cognition of empirical objects by the understanding (the givenness, conceptual, and consciousness conditions). I also briefly describe how mathematical and philosophical cognition are two kinds of ‘rational cognition’ (Vernunfterkenntnis) (B741) that can relatively straightforwardly satisfy these conditions and clarify how a third, namely, analytic cognition, remains problematic. In Section 3, however, I describe how analytic cognition satisfies variants of these conditions. In Section 4, to clarify how my reading of analytic cognition fits within broader Kantian commitments, I describe the place of analytic cognition in the Jäsche Logic’s grades of cognition passage and analytic cognition’s relation to analytic knowledge. In Section 5, I articulate and respond to objections to my interpretation. Finally, in Section 6, I highlight a central implication of this paper, namely, that all kinds of cognition share a minimal definition: a subject’s representation counts as a cognition iff (i) it renders available to a faculty a feature as an individual feature, (ii) it combines that feature with a concept, and (iii) the subject is conscious of what they are representing and thus of a standard of correctly combining the relevant feature and concept. This minimal definition sheds light on Kant’s suggestion that reason and the understanding share a ‘general root’ of cognition (A835/B863).

2. Empirical, mathematical, and philosophical cognitions

2.1 Empirical cognition: intuitions, concepts, consciousness

To understand the possibility of analytic cognition, we must understand a central concern of Kant’s Critique, namely, the nature of ‘empirical cognition’ (A92–3/B125–6). Kant claims that empirical cognition is a representation that results from the faculty of sensibility, which supplies empirical intuitions, working alongside the understanding, which supplies concepts. He writes, for instance:

There are two conditions under which alone the cognition of an object is possible: first, intuition, through which it is given, but only as appearance; second, concept, through which an object is thought that corresponds to this intuition. (A92–3/B125)

Let us consider each condition in turn, starting with givenness. Kant characterises the function of intuitions as giving (gegeben) objects to a subject. He has, at this point, previously explained that this entails that empirical intuitions, through the ‘sensations’ which they contain (A19/B34), are the kind of representations through which ‘we are affected by objects’ (A19/B33), where those objects are thereby available to our minds, particularly the understanding. An empirical intuition gives us objects because it is a representation that is ‘immediately related to an object’ and singular’ (A320/B377). Kant cites sight, taste, and colour as examples of the content of sensations contained in empirical intuitions (A28/B44).

Intuitions, according to Kant, are immediate insofar as they relate to an object without the presence of mediating mental content (e.g. concepts). There are several suggestions for how to further elucidate ‘immediacy’. Intuitions have been interpreted as being caused directly by the object being represented (e.g. Watkins and Willaschek 2017). Alternatively, they have been read as involving a phenomenological presence to the mind through a direct relation of acquaintance between consciousness and objects (Allais 2015: 158–60).

Intuitions are also singular (rather than general) representations through which empirical objects are given. This underpins the very distinction between intuitions and concepts. As Kant claims, ‘concepts differ from intuition by virtue of the fact that all intuition is singular’ (Log-W, 24: 905). Thus, an intuition’s content is of an individual object or an individual feature of an object. It picks out, for instance, a particular cup or feature of the cup in front of me, not ‘cup’ in general or some general feature of cups.

Thus, a function of intuitions in empirical cognition is to render available to the faculty of understanding individual empirical objects or features thereof (Allais 2015: 154–8). To say that a feature is rendered available to a faculty in this way means that a subject is in a position to attend to it as a discrete feature. ⁵ In turn, the subject will be able to hold that feature as a salient one in contrast to an object’s other features, or background, and track it across changing conditions. For instance, if sphericalness is rendered available to a subject’s faculty of understanding as an individual feature, then the subject will be able to attend to sphericalness as an individual feature. Thus, they will be able to (1) attend to that feature as a salient feature in contrast to (say) hardness or redness, (2) pick out sphericalness against a background, and (3) track sphericalness as lighting conditions change. Thus, intuitions – insofar as they render available to a faculty a feature as an individual feature – are distinct from more complex representations (e.g. a representation of an object as a ball, for which a subject would require the concept ‘ball’ and not merely the capacity to hold salient, pick out, and track the sphericalness of something).

Thus, for Kant, a subject’s representation counts as an empirical cognition only if it immediately represents particular features of an object such that those features are available as individual features, or ‘given’, to the understanding. I call this the givenness condition.

Second, let us consider the conceptual condition. Rejecting Leibniz’s conception of complete individual concepts, Kant thinks that all of our concepts are general representations: concepts are marks ‘common to several things’ (A320/B377) that allow us to think in generalisations (A320/B377, A19/B33, A50/B74, A713/B741): ‘he thought by red that which was common to many objects, and this was a concept’ (Log-W, 24: 905). Given that intuitions represent particular – as opposed to general – features, we cannot relate different objects to one another without concepts, which are representations common to many things. Accordingly, the function of concepts is to allow us to relate different objects to one another. For instance, possessing the concept ‘ball’ and employing it in the thought ‘this is a ball’ allows, and requires, that one may also think of something else, ‘that is a ball’. So, alongside the givenness condition, a subject’s representation of a ball, for instance, counts as a cognition only if it is a representation that combines particular representations (intuitions) with a general representation (a concept ‘ball’) that can be applied to other objects. This I call the conceptual condition.

Third, alongside the givenness and conceptual conditions, Kant holds that cognition of empirical objects requires consciousness of those objects. ⁶ In the A-Deduction, he writes that without consciousness, cognition of objects ‘would be entirely impossible’ (A104, A320/B376–7; JL, 9: 65; Log-Bl, 24: 133–4). ⁷ We can understand this requirement in three steps, as follows: (1) cognition requires a subject’s consciousness of what they are representing with the combination of intuition and concept. A subject’s representation that combines the particular feature sphericalness with the concept ball is not a genuine cognition unless the subject is conscious that they are representing the relevant object.

Consequently, (2) cognition requires consciousness that representing something involves correct standards of what one is representing because a subject’s awareness that they are representing something provides them with a necessarily correct way of representing that thing. If I am conscious that I am representing an object, then I am aware that I correctly represent that object iff I represent it with properties that it indeed possesses. As Kant puts it, ‘our thought of the relation of all cognition to its object carries something of necessity with it’ (A104). A representation lacking a consciousness of standards of representation is ‘opposed to our cognitions’ and is thus a combination of mental content ‘determined at pleasure and arbitrarily’ (ibid.).

As a further consequence, (3) the subject must be conscious of the material standards of correctly combining the relevant intuition and concept, namely, whether or not the representation agrees with the object. This is because the relevant standard arises from the subject’s awareness that they are representing an object, as noted by Kant when he claims that ‘our cognitions must … necessarily agree with each other in relation to [that object]’ (A104). If a subject is conscious of their representation as representing some object, then they will also be conscious of that representation having a material standard of correct representation. If I am (1) conscious that I am representing a ball, then I will be (2) conscious that doing so involves standards of correct representation and (3) conscious that my representation is correct iff that object is indeed a ball. Therefore, with respect of objects, to have a cognition is to be conscious of the material standards of correctly combining intuition and concept in a representation. I call this the consciousness condition.

The emerging account of empirical cognition is thus as follows:

1. Empirical cognition: A subject’s representation counts as an empirical cognition iff:

2. (i) the subject’s representation renders a feature of an empirical object (an empirical intuition) immediately available to the understanding (the givenness condition);

3. (ii) the subject combines that empirical intuition with a concept (the conceptual condition);

4. (iii) the subject is conscious that they are representing an empirical object and thus the material standards of correctly combining the relevant empirical intuitions and concepts (the consciousness condition).

We could further refine our definition of empirical cognition. For instance, as stated, the givenness condition does not explain how intuitions present us with objects. However, analytic cognition does not involve intuitions. So, curiously, the givenness of objects, as opposed to givenness simpliciter, will fall away from our analysis. However, this brief sketch allows us to see, in basic outline, what other instances of cognition must resemble to be recognisable as cognition within Kant’s theory of cognition.

2.2 Mathematical cognition

The above characterisation of cognition applies to cognition of empirical objects and therefore to those given in empirical intuition. The objects of pure mathematics lie, for Kant, beyond experience and are thereby lacking corresponding empirical intuitions. Accordingly, they are not cognisable by the faculties of sensibility and understanding working in tandem. Instead, Kant claims, besides cognitions of the understanding, we have cognitions of reason, or ‘rational cognitions’ (A713/B741), that include the a priori cognitions involved in pure mathematics, philosophy, and conceptual analysis. It will be useful, at this stage of the argument, briefly to highlight the ways in which mathematics and philosophy involve synthetic pure cognitions that satisfy variants of the above three conditions. This will allow us to see how cases of analytic cognition might likewise be defined by analogous conditions. I turn first to mathematical cognition.

In The Discipline of Pure Reason in Dogmatic Use, Kant characterises mathematical cognition as a priori rational cognition from the ‘construction of concepts’ (A713/B741; see Tolley 2020). To understand what this means, let us begin with the givenness condition.

Kant cites this condition as the reason that mathematical judgement yields ‘synthetic but rational cognition’ (A722/B750). Mathematical judgements yield rational cognition because they involve a priori mathematical concepts, such as ‘triangle’ and ‘magnitude’ (A716–7/B743–4). Mathematical judgements yield synthetic cognitions because, rather than merely involving the analysis of a concept, they involve the ‘addition’ of a priori intuitions to that concept through its construction in time and space (i.e. the mere form of appearances). Judgements in mathematical proofs require the construction of particular concrete mathematical objects in a priori intuition. As Kant argues, mathematical cognition involves going ‘beyond [a] concept, and indeed [going] to the intuition in which it is given’ (A721/B749), where the relevant intuitions are the a priori ‘form of intuition (time and space)’ (A723/B751). Mathematical cognition thus results from a ‘chain of inferences that is always guided by intuition’, allowing for the concrete representation of objects (A717/B745; see Friedman 1992; Hogan 2020). With respect to a triangle, this concrete representation permits the mathematician to attend to ‘three-sidedness’ by rendering that feature available to reason as a discrete feature.

With respect to mathematical cognition, the givenness condition is satisfied by a priori forms of intuitions of time and space. The conceptual condition is also satisfied in mathematical cognition given that mathematical concepts – the concept of ‘triangle’, say – are constructed in a priori intuition.

Moreover, the consciousness condition is satisfied in mathematical cognition. Through the construction of a mathematical object in a priori intuition, ‘anything unfounded and arbitrary instantly becomes obvious’ (A711/B739). That is, in constructing a concept in a priori intuition, one becomes aware of the correct material standards of constructing and thereby representing that concept in a priori intuition. Thus, we can summarise mathematical cognitions as follows:

1. Mathematical cognition: A subject’s representation counts as a mathematical cognition iff:

2. (i) the subject’s representation renders a feature of a constructed mathematical object available to reason by constructing a concept in a priori time or space (the givenness condition);

3. (ii) the subject combines that feature with a mathematical concept (the conceptual condition);

4. (iii) the subject is conscious that they are representing a mathematical object and thus the material standards of correctly combining the relevant feature and mathematical concept via construction (the consciousness condition).

Accordingly, it seems that all three conditions considered here are satisfied by mathematical cognition. I now turn to another kind of rational cognition, namely, philosophical cognition.

2.3 Philosophical cognition

Kant’s discussion of synthetic a priori philosophical cognition is also analogous to his discussion of empirical cognition, in ways that will help us understand the possibility of analytic cognition. Against the view of Wolff and other rationalists that the methods of mathematics can be used to achieve certainty in philosophy, Kant argues that there is a strict distinction between the methods of mathematics and those of philosophy, given that they concern different kinds of cognition (A713/B741). Unlike the construction of concepts in mathematical cognition, philosophical cognition is ‘rational cognition from concepts’ lacking empirical or a priori intuitions (A713/B741).

For Kant, the key distinction between philosophy and mathematics is that the former ‘confines itself solely to general concepts’ and thus is without objects, while the latter ‘cannot do anything with mere concepts but hurries immediately to intuitions’ (A715/B743). Given his commitment to the givenness condition and its intuition-involving nature, Kant seemingly contradicts himself in this passage in claiming that philosophical cognition is rational cognition from concepts alone. He then explains, however, the sense in which philosophical cognition counts as satisfying the givenness condition:

Now an a priori concept … either already contains a pure intuition in itself, in which case it can be constructed; or else it contains nothing but the synthesis of possible intuitions, which are not given a priori, in which case one can well judge synthetically and a priori by its means, but only discursively, in accordance with concepts…. (A719–20/B747–8)

Kant’s conception of the intuitions operative in philosophical cognition is all possible intuitions (cf. Tolley 2020). Philosophical cognitions count as cognitions partly because they involve the application of various concepts to ‘all possible intuitions’ as a feature of the concept ‘all possible intuitions’. In this sense, ‘all possible intuitions’ are rendered available to reason as an individual feature insofar as a subject can intellectually attend to it as a discrete feature. Thus, the philosopher attends, in philosophical cognition, to ‘all possible intuitions’ as opposed to certain possible empirical intuitions. Kant’s discussion of philosophical cognition thus suggests that, for givenness to be satisfied, a subject’s representation must render an individual feature of the concept ‘all possible intuitions’ available to reason.

Philosophical cognitions result from synthetic judgements about things in general, thereby abstracted from possible or actual objects and, in turn, empirical or a priori intuitions. Accordingly, connecting transcendental concepts – such as reality, substance, force, etc. – with ‘the synthesis of [all] empirical intuitions’ can count as ‘synthetic rational cognition in accordance with mere concepts’ (A722/B750), despite such cognition not being cognition of objects.

The givenness condition is therefore satisfied. This shows how the conceptual condition might also be satisfied. Clearly, Kant has the categories in mind when he discusses concepts being applied to all intuitions. For instance, he says that the concept of causality (everything that happens has a cause) relates to all possible intuitions because ‘it has the special property that it first makes possible its ground of proof, namely, experience’ (A737/B765). The consciousness condition may therefore be satisfied in philosophical cognition in that a subject’s representation counts as philosophical cognition only if their representation renders them aware of the transcendental standards of representation, that is, if they are aware of the standards by which a category relates to all empirical intuitions.

Accordingly, we may draw the following from Kant’s discussion of philosophical cognition:

1. Philosophical cognition: A subject’s representation counts as a philosophical cognition iff:

2. (i) the subject’s representation renders the feature ‘all possible intuitions’ available to reason (the givenness condition);

3. (ii) the subject combines ‘all possible intuitions’ with a transcendental concept (the conceptual condition);

4. (iii) the subject is conscious that they are representing a philosophical concept and thus the transcendental standards of correctly combining the relevant feature and transcendental concepts (the consciousness condition).

There is much more to be said about mathematical and philosophical cognition. But I think that Kant’s strategy for understanding them as satisfying the givenness, conceptual, and consciousness conditions is relatively clear. Analytic cognition, on the other hand, cannot count as cognition in the mathematical or philosophical sense because analytic cognition does not involve the construction of mathematical concepts in a priori time and space or the generalisation of the categories over all empirical intuitions.

However, with regard to philosophical cognition, ‘all possible intuitions’ is not an intuition but plays a functional role in cognition. It is a concept to which the categories are being applied, and thus Kant permits a kind of cognition that lacks intuitive content. This insight, I think, can serve as the basis of reading the three conditions on cognition more generally, to avoid dismissing Kant’s references to analytic cognition as loose talk.

3. Analytic cognition’s three conditions

3.1 Textual considerations

Throughout the Critique, references to analytic cognition are far less frequent than those to empirical cognition and the possibility of synthetic a priori cognition of the mathematical and philosophical kinds discussed above. In several prominent passages, however, Kant does refer to analytic cognition. For instance, as noted above, he says that analysis ‘affords us a multitude of cognitions’ (A5/B9) by way of characterising the central business of reason as the analysis of concepts in abstraction from intuitions. Kant also introduces the analytic/synthetic distinction by claiming that we ‘can cognize the concept of body analytically’ (A8/B12, my emphasis) and characterises his transcendental philosophy as ‘a science [that] would have to contain completely both analytic as well as synthetic a priori cognition [analytische Erkenntnis, als die synthetische a priori]’ (A12/B25, my emphasis). He describes the principle of contradiction as the supreme principle of analyticity by claiming that it counts as ‘the universal and completely sufficient principle of all analytic cognition [aller analytischen Erkenntnis]’ (A151/B190–191, my emphasis).

Furthermore, references to analytic cognition are not limited to the Critique. In the preamble to the Prolegomena, Kant says, ‘all analytic judgements … are by their nature cognitions a priori’ (P, 4: 267). In his logic lectures, Kant distinguishes between ‘[a]nalytic cognition … with [which we] make a given concept distinct’ and synthetic cognition that gives us a concept with distinctness (Log-W, 24: 845; cf. Log-Bl, 24: 131). And, in the Reflection quoted at the start of this paper, Kant appears to endorse the view that ‘one has analytic cognitions a priori’ (Refl 6050, 18: 437). In light of Kant’s consistent references to ‘analytic cognition’ across many texts, he appears committed to the idea that analytic cognitions are cognitions in some substantive sense, at least to the degree that philosophical cognitions count as cognitions. This holds despite the fact that analytic cognition appears to be an unlikely candidate for cognition within a Kantian framework.

Unfortunately, Kant never provides a systematic treatment of analytic cognition. We shall therefore need to assemble an interpretation of analytic cognition from his scattered remarks. A good way to start is with a brief discussion of the relation of analytic judgement to analytic cognition.

3.2 Analytic judgement, analytic cognition

In Kant’s theoretical philosophy, he defines judgements (Urteil) as conscious representations that are ‘functions of unity among our representation’ (A69/B94; JL, 9: 101). ⁸ He means that judgements require a specific ‘unifying’ structure to bring together concepts in such a way as to generate truth-apt propositions. This structure is expressed with a copula. For instance, in the analytic judgement ‘All bachelors are unmarried men’, one unifies ‘all bachelors’ with ‘unmarried men’ using the copula ‘are’. Thus, a judgement is a distinctive kind of representation necessarily involving predication.

Judgement and cognition are importantly and necessarily interrelated, and yet they are not equivalent. Each involves the act of unifying representations; however, Kant holds that one can both represent states of affairs with cognitions and form judgements about them, but one can cognise independently of forming judgements. As he writes, the presence of the copula explains the ‘relation of given cognitions in every judgement’ (B141; cf. P, 4: 298; see Willaschek and Watkins 2020). Accordingly, a subject may have a cognition and yet fail to form a judgement by failing to have a propositional thought about that cognitive content.

Therefore, when Kant claims that ‘all analytic judgements … are by their nature cognitions a priori’ (G, 4: 267), he cannot mean that all analytic judgements are merely identical to analytic cognitions; this would collapse the judgement/cognition distinction. Rather, he must be saying that predication in analytic judgement involves the representative content of cognition. When Kant is choosing his words carefully, he identifies analytic judgements with ‘draw[ing] out the predicate’ of what is already thought in the subject of a proposition (A7/B11), and conversely, he identifies analytic cognition with representation that makes ‘given concept[s] distinct’ (Log-W, 24: 845; cf. JL, 9: 64). ⁹ Yet the question remains: in what sense does the representational content of analytic cognition satisfy the givenness, conceptual, and consciousness conditions on cognition described above?

3.3 The givenness condition

When Kant briefly turns his attention to analytic cognition in the Aesthetic, he does so within a discussion of mathematical truth:

There is no other way [to form the necessarily true propositions of mathematics] than through concepts or through intuitions, both of which, however, are given, as such, either a priori or a posteriori … Concerning the first and only means for attaining to such [mathematical] cognitions, however, namely through mere concepts or a priori intuitions, it is clear that from mere concepts no synthetic cognition but only merely analytic cognition can be attained. (A47/B64)

Kant then shows how necessarily true mathematical propositions are possible only through the kind of synthetic a priori cognition described in Section 2.2 as mathematical cognition, involving a priori intuitions and concepts, as opposed to analytic or empirical cognition. However, the quoted passage above makes two further relevant claims. First, Kant expressly commits to the possibility of analytic cognition being ‘attained’. Second, he claims that intuitions as well as concepts can be ‘given’ (gegeben) (see also JL, 9: 142). My discussion in Section 2 would seem to lead to the conclusion that, in discussing givenness, Kant means to discuss minimally a property of cognition such that both intuitions and concepts are rendered available to the subject’s awareness. Indeed, he says that intuitions and concepts can be given a priori or a posteriori, suggesting that he is operating with a broader sense of givenness than merely one of objects given in intuition.

I therefore argue that Kant is operating with two distinct senses of givenness: the givenness of objects and the givenness of concepts. Kant repeatedly speaks of the givenness of objects; in the Aesthetic, he says that ‘the object is given to us … only if it affects the mind in a certain way’, and that ‘[objects] are therefore given to us by means of sensibility, and it alone affords us intuitions’ (A19/B33). A rarely noted point about this passage is that it is not clearly a characterisation of givenness simpliciter but rather of the givenness of objects. Kant does not say – or even imply – that givenness merely is the givenness of objects or that givenness is satisfied only by intuitions. Rather, he claims that a necessary condition of objects being given to us is that they affect our minds in an as yet unspecified way. Kant later specifies that objects are given to us via sensibility and that sensibility furnishes us with intuitions. Indeed, he often clarifies that ‘objects are given to us’ through the faculty of sensibility (A15/B29, cf. A62/B87, A139/B178, A155/B194) in a way that leaves conceptual space for other forms of givenness. ¹⁰

I think that Kant’s claim about the givenness of concepts, with respect to analytic cognition, is best captured by a functional variant of the givenness condition on cognitions of objects: a subject’s representation counts as an analytic cognition only if a feature of a concept (a mark) is rendered available to reason as an individual feature. In analytic cognition, a subject becomes aware of a concept’s features as individual features by considering a concept. A mark for Kant, as a feature of a concept, is a ‘partial concept’ (JL, 9: 58). ¹¹ Analytic marks, for him, are ‘partial concepts of [one’s] actual concepts’ (JL, 9: 59). In this sense, analytic marks are component representations of concepts. As such, ‘having colour’, ‘yellow’, ‘a particular shade of yellow’, and ‘metal’ are all marks of the concept ‘yellow metal’. Consider then the analytic cognition that would be involved in Kant’s example of the analytic judgement ‘Gold is a yellow metal’ (P, 4: 267). The givenness of the concepts condition would be satisfied only if the mark ‘yellow’ of the concept ‘yellow metal’ is available for the subject to intellectually attend to as an individual feature of the concept ‘yellow metal’. Thus, the subject will be able to hold as salient ‘yellow’ in contrast to ‘having colour’ or ‘having a particular shade of yellow’, thereby satisfying givenness for analytic cognitions.

What reasons are there for attributing to Kant a functional analogy between objects given to the understanding via sensibility on the one hand and, on the other, concepts given to reason via the understanding?

Kant implies a commitment to an analogous structure between the cognition of empirical objects (of the understanding) and cognition from concepts (of reason). This commitment is found in such statements as ‘pure reason is never related directly to objects, but instead to concepts of the understanding of them’ (A335/B392, my translation; cf. A680/B708). Reason, that is, connects concepts of the understanding, as opposed to the understanding that connects intuitions from sensibility via concepts. If we take Kant’s analogy seriously, then cognitions of objects and cognition from concepts necessarily have different kinds of content, as one concerns intuitions and the other concepts, but they are nonetheless representations with a similar structure: cognitions of objects require the givenness of objects to the understanding, while cognitions from concepts require the givenness of concepts to reason.

3.4 The conceptual condition

The way in which I intend to read the functional equivalent of the conceptual condition with respect to analytic cognition will, at this point, perhaps be unsurprising: a subject’s representation counts as an analytic cognition only if it combines one or more features of a concept available to reason with a more general concept.

This reading is consistent with Kant’s characterisation of analysis, which is the ‘combination’ of partial concepts (‘marks that I already think therein’) with ‘my actual concepts’ (JL, 9: 59). Analysis is a distinctive mental state of combining a concept with a mark to make that concept distinct: one can cognise the concept ‘body’ analytically, ‘through the marks of extension’ (A8/B12). In analysing ‘body’, one combines it with ‘extension’ without going beyond the original concept. This cognition serves as the representational content of an analytic judgement. Kant’s suggestion here then is that analytic cognition involves combining the representational content of a concept (here, ‘body’) with more a specific representation of a mark (‘extension’). This fits well with my claim that particular features of a concept functionally satisfy the givenness condition while concepts satisfy the conceptual condition.

3.5 The consciousness condition

This leads us to the consciousness condition on cognition, which states that a subject’s representation counts as cognition only if their combining of intuition and concept renders them conscious of the material standards of correct representation. Clearly, given their a priori status, analytic cognitions do not make us aware of the material standard of correct representation, but a ready counterpart here would be formal standards. Might it be that analytic cognition makes us aware of formal standards?

Tolley briefly notes this analogy between material and formal standards as opening up the possibility for developing a Kantian reading of analytic cognition. He argues that, in analytic cognition, we are conscious ‘only of concepts (representations) and their contents’ along with their ‘interrelations’ but lack consciousness of the relation between the content of these concepts and that which can be represented by said concepts (Tolley 2020: 3229). I concur with Tolley here. Just what this entails, however, will depend on the precise notion of formal standards at work.

The natural place to look for a precisification of the notion of formal standards is in the first Critique’s section On the Supreme Principle of All Analytic Judgement. There, Kant writes:

[I]f the judgement is analytic, whether it be negative or affirmative, its truth must always be able to be cognized sufficiently in accordance with the principle of contradiction … Hence we must allow the principle of contradiction to count as the universal and completely sufficient principle of all analytic cognition; but its authority and usefulness does not extend beyond this as a sufficient criterion of truth. For that no cognition can be opposed to it without annihilating itself certainly makes this principle into a conditio sine qua non, but not into a determining ground of the truth of our cognition. (A151/B191)

This passage makes several claims orthogonal to Kant’s discussion of analyticity but relevant to an analysis of analytic cognition. First, he claims that if something is an analytic judgement, then it must be cognisable in a particular way. More specifically, it must be cognisable according to the ‘merely logical’ principle of contradiction (A153/B192). Thus, Kant is claiming that some kind of cognition (presumably an analytic one) is required for analytic judgements. ¹²

Second, with respect to analytic cognition, Kant claims that the necessary and sufficient standard for truth is the principle of contradiction. ¹³ Recall that, for Kant, cognition is a kind of representation; he is therefore making a claim here about the correct standards of representation. The reason that the principle of contradiction is a principle or standard of analytic cognition is because ‘contradiction entirely annihilates and cancels’ contradictory cognitions (A150/B190).

Kant’s reference to the cancelling out, and annihilation of, a contradictory cognition is metaphorical, given that it seems possible to mistakenly combine a ‘triangle’ with a ‘four-sided shape’ in an analytically false cognition. He means that contradictory cognitions are ipso facto false cognitions regardless of any possible or actual relation to objects. In this sense, the principle of contradiction is a negative criterion of truth that holds for ‘cognitions merely as cognitions in general’ (A150/B190). Kant claims that we can form the judgement that ‘No unlearned person is learned’, but the ‘negative proposition follows immediately from the principle of contradiction’ because the ‘mark of unlearned’ is ‘comprised in the concept of the subject’ (A153/B192). He expresses the contradiction in decidedly cognitive terms: produced by combining the mark ‘unlearned’ with the concept ‘learned person’. This combination produces a false analytic cognition and thus a false representation in the sense that it necessarily cannot be true. It is because we recognise that this analytic cognition is false that the negative proposition (i.e. that ‘No unearned person is learned is false’) immediately follows. However, the representation is not materially false, given that one is not representing actual or possible persons. Instead, the representation is false because it violates a formal standard and thus annihilates such a cognition.

The third of Kant’s claims comes when he states that the principle of contradiction is merely a sufficient, but not necessary, criterion of the truth of cognition. This is important because, as we have seen, Kant allows for material standards of truth to act as correct standards of representation for the cognition of objects. It is also important because – against the views of Leibniz and Wolff that all necessary truths are analytic – Kant wants to show that some necessary truths are synthetic. He nevertheless holds that the principle of contradiction is a standard for correctly combining concepts in analytic cognition.

We can now summarise the above account of analytic cognition with the following necessary and sufficient conditions:

1. Analytic cognition ( analytische Erkenntnis ): A subject’s representation counts as an analytic cognition iff:

2. (i) the subject’s representation renders a feature of a concept (mark) available to reason (the givenness condition);

3. (ii) the subject combines that mark with a concept (the conceptual condition);

4. (iii) the subject is conscious that they are representing a concept in abstraction from intuitions and thus according to the formal standards of correctly combining that feature and concept, that is, the principle of contradiction (the consciousness condition).

A point of clarification. The above conditions may create the impression that, on my view, analytic cognition and judgement are about concepts, in the sense that their semantic content is exclusively concept-directed. ¹⁴ This would conflict with Kant’s multiple suggestions that analytic judgements ultimately involve objects, despite the fact that we know their truth values by analysing concepts (cf. Vanzo 2014; Longuenesse 1998: 86–90). For instance, he uses the example of the analytic judgement ‘All bodies are extended’ to show that both analytic and synthetic judgements involve objects (JL, 9: 111; see also A68/B93). However, that impression would be misleading.

On my view, analytic cognition and judgement are about what Kant calls the concept of the object in general, not empirical objects (see Section 2). Kant’s object in general is the thought of a basic schema: to x, to which belongs a, also belongs b (x, a + b, for short). ¹⁵ To illustrate, consider his explanation of the analytic judgement ‘All bodies are extended’: ‘To everything x, to which the concept of body (a + b) belongs, belongs also extension (b)’ (JL, 9: 111; original emphasis). In other words, in an analytic judgement, the x thought under the concept of ‘body’ is also thought under the concept of ‘extension’. Kant contrasts this with the synthetic judgement ‘To everything x, to which the concept of body (a + b) belongs, belongs also attraction (c)’ (JL, 9: 111; original emphasis), holding that both analytic and synthetic judgements share the x, a + b schema. In this way, analytic and synthetic judgements are related in a and b being thought of as belonging to x; this holds independently of, and logically prior to, their analytic or synthetic relation. This basic schema is Kant’s concept of the object in general.

With the x, a + b schema, Kant commits to the philosophical point that we should begin with the object in general, before inquiring whether that object can be judged or cognised as possible or impossible, analytically or synthetically (A290/B346).

Logically impossible objects are objects with x, a + b schemas where a contradicts b, as in a two-sided square; they are non-existent and thus said to be ‘nothing’ (A291/B348). Empirical objects are what Kant calls really possible objects: objects whose ‘concept[s] agree with the formal conditions of experience in general’ (A220/B267). These conditions are those under which an object can be given to us through empirical intuition (e.g. by way of sight, taste, and colour). For Kant, if an object cannot be given through empirical intuition, it is really impossible: ‘the object of a concept to which no intuition that can be given corresponds is = nothing’ (A291/B347).

Yet, the concept of the object in general encompasses logically possible and impossible, as well as really possible and impossible, objects. All such objects, that is, objects in general, share the x, a + b schema. The semantic content of analytic cognitions involves this schema. Consider the analytic judgement ‘All bodies are extended’. The concept of the object in general is in this case the philosophical thought of some x as something to which, if the concept body belongs to it, so does the concept extension (i.e. the thought of x as something to which belongs a + b). The analytic cognition is the representation that combines the concepts ‘body’ and ‘extension’, plus the consciousness of the formal standards of such a combination. Thus, first, analytic cognitions are representations, not philosophical notions. Second, analytic cognitions are not the mere combining of two concepts but of doing so with consciousness. Yet, the semantic content of analytic judgements and analytic cognitions involves objects, although not logically or really possible objects as such, but rather logically or really possible (or impossible) objects in general, and so not, as such, the objects of empirical or mathematical cognition.

With this clarification in place, in the next section, I show how the above account of analytic cognition fits within Kant’s broader theory of cognition and how analytic cognition can ground assents involved in analytic knowledge.

Figure 1 summarizes our discussion of the various kinds of cognition.

Figure 1. Summarizes our discussion of the various kinds of cognition.

4. Analytic cognition, grades of cognition, and knowledge

4.1 Grades of cognition

A canonical place to find Kant’s account of cognition is in his characterisation of the various ‘grades’ of cognition with respect to their ‘objective content’ in the Jäsche Logic (JL, 9: 64). This characterisation ranges from representation simpliciter to comprehension in ways that go beyond mere cognition of objects by the understanding. ¹⁶ If, as I claim, analytic cognition can be accommodated within Kant’s theory of cognition, one may wonder how analytic cognition finds a place in the grades of cognition passage.

In my view, it is natural to view analytic cognition as a species of insight. Kant claims that there are two grades of cognition beyond cognition of objects – insight and comprehension – and he associates these grades with reason, not with the understanding. Kant says that to have insight into something, or to ‘cognize something through reason’, involves seeking to ‘perfect [our cognitions] as to content’ (JL, 9: 64–5). ¹⁷ Accordingly, in Kant’s view, a distinctive feature of insight is that it involves the perfecting of cognitive content. Indeed, the Jäsche Logic passage concerning grades of cognition is situated in the context of a discussion of the perfection of cognition. But just what does it mean to perfect our cognitions?

Kant takes the perfection of cognitive content to be achieved through distinctness, which can be either aesthetic or logical (JL, 9: 62). The perfection of cognitive content through aesthetic distinctness involves ‘clarity through intuition’, by which Kant seems to mean the closer investigation of the target phenomenon and thus the identification of more of its ‘remote marks’ that are connected to intuition, albeit ‘through a long series’ (JL, 9: 62). This kind of perfection is distinctness that is merely a liveliness of the understanding. It therefore cannot be what Kant has in mind when he talks of insight as the perfecting of cognitive content, given that he takes insight to be cognition through reason and not the understanding.

Rather, for Kant, the perfection of cognitive content through logical distinctness involves ‘clarity through concepts’ (JL, 9: 62). Kant then remarks that clarity through concepts comes in two varieties: analytic and synthetic distinctness. In so doing, he returns to one of his central criticisms of the Wolffian logicians: that they recognise only analytic distinctness and not the possibility of synthetic distinctness, which extends the content of our concepts through ‘what is added as a mark beyond the concept in pure or empirical intuition’ (JL, 9: 63). Kant returns to the examples, analysed above, of mathematical and philosophical cognition as instances of perfecting cognitive content through synthetic distinctness. What remains clear, however, is that analytic distinctness, by ‘mere analysis’, is an important means of perfecting our cognitive content (JL, 9: 64). As he puts it, ‘the analytic procedure for creating distinctiveness … is the first and principal requirement in making cognition distinct’ (JL, 9: 64). And this despite the fact that a central achievement of the Critical philosophy is the discovery of the synthetic procedure for making concepts distinct.

Thus, we can see how analytic cognition fits with the grades of cognition passage and have arrived at a more complete understanding of cognition than is typically allowed in discussions of cognition in Kant. Accordingly, Kant’s conception of insight involves either synthetic or analytic distinctness. Mathematical and philosophical cognitions are instances of insight involving synthetic distinctness. By contrast, analytic cognitions are instances of insight involving analytic distinctness.

4.2 Analytic cognition and analytic knowledge

Central to Kant’s project is showing that there is a tight relationship between the cognition of objects and empirical knowledge. So a discussion of analytic cognition should shed light on an analogous relationship between analytic cognition and analytic knowledge. Knowledge (Wissen), for Kant, is an assent – literally a holding a proposition to be true (Fürwahrhalten) – that has objectively and subjectively sufficient grounds (A822/B851; see Chignell 2007). To have objectively sufficient grounds is, for Kant, to have sufficient evidence in the form of experience, testimony, mathematical proofs, and/or inferential reasoning (JL, 9: 70–1). On the other hand, Kant takes it that to have subjectively sufficient grounds is to be aware of the objective grounds on which one forms assents (A820/B848). Thus, knowledge is distinct from cognition because, while knowledge is a kind of propositional attitude, cognition is a kind of representation. Subjects can clearly have cognition without knowledge by failing to form an assent of the kind distinctive of knowledge.

However, cognitions of objects are an important source of objectively sufficient grounds. It is my cognition (a representation), expressed in a judgement, that provides the objective sufficient grounds (if all goes well) that will warrant a knowledge claim. It is precisely my relation to an object via an intuition (in cognition of an object) that underwrites my knowledge of that object. On this basis, Kant has been read as being committed either to the claim that analytic knowledge is not genuine knowledge or to the claim that analytic knowledge is knowledge without a cognitive basis. ¹⁸

Although Kant does not explicate the relation between analytic knowledge and cognition, he clearly commits to the possibility of analytic knowledge in the Phenomena and Noumena section:

For an analytic [assertion] takes the understanding no further, and since it is occupied only with that which is already thought in the concept, it leaves it undecided whether the concept even has any relation to objects; … it is enough for him to know [Wissen] what lies in its concept; what the concept might pertain to is indifferent to him. (A258–9/B314)

As this passage indicates, Kant holds that we can know what lies in a concept by mere analysis. According to Chignell, even though we can legitimately be said to provide analyses of the concepts that we already have of objects, the ‘resulting knowledge will not always be based in cognition of those objects’ (Chignell 2014: 578; cf. Willaschek and Watkins 2020). I am sympathetic to Chignell’s reading. Analytic knowledge cannot have objective grounds in the cognition of objects. As the above passage states, analytic knowledge is knowledge in abstraction from the way in which objects are given and from the cognition of objects. However, the qualifier, ‘of objects’, is important because, given the above analysis of analytic cognition, there is a natural way to read the relation between analytic knowledge and cognition from concepts alone: analytic knowledge is an assent that enjoys objectively sufficient grounds in the form of an analytic cognition alongside subjectively sufficient grounds in the form of awareness of that cognition. ¹⁹ Such knowledge will not be grounded in the cognition of objects, but rather in cognitions from concepts alone. In turn, even though, as explained above, their semantic content is not exclusively concept-directed, such knowledge is not knowledge of objects but rather knowledge of concepts. This way of seeing the relationship between analytic knowledge and cognition preserves the connection between the objective grounds of knowledge and cognition and therefore provides epistemic standards for analytic knowledge.

Thus, not only can we place analytic cognition within Kant’s theory of cognition, but we can see how analytic cognition can ground assents involved in analytic knowledge.

5. Objections and replies

Having presented the textual and conceptual case for my reading of analytic cognition in Kant, I will address what I take to be the two most substantive objections to my proposed interpretation. One objection might be that my proposed account of analytic cognition is at odds with some of the most canonical descriptions of cognition in the first Critique, as my proposal allows for a kind of cognition without intuitions, namely, analytic cognition. For instance, in an oft-quoted passage, Kant says:

Intuition and concepts therefore constitute the elements of all our cognition, so that neither concepts without intuition corresponding to them in some way nor intuition without concepts can yield a cognition. (A50/B74)

This passage appears to confirm that all cognition requires intuitions and concepts. Given the ubiquity of similar passages, the weight of textual evidence suggests that Kant’s official use of ‘cognition’ refers to this definition and thus rules out my proposed reading of analytic cognition.

In response, I insist again upon the importance of the distinction between cognition of objects and cognition simpliciter. We see this ‘of objects’ qualifier appear in several other important passages. In the Transcendental Deduction, for instance, Kant says that ‘two components belong to cognition: first, the concept, through which an object is thought at all (the category), and second, the intuition, through which it is given’. On close inspection, we see that Kant is talking about what it means to ‘cognize an object’ as opposed to merely think it (B146). Later, he similarly claims that ‘we cannot cognize any object that is thought except through intuitions that correspond to those concepts’ (B165, my emphasis). In the Prolegomena, Kant says that ‘it would be an absurdity for us, with respect to any object, to hope to cognize more than belongs to a possible experience of it’ (P, 4: 350, my emphasis; cf. B1, A19/B33, A90/B122, A95/B129, B289, B166). Once we are aware of the qualifier, ‘of objects’, we see that Kant often employs it in such a way as to leave conceptual space for the possibility that analytic cognitions are cognitions from concepts.

A second objection is that analytic cognition in Kant’s work always arises in the analysis of empirical concepts (or so the objection goes), but my proposed conception of analytic cognition includes both a priori and empirical concepts. Hanna, for instance, claims that analytic cognition counts as cognition because analytic propositions contain ‘only empirical concepts’. Thus, he claims, the analytic proposition ‘Bachelors are males’ has ‘primitive objective validity’, by which he means that even though the proposition is analytic its concepts relate ‘by means of our sensory empirical intuition’ to possible or actual objects of experience (Hanna 2001: 93–4). The force of this reading is that it can provide us with a way of reading analytic cognition as retaining intuitive content via empirical concepts and thus enable us to preserve the view that cognition must involve intuition while accounting for analytic cognitions.

There are strong reasons to resist this objection. First, Kant claims that all analytic judgements are a priori cognitions whether ‘their concepts are empirical or not’ (G, 4: 267; A47/B64; cf. JL, 9: 142; Refl 6050, 18: 437) and that we can ‘first cognize’ concepts analytically (A8/B12). These are explicit endorsements of the possibility of analytic cognition of non-empirical concepts. Second, Kant repeatedly claims that we can analyse empirical concepts. In the Prolegomena’s discussion of analytic cognition, he claims that ‘all analytic propositions are still a priori judgements even if their concepts are empirical, as in: Gold is a yellow metal’ (G, 4: 267). Such claims, however, suggest that empirical concepts are a sufficient, but not a necessary, condition for analytic cognition. Kant does not claim that we can analyse only empirical concepts and he therefore seems committed to the possibility of the analytic cognition of non-empirical concepts.

Furthermore, one of Kant’s preferred examples of analysis involves a non-empirical concept, namely, virtue (Log-Bl, 24: 131; JL, 9: 35; Log-W, 24: 834). In the Vienna Lecture, Kant says that ‘with analytic cognition, [one makes] a given concept distinct’. He continues, ‘E.g., I cannot explain virtue synthetically. For I am supposed to say what we all think under the concept of virtue, not what I perhaps understand under this concept in accordance with my own caprice’ (Log-W, 24: 845). This ‘caprice’, Kant thinks, involves viewing virtue as an empirical concept, rendering it ‘an ambiguous non-entity, changeable with time and circumstances’ (A315/B371). Thus, for Kant, virtue must be a non-empirical concept. So the relevant analytic cognition here involves a theoretical analysis of virtue’s component parts, which Kant lists as ‘readiness in lawful actions … a readiness in actions, … and domination of the inclinations, and so on, and so on’ (Log-W, 24: 847). Kant’s discussion here strongly supports my view that analytic cognition can involve both empirical and non-empirical concepts. Thus, we have strong reasons to reject the view that Kant’s references to analytic cognition can be understood merely in terms of the analysis of empirical concepts.

6. Conclusion: cognition’s general root

In this article, I have aimed to show how we can account for Kant’s references to analytic cognition. I have argued that analytic cognition (1) is a unique kind of representation that satisfies forms of the givenness, conceptual, and consciousness conditions, (2) is what Kant characterises as insight involving analytic distinctness, and (3) can ground assents involved in analytic knowledge. Central to my account was reading the givenness condition on analytic cognition: thus, the subject’s representation renders a feature of a concept (mark) available to reason. In conceptual analysis, a subject becomes aware of a concept’s marks by considering that concept.

To conclude, I wish to highlight an implication of inquiring into the possibility of analytic cognition in Kant: we can derive a general form of the givenness, conceptual, and consciousness conditions encompassing empirical, mathematical, philosophical, and analytic cognition. This explains Kant’s suggestion in the Architectonic of Pure Reason that reason and the understanding share a genus: ‘the general root of our cognitive power [allgemeine Wurzel unserer Erkenntniskraft] divides and branches out into two stems, one of which is reason’ (A835/B863) and the other of which is presumably the understanding. ²⁰ Kant does not explain what he considers to be the genus of all cognition. However, by spelling out the three conditions on cognition in their most minimal sense, we can define cognition’s general root.

1. Cognition’s general root: A subject’s representation counts as a cognition simpliciter iff:

2. (i) the subject’s representation renders available to a faculty a feature as an individual feature (the givenness condition);

3. (ii) the subject’s representation combines that feature with a concept (the conceptual condition);

4. (iii) the subject is conscious of what they are representing and thus of a standard of correctly combining that feature and concept (the consciousness condition).

The subject’s representation renders available to a faculty a feature as an individual feature (of an object or concept); thus, the subject will be able to attend to that feature and can thus conceptualise with regard to it with consciousness of the (material or formal) standards of correct representation. Cognition’s general root is a broader notion than empirical cognition, and we should thus affirm, as many do, that Kant operates with broad and narrow definitions of cognition. My analysis, however, also specifies this broader notion in a technical sense. Kant’s suggestion that all cognition shares a general root remains underexplored in current scholarship, which presumably explains why analytic cognition is so often read as an instance of loose talk. With regard to ‘analytic cognition’, I see no reason to think that Kant is not talking about cognition’s general root – in this broad, yet still technical sense.

The possibility of uncovering cognition’s general root would have excited many of Kant’s immediate successors. One of the chief projects, in post-Kantian philosophy, was to discover the root of our various faculties. This possibility ought to excite us today, given that the spirit of Kant’s Critical philosophy is an unfettered attempt to search for unity in any assumed division.