DENNIS SCHULTING | Kant’s Radical Subjectivism. Perspectives on the Transcendental Deduction | Palgrave Macmillan 2017


By Dennis Schulting

I thank Sacha Golob for his challenging and interesting notes on my stance on Kantian nonconceptualism, in particular also on the concept of objectivity, one’s view upon which is crucially related to how one positions oneself in the debate about nonconceptual content in Kant. It may seem from reading Golob’s criticisms that I’m not at all sympathetic to the core idea of nonconceptualism, namely the possibility of one’s representations having nonconceptual content in terms of being directed at given spatiotemporal objects independently of the application of concepts, or indeed of animals having some kind of consciousness of objects in their surroundings, with which they interact in multiple complex ways. My actual position is much more nuanced though.

However, it is true to say that my interpretation of Kant’s position does not allow for the strong form of nonconceptualist objective* intentionality that Golob argues for in his commentary and elsewhere (see Golob forthcoming), or for any strong form of non-categorially-constituted objectivity.

I find Golob’s analyses of the passages in the Second Analogy that he quotes in defence of an account of the ‘various senses of objectivity’ that Kant supposedly espouses, in particular objectivity*, not convincing and I cannot subscribe to the Longuenessian claim that he adopts regarding the relation between the categories and judgement: in my view, there is no possible ‘over-alignment’ of the categories and the logical forms of judgements, for one cannot have categories without the logical forms of judgement—whichever way one interprets their relation—since categories are nothing but the logical functions insofar as they determine an object for an intuition (B143) (see my critique of Longuenesse on this point in Schulting 2012/2018, ch. 3). Categories either are or are not instantiated, and they are instantiated in judgements only. There is no room, in my view, for so-called ‘first’ and ‘second applications’ of the categories as on Longuenesse’s reading, on pain of inviting an explanatory regress.

I should also like to note straight out that some of the passages that Golob quotes from my book as putative evidence for the claim that my account stands in tension with the reality of animal perception of spatiotemporal particulars are taken out of a context where animal perception is not at all discussed, let alone denied. I shall come back to this in Section III below.

Furthermore, from his proposals in Section 3 of his essay it seems that Golob is not sufficiently appreciative of one of the key claims of Kantian idealism, namely that the objects of our experience are a function of our cognition, and so do not exist apart from it; as such, they are not given independently of the activity of a priori synthesis (a point that I discussed in some detail in defence of my thesis of Kant’s radical subjectivism), though from an empirical point of view, objects are of course necessarily given as that on which given intuitions depend (cf. V-Met/Schön, AA 28:484 and B72). I also think he mixes up the transcendental and empirical levels of reasoning in Kant’s story and the different kinds of consciousness that are at play, namely transitive and intransitive consciousness. It seems to me that the kind of object-directed mode of experience for animals that Golob argues for would require a form of transitive consciousness that, at least for Kant, is not the kind of consciousness that animals have;[1] I agree that, in a purely phenomenological sense, animals could be said to be conscious of the objects that they perceive—but this does not involve a form of Kantian object-directedness that I take Golob to be arguing for. I do not believe that such an argument is available within the Kantian framework.

So there is no getting round it: there is fundamental disagreement between our positions on the prospects for Kantian nonconceptualism. But there are elements in Golob’s account which I agree with and are congruent with some of my articulations in the book, though Golob seems to present them as contradicting my interpretation, such as the fact that Kant allows for forms of binding that are not that of ‘necessary combination’. Nevertheless, Golob’s critique presents me with the opportunity to shed more light on my take on the debate on nonconceptualism, and how Kant should be situated in it, in particular concerning the concept of objectivity, its role in human experience, and what implication this has for Kantian takes on animal perception. Before I address Golob’s criticisms in the detail that they deserve (in Sections II and III), I first want to make a few general comments on the debate as such (Section I).

I.  Issues about Kantian nonconceptualism

Very recently, in a generally sympathetic review of the book that is here under discussion (henceforth KRS), Tim Jankowiak wrote the following on my view of Kant as a moderate conceptualist:

It is worth noting that moderate conceptualism might as well also be called a moderate non-conceptualism, since it ‘leaves room for non-conceptual mental content in some minimal sense’ (Schulting 2017:198). Here again his position can be difficult to pin down. On the one hand, he wants to argue that all intuitions that contribute to objectively valid cognition and reference to objects will necessarily be combined with the categories:

Necessarily, if intuitions are to be seen as contributing to possible knowledge of objects, then intuitions are subsumed under the categories as the conceptual conditions under which knowledge of objects is possible. (2017:236)

This leaves room for the possibility of non-conceptual intuitions that are not subsumable under the categories and which won’t enable reference to objects (see Schulting 2017:197, 236). Thus the position seems to be that the majority of our intuitions are conceptual, even though some are not. Yet on the other hand, he also argues that ‘intuition as such is not necessarily or at least not yet categorially determined. […] Intuition is not in and of itself conceptual or even proto-conceptual’ (2017:239–40). Rather ‘it is the relation between intuition and concept which is conceptual’ (2017:239). Putting these two elements of his interpretation together, it might sound like Schulting is saying that (objectively referring) intuitions both are and are not necessarily conceptual. However, a more charitable reading would take him to mean that intuitions cannot perform their function of referring to objects except insofar as they are combined with concepts in actual judgments. Nevertheless, when the position is put this way, I worry that it turns the conceptualism debate into a mere verbal dispute. For the position is that intuitions i) are not in themselves conceptual, ii) are not even all conceptualizable, and iii) succeed in referring to objects only when combined with concepts in judgment. This position ‘could’ be called a form of conceptualism, but I suspect that many self-proclaimed non-conceptualists would be happy to embrace all three claims as well. For I would have thought that the conceptualism debate was about the question whether intuitions ‘in themselves’ are conceptually structured (perhaps because they are formed by a conceptually guided process), not whether intuitions need to be combined with concepts in judgments in order to enable objectively valid reference to objects. (Jankowiak 2018:121–2, my underlining)

Though in a standard review one is of necessity constrained by space limits, which makes substantiating one’s critique quite a feat, Jankowiak’s rather summary portrayal of my position on Kantian nonconceptualism might give a reader the wrong impression of my views: clearly, I am not saying nor even suggesting, anywhere in the book, that ‘(objectively referring) intuitions both are and are not necessarily conceptual’. Jankowiak has selectively taken a few text fragments out of their contexts, so that prima facie my view looks contradictory. But even based on the quotations he provides, the contradictory conclusion does not follow. He says further that on my reading the non/conceptualism debate is vulnerable to becoming ‘a mere verbal dispute’ (2018:121), for my reading glosses over, as Jankowiak appears to suggest, the very question around which the debate circles: the possibility of nonconceptual representational content, which is either denied or affirmed depending on which side of the debate one chooses. Golob charges the same: my alleged

binary approach with respect to nonconceptualism […] fails to engage with what was really at stake in the existing debate. Both conceptualists and nonconceptualists agree that what is at issue is the possibility of nonconceptual intentionality.

But I find this a puzzling charge. I repeatedly argue, as Golob rightly notes, that nonconceptual intentionality is incompatible with Kant’s views because objective validity is a function of apperception/synthesis/judgement only, so how is that failing to engage with what is really at stake in the debate?

Moreover, one of the central elements of the debate about Kant’s putative conceptualism or nonconceptualism is the important question, raised by the discussion in the run-up to the Deduction (§13), whether intuitions require the functions of the understanding (the categories) to be intuitions. Conceptualists argue that they do, nonconceptualists deny this. While I concur with the nonconceptualists that intuitions do not require the categories to be intuitions, I argued that the issue of the precise relation between them is a bit more complicated. The crucial issue is how we should understand the modal nature of that relation, which is a direct result of the Copernican question about the conformity of objects to the a priori forms of our intuition and those of our understanding. That is, the objects of our intuition and the understanding are a function of those forms: those forms constitute the possibility of the objects as objects of our intuition and understanding, and only as such. Further, Kant explicitly distinguishes between our intuition and our conceptualisation of objects—this distinction is absolute and non-negotiable, for it relates to the different definitions that Kant gives for intuitions and concepts respectively based on his fundamental critique of the rationalist belief that conceptual relations define the world down to the lowest species. That means, by implication, that the relation between the intuition of the object and the a priori concepts that are applied to it must be read as a conditional, not as absolute.

This is the interpretative-systematic question that is important for an assessment of the very thrust of the Deduction (also referred to as the so-called ‘two step’ problem) and which I discussed at length in Chapter 5 of the book, as well as in Chapter 7. The further question of the possibility of nonconceptual intentionality is of course relevant, but it is one requiring an evaluation of what an intuition actually is. In Chapter 5, I focused on the interpretative question about the Deduction’s argumentative thrust in relation to the conceptualism/nonconceptualism debate, not specifically on the meaning of ‘intuition’. But the question of how nonconceptualists and conceptualists see the relation between intuitions and categories has of course a direct bearing on the question whether there can be intentionality or object-directedness independently of the categories or concepts. I believe, and argued, that the latter is not possible within the framework of the Deduction because objective reference is a function of categories being applied only.

My rejection of the possibility of nonconceptual intentionality might thus appear to imply that I explicitly come out on the conceptualist side of the debate because in denying this possibility my position cannot in any straightforward sense be considered nonconceptualist (in the definition of the debate). Hence, Golob thinks my interpretation is a reflection of the ‘dominant’ position, i.e. ‘it is just another conceptualism even if it does avoid the most extreme versions of that doctrine’—which is rather ironic, since Jankowiak seems to think, in the above quotation from his review, that the moderately conceptualist claims I’m making are entirely palatable to any nonconceptualist, and in fact do not differentiate my reading as particularly conceptualist. But as I shall point out in Section II, my denying the possibility of nonconceptual intentionality does not imply that I must be a card-carrying conceptualist full stop (such as Hannah Ginsborg, whom Golob quotes). Neither—and this is important to stress—does it imply that my moderate conceptualism cannot accommodate the real possibility of animal perception or consciousness or indeed nonconceptual human consciousness.

I should like to stress—and I thought I made this clear in the book, and especially in the chapter that is entitled ‘Problems of Kantian Nonconceptualism and the Transcendental Deduction’[2]—that my aim was and is not ‘a compromise’, to use Golob’s words, between nonconceptualism and conceptualism; nor can my position be pigeonholed in either one of the categories, either non-conceptualism or conceptualism. I am not interested in the question whether Kant is to be considered a conceptualist or a nonconceptualist. If the question is put this simplistically, the answer seems to me quite clear and simple: neither, for both extremes are wrong. If this makes my position ‘difficult to pin down’, then this might be because Kant himself is more complex than the debate allows, and cannot be pinned down in the terms of the debate. Unfortunately, the recent surge in articles that argue explicitly either pro or contra Kantian nonconceptualism (see the references in the introduction to my Kantian Nonconceptualism; Schulting 2016) does make it seem that Kant can be pigeonholed as either a conceptualist or a nonconceptualist.[3]

That Kant’s position is more complex has been shown by the fact that Lucy Allais, in the article that initiated the debate (Allais 2009),[4] made a point—quite rightly—of the fact that Kant is not a strong or extreme nonconceptualist (Golob also notes this, by acknowledging that there ‘are, of course, multiple competing theories of intentionality, on both the relationalist and representationalist ends of the spectrum’). The relative nonconceptualism she espouses is, to an extent, compatible with my moderately conceptualist reading. As Jankowiak notices, my position can be considered one the claims of which ‘many self-proclaimed non-conceptualists would be happy to embrace’, and indeed a relative nonconceptualist agrees with those three above-mentioned claims (as I made clear in the book) (see e.g. Allais 2009:386, 400), and to that extent my moderate conceptualism is logically also a moderate nonconceptualism. My view concurs with Allais’s on this point. The truth is of course always somewhere in the middle, never in the extremes. But Jankowiak uses this apparent equivocation to show up an apparent contradiction in my reading. Supposedly, my moderate conceptualism is not at all differentiatable from nonconceptualism insofar as these three claims are concerned, and I skirt around the one central issue in the debate: are ‘intuitions “in themselves” … conceptually structured’ or not? But Jankowiak fails to notice that again and again I make a point of denying that intuitions are in and by themselves conceptually structured, which brings my position closer to nonconceptualism, but also—and this is one of the pillars of my critique of nonconceptualism, including Allais’s though she is much more careful in her formulations than others[5]—that nonconceptualists are wrong to think that intuitions have objective validity all by themselves, independently of the categories. The contradiction, if there is one, is not an inherent fault of my position, but it stems from what I regard as the central mistakes of both nonconceptualism and strong forms of conceptualism: the former mistakenly assumes that intuitions refer objectively all by themselves, the latter wrongly believe that ‘intuitions “in themselves” are conceptually structured’. If those were the two options, then neither of conceptualism and nonconceptualism are true. End of story.

My position is a critique of both readings; any moderating feature implies that one’s position is always both to an extent nonconceptualist and to an extent conceptualist. So what is so confusing about my moderate conceptualism? If that critique shows up contradictions, then that demonstrates that, more than anything else, there is something fundamentally wrong with foisting the debate surrounding the possibility of nonconceptual representational content on Kant. The extreme conceptualist and nonconceptual readings seem both uncharitably rigid readings of Kant,[6] so it is not surprising to see so many modifications of the extreme positions (strong and moderate or absolute and relative conceptualism or nonconceptualism, intellectualist and sensibilist interpretations, etc.), and my attempt at presenting a moderate form of Kantian conceptualism might be seen as one more such modification, which only leads to more contradictory or confusing views rather than to a consensus view. Perhaps it would be best to leave the debate for what it is and start again reading Kant’s text without imposing any preconceived frame or model!

At any rate, my critique of nonconceptualism is crucial, namely that intuitions do not all by themselves refer objectively, and so my moderate conceptualism would not at all be acceptable to standard nonconceptualists (relative nor strong nonconceptualism), as this element is fundamental to nonconceptualism. My critique of strong forms of conceptualism is equally pivotal: I deny the claim that strong conceptualists endorse, namely the modally intemperate claim that all intuitions must be conceptually (categorially) structured to be intuitions, or even less intemperately, the claim that all intuitions must at least be ‘formed by a conceptually guided process’, as Jankowiak puts it. Why this two-pronged critique does not lead to ambiguity let alone contradiction is made clear by my claim, fleshed out in Section 5.6 of the book, that Kant’s argument must be read modally. Jankowiak quoted a phrase of mine that points that out (see the quotation above), but he ignores the modal element and its implications for evaluating the Kantian nonconceptualist and conceptualist positions. This claim can basically be put in the following way:

1. For intuitions to refer objectively they must be categorially determined.

2. Intuitions are categorially determined if and only if they are so determined by an act of the understanding.

This dual claim is not a trivial claim that nonconceptualists and conceptualists alike agree upon. Conceptualists standardly interpret the above claims in a modally intemperate manner, namely as implying that

3. Necessarily, intuitions are categorially determined

for conceptualists claim that, as Jankowiak puts it,

4. ‘intuitions “in themselves” are conceptually structured’.

The claim is that the conceptual (i.e. categorial)[7] structure is what makes intuitions intuitions (and not just a bunch of sensations).[8] But 3 and 4 do not at all follow from 1 and 2. A modally temperate claim, on the other hand, does agree with 1 and 2, articulated for example in this passage in my book that was quoted by Jankowiak:

5. ‘Necessarily, if intuitions are to be seen as contributing to possible knowledge of objects, then intuitions are subsumed under the categories as the conceptual conditions under which knowledge of objects is possible.’ (p. 236)

Standard (‘dominant’) conceptualists cannot accept claim 5 because the conditional leaves open the real possibility that some intuitions are not categorially determined (i.e. subsumed under the categories) and so intuitions are not ‘in themselves’ conceptually structured. By contrast, nonconceptualists argue indeed that some intuitions are not or must not be categorially determined (conceptualised) and I entirely agree (this makes my reading in some minimal sense nonconceptualist). However, while nonconceptualists can agree on the modally temperate interpretation of claim 2, they reject claim 1.

As is now evident, 5 in conjunction with both 1 and 2 is a genuine alternative to both the conceptualist and nonconceptualist interpretations of the relation between intuition and the categories. This is my moderately conceptualist interpretation, which concerns a modally temperate claim about the way in which intuitions entail categorial determination, namely, only so as to contribute to knowledge, given that there is knowledge (the Faktum of knowledge being the premise of Kant’s analysis). My interpretation is not a compromise, but an alternative.

To put the above somewhat differently by way of summarising my viewpoint, I agree with the conceptualists that objective validity is a function solely of transcendental apperception/the understanding/categories/judgement, and not a function of intuition independently of transcendental apperception/the understanding/categories/judgement. There are of course variations possible with respect to the question of how intimately apperception, the understanding, judgement and the categories are in fact related. I side with the strong conceptualist here, who sees these as all inseparably connected. But it seems that Golob agrees here (for the most part at least). So if Golob accepts that, as I argued, figurative synthesis cannot be decoupled from apperception/the understanding, he would argue that intuitions can have objective validity independently of figurative synthesis (and that is what I take him to argue).[9] Golob diverges from the standard (strong) nonconceptualism espoused by for example Robert Hanna, who argues that figurative synthesis figures in the objective reference of intuition independently of apperception /the understanding /the categories /judgement. On the other hand, Golob finds that my reading tends to ‘over-align’ the categories and the logical functions of judgement. So there is some ambiguity in Golob’s critique. The categories are nothing but so many functions of a priori synthesis or acts of apperception, so they cannot be divorced from figurative synthesis, but since categories are functions of judgement and the unity of apperception in fact defines judgement, it is hard to see how they are operative outside the context of judgements.

In any case, I agree with the nonconceptualists that intuitions do not necessarily entail their being subject to the categories, synthesis, the understanding, or apperception—i.e. to being ‘conceptualised’ or categorially determined. Some intuitions are not or even cannot be subsumed under the categories. But I disagree with the nonconceptualists on the objective validity or reference of intuitions independently of the categories /synthesis/ understanding/ apperception. This means that my position cannot be a compromise between standard or ‘dominant’ conceptualism and standard nonconceptualism because I disagree with a fundamental aspect of both positions. In the book (and certainly in Schulting 2015b), I make abundantly clear that both the conceptualist and nonconceptualist construals of Kant’s Deduction face problems that stem from a misunderstanding of the core argument in the Deduction. I thus disagree with both the standard or ‘dominant’ conceptualist position and all varieties of Kantian nonconceptualism on offer in the literature, since especially the latter fail to see the ‘radically subjective’ cornerstone of the Deduction, and Kantian nonconceptualism quintessentially espouses the mistaken view that objectivity is also a function of intuitions (i.e. that intuitions are objectively valid or objectively refer). Intuitions depend on given objects (B72), but that does not make objectivity a function of intuition. Precisely the idea that Golob rightly singles out as characterising the nonconceptualist position, namely the objective intentionality of intuitions independently of apperception and the categories, I find hugely problematic.[10]

Concluding this section, my ‘radically subjective’ interpretation of the Deduction is thus neither conceptualist nor nonconceptualist as defined by the debate on Kant’s alleged conceptualism or nonconceptualism, and since I criticise fundamental elements of both, my position cannot be considered, and wasn’t intended as, a compromise. In fact, I seriously think that the whole debate is misguided if it is meant as a way to understand Kant’s position.[11] The truth about Kant’s position lies not somewhere in the middle (a ‘compromise’ between conceptualism and nonconceptualism), but somewhere else entirely. I believe Kant’s position is defined by neither conceptualism nor nonconceptualism, or to put it somewhat harshly, the parameters of the debate are ill-suited to understand the Deduction.

But, rather than carrying on with waging a war of words over the definition of the terms of the debate and whether or not I fail to address its principle concern, I want to look now at the far more interesting and challenging substantial critique of my position on nonconceptual intentionality that Golob has to offer. This concerns specifically the possibility of animal perception of ‘spatiotemporal particulars standing in quite specific relations’ that Golob raises as a challenge to my ‘moderately conceptualist’ stance, and the multiple senses of objectivity Golob says are at work in Kant’s text. I address these in the next and following sections (in reverse order).

II. Objectivity and consciousness

Let me first say something about the notion of ‘objectivity*’ that Golob differentiates from the notion of objectivity as it is defined in the Deduction (§17). Golob defines objectivity* in the following way:

A visual experience E is objective* iff E represents a distinction between spatiotemporal particulars and the mental states of the subject of that experience.

This notion of objectivity is supposed to enable a conception of objectivity that is not a function of the unity of apperception (and thus of judgement)—which it is on Kant’s understanding in §17 in the Deduction—but nonetheless guarantees a particular perspective by a subject or an ego-centric perspective vis-à-vis an object that can be differentiated from that perspective. On the one hand, I suspect that what Golob means by objectivity* is partly covered by what I referred to as Kant’s Gegenstand, which lies at the root of any intuition of a given object (A19/B33), is not a mental state but is rather that which is outside us spatially, and provides us with impressions, which are causally related to it. On the other hand, I do not think that objectivity* as a kind of primitive object-directed intentionality—animal perception or the intentionality of infants, say—is something that is compatible with the Kantian transcendental framework. Object-directed intentionality is the prerogative of the subject that, in Copernican terms, sees objects in terms of their conformity to the forms of her intuition and understanding. Anything outside that purview cannot count as objective or as having object-directed intentionality. But this does of course not mean that animals and infants, say, have no immediate awareness of their surroundings as a result of sensory input. Succinctly put, in my view it is either objectivity, as Kant defines it in the Deduction, or it is pure immediacy—i.e. any set of representations that are sub-objective or merely subjectively valid do not qualify as objective or object-directed in the strict sense, but can and do involve being affected by objects and the behavioural responses that result from it. The Kantian nonconceptualist cannot have his cake and eat it. (I shall elaborate further below on the meaning of ‘immediacy’.)

Though it would be perfectly legitimate in general—Golob suggests, with good reason, that ‘objectivity’ is a technical term that has different meanings for different philosophers, and even, as Golob notes, Kant himself points to a loose sense of the term ‘object’ at A189–90—to differentiate between various senses of objectivity and to use the same term in all these cases (even if annotated by an * or whatever other signifier is preferred) undermines one of the central arguments of the Deduction (in §17 of the B-Deduction, and A103–10 in the A-Deduction), namely that objectivity is constituted first and only by the unity of apperception, and that there is no objectivity, strictly speaking, apart from the unity of apperception. To argue otherwise, or to weaken that link, is in my opinion to undermine the thrust of the Deduction, and the core claim of the Copernican turn: that objectivity is first constituted by the conformity to the forms of our intuition and understanding. There is also a philosophical reason not to go the route of redefinition and modification in order to fit any sub-transcendental notion of objectivity into the Kantian transcendental framework (see below). But, as Golob observes, the issues are of course not merely definitional—on second thought, I should have refrained from giving the impression, on one occasion, that Kant rigorously distinguishes between the terms Objekt and Gegenstand (p. 21); as Golob correctly notes, Kant does not stick to such a strict differentiation and uses both terms interchangeably.

If we look at the issues form a more systematic point of view, to posit an objective* intentionality for animals, in the sense that Golob proposes, strikes me as anthropomorphic and as revealing a conflation of the transcendental and empirical levels of reasoning. These levels are to be distinguished strictly on account of the Copernican turn, which proposed to look at the justification of the use of concepts from the perspective of the self-knowledge of reason, that is, from the perspective of rational self-legitimation without having to have recourse to empirical verificationism. It might be that Kant is not sufficiently phenomenologically sensitive to possible kinds of intentionality other than the one that is governed by transcendental laws (see Golob, forthcoming). But it is important to point out, I think, that to try and imagine a non-human ‘perspective’ or ‘intentionality’ is still always an attempt from our human rational perspective, namely the perspective in which the very idea of a perspective is fundamental, that is, the perspective of the self-knowledge of human reason. I do not think an animal has a viewpoint—in the sense of what it is for something to be for it[12]—in the way that we have a viewpoint on some object. Obviously, animals engage with their environment, often in very intricate ways (an eagle, say, as Golob explains—I return to this example below), but they do not represent to themselves what they engage with in the way we represent to ourselves what we see and interact with—and this representing to oneself one’s own representing, being aware of one’s representing is,[13] I believe, essential to ‘having a viewpoint’.

In other words, animals lack reflexivity, namely the act of representing one’s representing to oneself, which first differentiates the subject of representation from the object of representation and vice versa, as Christian Wolff already made clear (this is an important element that Kant picks up from Wolff). Udo Thiel has rightly pointed out that for Wolff self-consciousness is ‘doubly derivative’, that is, ‘it depends on the consciousness of objects, and on the consciousness of our mental act of distinguishing that is involved in the consciousness of objects’ (Thiel 2011:308). But it also means that, reciprocally, we could not be conscious of objects if we were not self-conscious. This element of reflexivity is fundamentally lacking in animals.[14] See for example this early passage from the Metaphysik Herder:

Animals (by hypothesis) have a faculty for acting according to choice, but they cannot represent to themselves the motivational grounds: they are not self-consciously acting on these at will. (V-Met/Herder, AA 28:99, emphasis added)

Of course, one could say that animals also have a particular outlook from which they have awareness of their surroundings, and that is certainly true, but this would not yet be an object-directed intentionality nor would it imply that animals have mental states which contain an intended object within themselves. An animal represents, but it is a substantia bruta repræsentativa, not an intelligentia. That animals do not have intentionality, according to Kant, is clearly shown by a passage in a lesser work of the mid-80s (hence, in the Critical period):

For there is in cattle, as well as in the human being, that remarkable faculty we call imagination, the principle of perception and motion by which things that are absent can really exist in the mind as though they were present, as can things that never have been and perhaps never can be. But in cattle, this force is not directed by any choice or deliberate intention of the animal [haec vis non arbitrio quodam ipsius animalis et deliberato proposito[15] regitur], but is put into play by stimuli and impulsions implanted by nature itself, apart from any influence of the will. (De medicina corporis, AA 15:944, emphasis added)

As to animal consciousness, Kant often seems to deny that animals have consciousness. In a passage in the Philosophical Encyclopædia of 1775, he writes:

The main, and nearly the only, difference between animals and humans is consciousness, but that is also so great that it can never be replaced with something else. Many animals behave and build so craftily that they come quite close to humans, but all are without consciousness. (PhilEnz, AA 29:44–5)

But it should be noted that the consciousness Kant likely means here is transitive consciousness (see Schulting 2015a). See also the following passages in the Pölitz lectures and in a later lecture note:

Accordingly we attribute to [animals] a faculty of sensation, imagination, etc., but all only sensible as lower faculties, and not connected with consciousness. We can explain all the phenomena of the animals from this outer sensibility and from the mechanical grounds of their bodies, without accepting consciousness or inner sense. The philosopher must not increase the principles of cognition without cause. (V-Met-L1/Pölitz, AA 28:277)

[C]onsciousness is wholly lacking in animals, their behavior occurs according to laws of the power of imagination, which nature has laid in them. (V-Met/Dohna, AA 28:689–90)

In a late letter, Kant writes:

[First, in] the division of the faculty of representation, [there is the sphere of] mere apprehension of the representation: apprehensio bruta without consciousness (which is solely for animals), and the sphere of apperception, i. e. the concept; the last comprises the whole of the sphere of the understanding. (Br, AA 11:345)

When Kant denies animals consciousness, it is self-consciousness (or inner sense) or transitive consciousness what he denies them, not creature consciousness and a certain awareness of their surroundings. Objectivity in Kant’s sense is, however, a function of apperceptive subjectivity (this is basically the thesis of subjectivism that I defend in the book). Animals are not self-conscious in Kant’s sense, i.e. apperceptively subjective. Because they are not subjects, animals by implication do not have a sense or ‘idea’ of objects in the strong Kantian sense—objects qua objects; they perceive objects, but do not perceive them as objects. But it should also be borne in mind that there are in fact no objects strictly speaking outside of the objectivity that is born of apperceptive subjectivity, that is, objects external to the transcendental self. It concerns here the essential, philosophical meaning of the term ‘object’ (for Kant at least): that which is positioned ‘over against’ (A104–5); an object is as it were projected over against the subject by that subject,[16] and does not exist outside that subject—qua object, that is, since as a thing in itself it does of course exist outside the subject (see the references in KRS, ch. 1 ).

This is the whole point of Kant’s transcendental (Copernican) turn: there is no object just given over there, to which we, perceivers, cognisers, knowers, subsequently simply latch on, let alone could have a priori knowledge of if it were given beforehand. Of course, in a way something is given that we can label ‘an object’ in the ordinary (manifest) sense of the word (what I referred to as Gegenstand), but specifically not in the philosophical sense of the word. To put it differently, we cannot just assume that there is an object sensu stricto given, with respect to which animals, infants, human beings or what have you adopt, each in their own way, a subjective perspective.

The transcendental perspective does not apply to the animal case. To say that animal experience or perception refers to an object in a way that differentiates the object from their perspective on it, as Golob suggests, is, as I pointed out earlier, to mix up the transcendental and the empirical levels. On the empirical level animals are naturally as much part of the objective world as human beings are, and they interact with objects in similar or not so similar ways as human beings, but they are not subjects in the transcendental way, that is, they do not interact with objects qua objects of experience, as differentiatable from subjects.[17] Though I am not committed to a particular standpoint about which kinds of consciousness can be attributed to animals—and I think we need to differentiate here between types of animals (ants are sentient as much as dolphins are, but their levels of consciousness are uncontroversially very different)—I think that from a Kantian point of view animals cannot be said to have a type of consciousness other than intransitive. Animals have creature consciousness and some form of awareness of objects simply by being sensibly affected by them, but no transitive consciousness. They are not conscious of an object as differentiated, by themselves, from themselves. This does not of course imply that as creature-conscious animals they do not interact with objects in multiple complex ways. They are surely aware of the particulars, animate and inanimate, that they encounter and with which they interact in their surroundings by tracking these particulars in their different spatial locations (Golob calls these particulars ‘objects3‘ in Golob, forthcoming). So I concur with Golob that as conscious beings animals have a spatial phenomenology, but I diverge from his view in that I do not think that this entails animals have an awareness of the distinction between a perceived object and themselves (see further Section III).

I should point out that denying animals a higher-order form of consciousness might be seen as at least controversial because it has been demonstrated in research that certain higher mammals, such as elephants, dolphins and hominids, apparently show signs of some form of self-reflexivity—that is, they seem to have some ‘bodily self’-awareness when they see and react to a mirror image of a blob of paint that is painted on a particular side of their bodies in a way that suggests that they recognise it to be a blob of paint on their own bodies. But I think this ‘bodily self’-awareness still falls way short of the kind of self-consciousness that is conditional on the possibility of transcendental self-consciousness as mutually constitutive of explicit consciousness of self qua self and experience of objects qua objects.[18] The awareness of their own bodies might be likened to Sartre’s ‘non-thetic consciousness (of) the body’, which is of course in the first instance applicable to our own bodily self-awareness, but can be extrapolated to animals.[19]

Obviously, animals can make a distinction between, say, a stall and its door, as Kant himself points out (DfS, 2:59–60; see the discussion in KRS, p. 103, and Golob’s own excellent discussion in Golob, forthcoming, pp. 10–12), but that differentiation is not allied to an awareness of themselves; or, as Kant says, the ox can differentiate the stall from its door, but it does not recognise the differentiating as its own act: Kant makes a distinction between physically and logically differentiating. Now Golob would of course immediately rejoin that while an animal obviously does not make judgements about the objects it sees (by making logical distinctions), it does make physical distinctions, not just between objects, but also between the objects and it itself. But while we agree on the fact that animals are capable of making physical differentiations, on the point of what it means to make logical differentiations I think we disagree fundamentally, for in my interpretation apperception and the capacity for self-awareness in that strict reflexive sense cannot be decoupled from the capacity to judge, i.e. the capacity for making logical distinctions. So even if it is true that explicit judgements need not—and in the animal case, cannot—be made, the capacity to judge must still be there for beings to have the capacity for transcendental apperception as the necessary condition for strict awareness of self as well as objects as different from self. And, as we agree, animals do not have that capacity. Physically differentiating, by contrast, is just the empirical act of differentiating, by locating and tracking spatially arrayed objects, without having the capacity for being aware of, or recognising, that one is doing so—which, as Golob nicely points out, involves the ability to ‘recognise generic properties or marks’ (Golob, forthcoming). This last act Kant calls, in this early work, the act of logically differentiating, which later, in the Critical period, is more accurately called the recognitive synthesis of the imagination, which is the necessary ground of the capacity for conceptualising (and so for judging). While Golob appears to agree that this synthesis is not a capacity that animals share with humans, he does appear to believe, as said, that animals have some way of differentiating, not just between different objects such as a stall and its door, but also between objects and themselves. But I think that the capacity for physical differentiating is insufficient for this latter type of recognition of differences, and by implication the recognition of an object as an object (i.e. intentionality, or object*-directedness). And it is here that Golob and I disagree. Let me elaborate in the next section on the point of determinate space that Golob brings up, which is relevant to having the capacity to differentiate between objects.

III. Animal perception and spatial determination

When I wrote about the elephant calf taking a shortcut through the wooden fence of an elephant sanctuary (p. 308), this was meant to underline that it is not as if animals cannot perceive, in their own characteristic ways, complex arrangements in nature just because they do not share our transcendental, i.e. objectively valid, perspective in terms of being subjects of experience—animals are of course part of the category-governed mechanical nexus of nature. My point was intended to counteract arguments by strong conceptualists, who tend to argue that outside the transcendental perspective, so absent transcendental apperception, there is just ‘chaos’. But both strong conceptualists and nonconceptualists appear to mix up the empirical and transcendental perspectives. Obviously, purely seen from the empirical perspective, it is undeniable that animals as well as human beings interact in multiply complex and variant ways with objects. But, as I said above, it is also undeniable that animals do not have knowledge of objects qua objects and qua their interactions with them: they just interact with them in accordance with the laws that govern the phenomenal world.

The reason why the transcendental perspective is introduced by Kant is because in order to understand, from an a priori angle, how it is possible that we have knowledge of objects and why the judgements we make about them are objectively valid, it needs to be shown, in the Transcendental Deduction and then in the Analogies, that objects are functions of our very thought, and not things that are just out there for us to subsequently latch onto so that as a result of this we acquire knowledge of them as objects. For animals, by contrast, objects are not functions of their experience, let alone of their thought. Animals merely encounter objects with which they are programmed, as it were, to interact in specific ways. The entire Copernican perspective is not relevant for the case of animals. For us, of course, the way that animals do so interact is part of the natural realm that is constituted by the transcendental perspective because everything that happens in the phenomenal realm, in nature, is part of the nature that is determined by transcendental laws. But it would be a category mistake of the highest order to believe that since this is so, animals too have a subjective (let alone a transcendental) way of looking at things, or a close analogue of it. It would be anthropomorphic to project the conditions of our necessary way of experiencing (which, importantly, are also the conditions of the objects of our experience) and of cognising phenomenal nature on animals. The creature consciousness that animals have is nothing like our subjective, transitive kind of consciousness that we have when we experience or cognise our surrounding environment, which entails the possibility of making judgements about it and accounting for our beliefs (it should be noted that there is also a level of creature consciousness in ourselves, which is the kind of immediate awareness that we have in what Kant calls the ‘feeling’ of our bodily existence, i.e. proprioception—see the excellent analyses in this regard by Longuenesse 2017).

Golob misconstrues the context in which I wrote that it is ‘difficult to understand [note the emphasis!] though what it could mean for one to have intuitions that are not synthesised, that is, to have “merely [a] manifold” […] of representations’ (p. 326). The context here is the conclusion of the proof in the B-Deduction, namely that categories are necessary conditions of the perception of spatiotemporal objects, and that thus the objective validity of the categories has been proven. But this is clearly the conclusion of an argument that aims to explain the possibility of human cognition of objects, not the impossibility of animal perception. The combination or synthesis that is the necessary enabling condition of such cognition is necessary combination, not just any contingent binding of perceptions. I have analysed the difference between necessary combination and categorially arbitrary binding, and also why the argument in B130, which Golob cites, should not be read as if the two were conflated by Kant, in detail in my previous book (see Schulting 2012:141–7/2018:186–93). When Golob quotes me as denying, supposedly, that ‘animals “can have relation to an individual object, even if only indeterminately” (p. 21)’, and that ‘animals are unable to perceive “objects as determinate spaces” (p. 22)’, in both cases he interpolates the word ‘animals’, where in the context of those quotations, which is the introduction to the book, I do not at all talk about animals (animals first briefly appear in Chapter 3), but about possible human perception independently of the workings of synthesis. In fact, Golob’s charge is directly contradicted by another passage, where I do discuss animal perception. It is useful to quote this in full:

This does not imply that, in the absence of the categories, there cannot be a nonconceptual intuition of particulars, e.g. when an animal or an infant sees a spatially located thing, which it is able to discriminate from among other things, and which it can track (see e.g. Allais 2011:103). The animal or infant does not need to think the thing “as a persisting and causally unitary substance” in order to perceive a relatively distinct particular.This is just because animals or infants have no conception of objects in the strict Kantian sense of persisting and causally related substances. And mere perception is not dependent on the employment of the categories, or indeed the application of empirical concepts or ascription of properties. I concur with Allais on this (see Chap. 5). But it would be wrong to conclude from this—it being contrary to the specific goal of TD—that for adults with properly functioning cognitive abilities, who do employ concepts in judgements, categories would merely be conditions for thinking about objects, by virtue of “apply[ing] empirical concepts to objects in judgments” (Allais 2011:103). Rather, the goal of TD is precisely to explain how categories, not only enable the application of empirical concepts to given objects, but also first determine manifolds of representations in an intuition as objects, namely, as determinate appearances. The categories must thus be seen as also establishing the necessary conditions under which there can be said first to be objects, to which then empirical concepts can be applied. (pp. 170–1, underlining added)

Another passage he quotes from KRS on p. 266, where I talk about ‘a recognition of the manifold […] as qualitatively or quantitatively complex’, is linked by Golob to animal perception, but I do not discuss animal perception at all there. Again, the discussion there is to be seen in the context of the possibility of human perception of objects in the analysis of the Deduction (specifically, the A-Deduction), and what the counterfactual possibility of the categories not being instantiated in intuition would imply for human perception. Golob’s linking of that counterfactual possibility and the animal case might seem justified, but that assumes that our mere sensibility is equivalent to animal sensibility—this is an assumption underlying what James Conant (2016) has recently dismissed as ‘the layer-cake-conception’ of the relation between sensibility and the understanding; Golob indeed appears to endorse the layer-cake-conception, whereas I concur with Conant’s critique of this view. The difference is crucial: even if our sensibility were comparable to that of animals, for us it is difficult to imagine what it would mean to have mere sensibility, because doing so lands us in an inevitable circle: to be able to imagine to have mere sensibility we would need the very tools, namely, the synthesis of the imagination, that deny us that immediacy. By contrast, sensible immediacy is precisely the default disposition for animals. To try and imagine the animal’s situation is just that, an attempt to put oneself imaginatively in the position of a being that we cannot be (we can of course act in ways similar to animal behaviour in some of our subcognitive activities or sensible ‘coping’) (see further below).

Again, Golob himself appears to conflate two levels here: the transcendental or constitutive level and the empirical level. And I think this partly explains the problem he has with my reading, and also partly allays the worries he has. Though Golob makes it seem we do, in fact we do not disagree on the fundamental, empirically provable fact that animals and human beings alike—as holds for any object in nature—interact with objects in space in variant and multiple complex ways. There is nothing arbitrary (in an empirical sense) about the manner of these interactions: these are bound by the empirical laws of nature and diverse species-related biological embodied dispositions and behavioural patterns. Thus, if we take the excellent example that Golob takes, namely the ‘eagle who picks out the rapidly moving body of a mouse, who tracks that body as it passes bushes and grasses and strikes precisely it and not the similar coloured rock next to it’: the eagle is aiming at ‘a determinate space’ in the sense that the determinate space in question concerns another animal that is distinct from the object adjacent to it so that, wholly in virtue of contingent empirical laws of nature,  the predator can ‘aim’ accurately (or indeed less accurately) at its prey. This is entirely independent of the transcendental question about the possibility of determining a space from the vantage point of human reason for the purposes of analysing knowledge, the issue that Kant is interested in in the Deduction and the Analogies.

There is often the misconception that on Kant’s view, the world would de re not be the phenomenal, natural world that we know if it weren’t for the transcendental laws of human understanding. This misconception, again, conflates the transcendental and empirical levels of the analysis in the Critique. What is being analysed in Kant’s transcendental analysis is the possibility of empirical objects and events insofar as they are objects and events, not with respect to their factual existences and occurrences. The question is always and across the different sections of the Aesthetic and the Analytic in the Critique the question of how something is the case or exists, never that or if it is the case or exists, and then only from the specifically human standpoint of reason (to be sure, this is not a relativist premise on Kant’s part, but the premise of an analysis that purports to be rational and a priori, and not merely descriptive or empirical).

This means that criticisms, such as are currently en vogue in ‘speculative realist’ circles, that the Kantian transcendental perspective is a limited (and biased, reductionist) humanist perspective wholly miss the mark. If human beings were no longer around, or in ancestral times when there were no humans around, the eagle would still be able to precisely aim at the mouse, and not accidentally hit at the rock, because it is able, in virtue of empirical laws of association and location in space, to accurately track the body of the mouse that is in a distinct place in space; the associated connections directly map onto the way in which the prey is a spatially arrayed body in phenomenal nature relative to the predator’s own body. And to be sure, in those situations void of human beings there to observe events of interacting objects the transcendental perspective would still be universally valid, for that perspective is about the possibility of our cognition of objects and their conformity to our forms of cognition; spatiotemporal objects and events happen independently of our actual observations, but they do not and cannot happen independently of possible experience.[20]

This is why we argued in Onof & Schulting (2015) for the nonconceptualist position that spatial relations per se must not be seen as necessarily grounded on the a priori synthesis of the understanding. When Kant talks about ‘determinate’ space in the context of the Deduction, he means space as determined (bestimmt) by the human understanding (via the synthetic act of apperception, i.e. via imagination). Determinate space (bestimmter Raum or a ‘formal intuition’ [B160n]) strictly speaking, i.e. as such, exists only on condition of there being a determination by the understanding. There is no determinate space absent determination by the understanding: these are two sides of the same coin that cannot be decoupled on pain of misapprehending the implications of Kant’s idealism that is born of the Copernican turn. (And likewise, nature formaliter spectata exists only on condition of there being a determination by the understanding, but the empirical laws of nature, nature materialiter spectata, are of course as such independent of such an a priori determination, though they must conform to it.)[21] But determinate space as defined by the determinative act of the understanding is not metaphysical space, nor does it define the sui generis properties of space. Of course, the two are not two different kinds of space, as ontologically speaking there is just one space; but metaphysical space is as such independent of the way it is necessarily determined by the understanding if space is to be determined as a particular place adjacent to another. Determinate space is always finite with respect to the larger space of which it forms a part.

The ability of an animal to track its prey in space is itself therefore independent of the determination by an act of a priori synthesis, but the space in which it tracks its prey can be determined by proxy, i.e. by us, to be determinate space (on the animal’s behalf, as it were); in fact, it must be determinate space, for animals interact with phenomenal objects in the same (metaphysical) space in which we human beings do so.

But, by analogy with the thought experiment of which Kant speaks in his 26 May 1789 letter to Marcus Herz, if we ‘imagine [ourselves] to be an animal’, our (animal) representations of something in space would still ‘carry on their play in an orderly fashion, as representations connected according to empirical laws of association, and thus even have an influence on my feeling and desire [which Kant allowed for animals],[22] without my being aware of them’, i.e. not being aware of the representations’ ‘relation to the unity of representation of their object, by means of the synthetic unity of their apperception’ (Br., AA 11:52). Thus, animals have awareness of spatial objects, but not in virtue of an intentional awareness of the object of a unified representation that constitutes an object as ‘a determinate space’ (B138), but rather in virtue of their ability to associate merely according to empirical laws.

Where Golob and I diverge is with respect to the fact that I firmly believe that, on Kantian grounds at least, one cannot state that animals themselves, on account of their putative sui generis intentionality (object*-directedness), are responsible for the determination of an object in space or have such an intentional object in their mental states; that is, they are not reflexively directed at a determinate space such that they become aware of themselves as the agents of their perceiving a determinate space; they rather just interact with what are in effect, on account of a transcendental analysis, determinate objects in space (or, put differently, they interact with determinate spaces). Animals do not apperceive, synthesise, nor cognise, let alone apply categories, which means they do not have an intentional object. They simply perceive, with more or less accuracy, in virtue of binding certain representations that map one to one onto the objects they come to represent (these binding relations can be perfectly studied in zoology, or ornithology for that matter, to remain with the example of eagles).

The animal example I gave in the book was meant to illustrate the fact that, contrary to what Thomas Land (whose conceptualist views on spatial determination I discuss critically in that section) believes, there can be an intuition of something spatial without a priori synthesis being involved. I used the example of an elephant calf to illustrate this. Not all perception requires synthesis. But there the analogy stops with human perception; I certainly did not want to suggest that our sensibility is in all or even most respects like that of animals. And Golob’s critique of my position suggests that I accept the structural similarity between animal sensibility and our own, as I said earlier; but I deny this. So when I gave the example of hearing a sudden noise coming from a particular direction (p. 309), this was, again, of course an example of our human perception that is not (yet) a determinate perception of a distinct object—though I admit it might be a ground for confusion that it directly followed the paragraph in which I provided the animal example. But there is a similarity between that example and how an animal reacts in the same way but then does not, unlike human beings, apply categories (synthesise representations, apperceive) to become aware of a distinct object as distinct from themselves. But that does not mean that the animal only sees chaos, and cannot make distinctions on an empirical level, just as a human being would be able to do upon hearing a sound from the left without thereby directly invoking the categories.

That Golob thinks my example of hearing a sudden noise as indicative of the real possibility of an intuition of a spatially located but perceptually indistinct object independently of the activity of synthesis entails a ‘surely empirically false’ claim about animal perception indicates that he conflates the necessary combination that is required for human cognitive perception of objects (qua distinguishable objects) and the contingent empirical nature of the connections that hold between objects (including the specific ways in which an animal tracks an object) regardless of our determination of the necessary character of these connections which makes it possible for us to cognize objects in nature (this includes the determination of them qua objects as functions solely of our determination). The empirical world does not need the categories to exist or for some causal event to happen, nor are empirical laws dependent on the categories to obtain; they obtain regardless of whether we apply the categories or not. The determination of the necessity in them does of course require the categories: nature is nature qua nature only in virtue of the categories. But no animal or inanimate object for that matter needs the categories to interact in the specific contingent empirical ways that they do. The fine-grainedness argument that Golob brings to the discussion is irrelevant. That animals are capable of ‘tracking far finer-grained spatio-temporal relations than we are’ is besides the point of the discussion in the Deduction (and the Analogies) concerning the determination of spatiotemporal relations. The sharper eyesight of an eagle, compared to that of our eyesight, is not a ground for claiming that surely the eagle can aim at a determinate space as much as a human being, if not more so. For the determinacy in question in the Deduction is not empirical determinacy but necessary combination. But, again, the eagle does in fact aim at a determinate space, however not in virtue of its own intentionality, but because it is part of phenomenal nature that is constituted by transcendental as well as empirical natural laws.

Golob ignored my discussion, in Chapter 6 as well as Chapter 7, of the distinction that needs to be made between determining a complex manifold as complex and the complexity of the manifold as such. Golob’s critique of my position makes it appear as if the complexity is solely born of the determination in virtue of the synthesis of the imagination, and not already there in the manifold, as if indeed the spatially arrayed manifold that an animal perceives would not allow of very specific complex, fine-grained relations that the scientific evidence shows an animal such as an eagle can track. As I noted in my analysis of the synthesis of apprehension, a ‘mere manifold of isolated representations is not ipso facto qualitatively or quantitatively simple’, but ‘unsynthesised representations of outer sense are already quantitatively complex’, that these are ‘not recognised as complex’ (p. 287n.13), and that the complexity of a mere manifold does not depend on the synthesis of apprehension (p. 265). This is in line with for example Stefanie Grüne’s view that sensibility ‘delivers complex representations, whose content is however not represented as something complex’ (Grüne 2009:161).

IV. Concluding remark: Either objectivity or immediacy

I want to close with a remark on the ‘immediacy’ of intuition, which bears on animal perception. Absent transcendental apperception and the application of categories the object of intuition is still categorially indeterminate, or indistinct, but, importantly, given the absence of transcendental apperception there is no me strictly speaking, and hence there is no clear transcendental distinction between me and the object—in other words, there is no for me, so there cannot be an object for me. This follows from the analysis of Kant’s concept of self-consciousness and the related definition of an object in general in the first half of the B-Deduction (cf. the ‘transcendental object’ in the A-Deduction). The relation to the object of intuition I have in intuition is immediate, and therefore as such indistinct because it is by definition not mediated by a conceptual differentiation—that does not imply that the object is not something spatially arrayed outside me, with specific determinable boundaries, at a specific distance, with which other objects in the adjacent spaces interact in ways that are governed by universally valid natural laws. Animals are likewise orientated towards objects that are outside them in space, and in this sense I concur with Golob’s claim that ‘a three dimensional egocentrically orientated awareness of space within which something is seen as more or less distant is sufficient to sustain a distinction between spatio-temporal particulars and the subject’s own states, such as sensations’, so long as this is seen purely on the empirical level (though I would not use the term ‘egocentrically’). And this is an important proviso: we should not mix up the transcendental and empirical levels; only the former constitutes objective intentionality or objectivity in the strict sense.

What is important here is to realise that Kant, as well as his successors in German idealism who built on this aspect of the Critical philosophy, were keenly aware of the mutually conditioning relation between an object and the subject of experience in the philosophically precise sense of those terms. If there is no explicit subject, i.e. no category application and a priori synthesis, such as in the case of animals, there is no object strictly speaking either. This follows from the transcendental logic of what it means to have an intentional object: ‘object’ is a function of subjectivity, the intentional object is a transcendental object that is nothing but a synthetic unity of representations, an object in the mind (the transcendental subject) (see e.g. A104–9).

Does that mean that there is no object simpliciter, a thing or things de re? No, of course not—this is what I tried to convey by making the distinction Gegenstand/Objekt (p. 21), pointing to Kant’s own use of the generic, non-technical sense of Gegenstand that he employs at B1, where he states the empiricist principle that all cognition starts with experience. However, without category application/a priori synthesis/transcendental apperception there is no object for a subject just because there is no subject for which there would be an object. There is of course a certain orientation in space, necessarily for any rational or non-rational animal that is part of nature. However, while the orientation in space for humans is necessarily guided by the understanding—by way of the figurative synthesis—in order to determine the objects towards which one is orientated for the purpose of cognitive experience or judgement, for animals the orientation is sufficiently guaranteed by the way they are wired to interact with spatially arrayed objects; their orientation does not depend on a subjective act of synthesis or the binding of representations in such a way that the objects are determinate for them, as it were. The objective space within which they track other animals and objects is just the space of which they are part as much as objects are. To say that there is something for an object to be for an animal other than that an object sensibly affects that animal, and prompts sensations in the animal and consequently certain behavioural patterns, is an anthropomorphic projection. Rather, the animal’s ‘experience’, if it can be called that, is one of pure immediacy.

Pure immediacy must be understood in the way that Hegel understood it: the identity that is constitutive of both subject and object, and thus at the same time enables their differentiation, is ‘in intuition […] totally immersed in the manifold [im Anschauen ganz und gar in die Mannigfaltigkeit versenkt]’ (Hegel 1977:70, GW 4:327). Hegel, just as Kant, has of course human sensible intuition in mind here, but this can be extrapolated to the animal case. Pure immediacy means that having a mere intuition of an object does not differentiate the object from its being represented; the representer’s identity and the relative difference of the represented object are ‘immersed’ in the manifold of representations; but that does not mean that the difference between the representation or sensation (mental states) and an object that is being represented is not there, it is rather ‘immersed’ in the manifold representations. The metaphor that Hegel uses here is to make it clear that only the understanding is able to lift the relative difference (antithesis), and thus the relative identity, of subject and object out of that ‘immersion’. Hegel refers to the blindness of intuition of which Kant speaks at A51/B75. That an intuition without concept is blind does not mean (as might be suggested by the term ‘blind’) that an intuition lacks awareness, but it means that without the involvement of the capacity for conceptual differentiation (and synthesis), i.e. apperception, an intuition cannot make out the distinction between representation and represented, and is thus blind with respect to the ‘antithesis’, as Hegel calls it. Hegel writes:

Kant is therefore quite right in calling intuition without form [i.e., concept] [Form (des Denkens)] blind. For in [mere] intuition [without form] there is no relative antithesis [Gegensatz], and hence there is no relative identity of unity and difference. This relative identity and antithesis is what seeing or being conscious [das Sehen oder das Bewußtsein] consists in; but the identity is completely identical with the difference just as it is in the magnet. (Hegel 1977:70, GW 4:327)

I am not interested here in the validity of Hegel’s interpretation of Kant, but one thing is right about this reading, namely that in intuition the ‘relative identity’ that exists between the unity of the understanding, with its characteristic forms (categories), and the difference (plurality) of the manifold of representations—an identity that is semantically expressed in a judgement (the very structure of a judgement S is P expresses the relative identity and antithesis of intuition and concept, “something S is P”)—has not yet come to the fore: this identity is the unity between, on the one hand, an experiencing subject that applies concepts and, on the other, the object to which, by way of the synthesis of an intuition of that object, those concepts are applied. Rather, the antithesis in empirical intuition ‘besteht … in dieser Form des Versenktseins’.

This, then, is what nonconceptual content in a Kantian sense would mean: a content that has not yet been given the conceptual form which makes it a determinate content, whereby the opposition (and relative identity) between the subject and object of experience is first made manifest, distinct. Now of course we can redefine the terms of what we mean by objectivity or, more interestingly, differentiate various meanings of objectivity—which is what Golob aims for in his Section 3, and more in detail in his article—but then we are no longer in the territory of interpreting Kant. The Copernican object is always an object for us, a transcendental object, and not something that exists or is already given apart from us. On Kant’s view, animals do not have the Copernican perspective, so their perceptions involve neither a transcendental object nor some kind of intentional object-directedness. Animals are just part of the phenomenal realm as objects among other objects, with which they interact ‘immediately’, not as if they were intentionally aiming at them.[23]

Acknowledgements: I thank my co-editors Christian Onof and Jacco Verburgt for their helpful comments on an earlier draft of this essay.


[1] It is difficult to say what Kant thinks of animal consciousness, as he does not say that much about it; at first blush, he seems to deny animals consciousness (though not representations) and certainly he denied them self-consciousness (see Naragon 1990:6–7). But an informed view can be extrapolated from some of his enunciations in this respect; see further below in Section II and Schulting (2015a). Cf. McLear (2011).

[2] In Schulting (2015b), I’m at any rate very clear about my dismissal of both nonconceptualism and conceptualism, and about the fact that the whole debate is not helpful to understand the thrust of Kant’s Deduction.

[3] An exception here is Robert Hanna, the originator of the debate, who argued that Kant is both a nonconceptualist and a conceptualist (Hanna 2005, 2008), and that this contradiction is exhibited by Kant himself. Even though I disagree with Hanna, I find his position more consistent than all those that try and argue that Kant is either a nonconceptualist or a conceptualist, whether or not with specific emendations. Lucy Allais (2009) is perhaps an exception in this, for she explicitly espouses a relative nonconceptualism.

[4] To be precise and fair, it was actually Robert Hanna who initiated the debate with his pivotal article in the European Journal of Philosophy in 2005 (Hanna 2005). But it was only in the wake of Allais’s article, published in 2009, that quickly many more followed. For bibliographic references see the introduction to Schulting (2016).

[5] As I wrote on p. 220 of the book, Allais makes an important distinction between ‘being presented’ with objects and the presentation of an object as object, which ‘the non-concept-having creature is arguably not in a position’ to do (Allais 2009:405; cf. 2009:407, 413). This position is defensible as an interpretation of the textual evidence there is with respect to the relation between given objects and intuition, and in principle compatible with my moderate conceptualist interpretation, but, I think, not with Golob’s nonconceptualist reading (at least as presented in his critique).

[6] Hanna thinks Kant is both a nonconceptualist and conceptualist, but, consistently, he acknowledges the contradiction, which he believes there to be in Kant.

[7] To say that intuitions are conceptually structured really means that they are categorially structured. This is, I think, important to notice, since one could argue that though intuitions are categorially structured, empirical concepts need not be applied—this is e.g. the position of Griffith (2012) (see my critique in Schulting 2015b:566n.13).

[8] I do not endorse the view that mere intuitions are different from ‘a bunch of sensations’. Often it is suggested that sensations do not have pure forms of intuition, but sensations must at least have the form of time, so sensations are the empirical content of an intuition which has time as pure form.

[9] This is in contrast to, say, Robert Hanna and the early Allais. See my discussion of their accounts in KRS, ch. 5.

[10] As Golob says, ‘it is standardly understood as the thesis that a subject may possess empirical intuitions of spatiotemporal particulars even if that subject entirely lacks conceptual capacities’; I agree with this, but I do not take this to imply that intuitions refer objectively, in the strict sense. See further below.

[11] I am aware of course of the irony that I edited a  book on Kantian Nonconceptualism (Schulting 2016).

[12] I do not think Thomas Nagel’s view of something ‘what it is like to be’ can be transferred to the Kantian context. I also think it is important to make a distinction between viewpoint in terms of ‘what it is for x to be for y’ and in terms of ‘there being something that it is like to be’. Nagel seems to conflate these two. Nagel believes that ‘fundamentally an organism has conscious mental states if and only if there is something that it is like to be that organism—something it is like for the organism. We may call this the subjective character of experience’ (1991:166). Obviously, one of Nagel’s points is to suggest the fundamental alienness of non-human forms of experience. But the way he defines consciousness conflates intransitive and transitive consciousness. For an organism to have some form of consciousness does not entail that there is something it is like to be that organism for itself, that it has subjective experience and thus consciousness of an inner life. A non-human animal (depending on the species) can be attributed creature consciousness (some lesser or greater degree of sentience, responsiveness to stimuli), but not a higher-order form of consciousness, only the latter of which indicates a consciousness of something for one, and not just being sentient or responsive, in whichever intricate ways, to sense stimuli.

[13] Sometimes Kant also defines consciousness in this way, e.g. in the Logik, namely ‘Bewußtsein [ist] eine Vorstellung, daß eine andre Vorstellung in mir ist’ (Log, AA 9:33; cf. V-Met-L1/Pölitz, AA 28:227), but here, as so often in Kant, transitive consciousness is meant.

[14] Cf. Naragon (1990:13).

[15] Kant does not employ the term intentio, but the Latin propositum is a close analogue, since it denotes ‘intention’, ‘plan’, ‘aim’.

[16] This is the root sense of ‘object’, which is the noun form of the verb obiicio, which denotes ‘put in front of’, ‘put before’. Object as antonym of subject is of course of relatively late, medieval origin.

[17] This is the crucial aspect of Kant that Strawson got right with his interpretation of Kant’s argument as a transcendental argument about the possibility of experience, namely, that there exists a reflexive relation between the subject and the object of experience. Golob also notes this, but he does not mention the element of reflexivity that is involved and which Strawson points out too. Experience itself expresses this reflexive relation between a subject of experience and the object that it experiences. It is hard to see how an animal has this reflexive awareness, whereby it self-ascribes the representations that it has of objects to itself.

[18] Also, animals’ awareness is not mutually recognitional: they do not recognise each other as having the type of awareness that would characterise them as self-conscious agents.

[19] Cf. Longuenesse (2017), who relates Sartre’s concept of ‘non-thetic (self-)consciousness’ to Kant’s implicit notion of bodily awareness in the ‘feeling’ of existence that is implied by Kant’s ‘I think’-proposition. She also rightly differentiates this consciousness from the consciousness of determinate objects, as well as from the consciousness of self as the agent of thought.

[20] Suppose that we were transported back to an ancestral time t, in which dinosaurs roamed the world. The transcendental perspective would ensure that the empirical objects we cognised conformed to our forms of intuition and the understanding so as to enable a recognition of those natural objects and events as such in conformity with both the empirical and transcendental laws of nature. Objects are subject to those forms no matter whether humans are around or not to actually observe them.

[21] See KRS, ch. 7.

[22] See V-Met/Herder, AA, 28:117.

[23] My defence of Kant’s position on animal perception may be seen as futile when one has seen this video, and wonders if animals do not have a rather sophisticated form of intentional consciousness after all, just as Golob argues.


Allais, L. (2009), ‘Kant, Non-Conceptual Content, and the Representation of Space’, Journal of the History of Philosophy 47(3): 383–413.

Conant, J. (2016), ‘Why Kant is not a Kantian’, Philosophical Topics 44(1): 75–125.

Golob, S. (forthcoming), ‘What Do Animals See? Intentionality, Objects and Kantian Nonconceptualism’, in L. Allais & J. Callanan (eds), Kant and Animals (Oxford: Oxford University Press).

Griffith, A. (2012), ‘Perception and the Categories: A Conceptualist Reading of Kant’s Critique of Pure Reason‘, European Journal of Philosophy 20(2): 193–222.

Grüne, S. (2009), Blinde Anschauung. Die Rolle  von Begriffen in Kants Theorie sinnlicher Synthesis (Frankfurt a.M: Klostermann).

Hanna, R. (2005), ‘Kant and Nonconceptual Content’, European Journal of Philosophy 13(2): 247–90.

——— (2008), ‘Kantian Non-Conceptualism’, Philosophical Studies 137: 41–64.

Jankowiak, T. (2018), Review of Dennis Schulting: Kant’s Radical Subjectivism: Perspectives on the Transcendental Deduction, Revista de Estudios Kantianos 3(1): 119–22.

Longuenesse, B. (2017), I, Me, Mine. Back to Kant and Back Again (Oxford: Oxford University Press).

McLear, C. (2011), ‘Kant on Animal Consciousness’, Philosophers’ Imprint 11.

Naragon, S. (1990), ‘Kant on Descartes and the Brutes’, Kant-Studien 81: 1–23.

Nagel, T. (1991), ‘What Is It Like to Be A Bat?’, in T. Nagel, Mortal Questions (Cambridge: Cambridge University Press), pp. 165–80.

Onof, C. & D. Schulting (2015), ‘Space as Form of Intuition and as Formal Intuition. On the Note to B160 in Kant’s Critique of Pure Reason‘, the Philosophical Review 124 (1): 1–58.

Schulting, D. (2012), Kant’s Deduction and Apperception. Explaining the Categories (Basingstoke: Palgrave Macmillan).

——— (2015a), ‘Transcendental Apperception and Consciousness in Kant’s Lectures on Metaphysics’, in R. Clewis (ed.), Reading Kant’s Lectures (Berlin and New York: de Gruyter), pp. 89–113.

——— (2015b), ‘Probleme des „kantianischen“ Nonkonzeptualismus im Hinblick auf die B-Deduktion’, Kant-Studien 106(4): 561–80.

——— (2016), Kantian Nonconceptualism (London and New York: Palgrave Macmillan).

——— (2017), Kant’s Radical Subjectivism. Perspectives on Transcendental Deduction (London and New York: Palgrave Macmillan).

——— (2018), Kant’s Deduction From Apperception. An Essay on the Transcendental Deduction of the Categories (Berlin and Boston: de Gruyter).

Thiel, U. (2011), The Early Modern Subject. Self-Consciousness and Personal Identity from Descartes to Hume (Oxford: Oxford University Press).

© Dennis Schulting, 2018.

Dennis Schulting is a former Assistant Professor of Philosophy at the University of Amsterdam. His work concentrates on Kant and German idealism. His most recent journal publications were published in Kantian Review, Studi Kantiani KANT YEARBOOK, KANT-STUDIEN, and the PHILOSOPHICAL REVIEW. He is the editor of Kantian Nonconceptualism (Palgrave Macmillan 2016). Two new books are forthcoming: Kant’s Deduction from Apperception. An Essay on the Transcendental Deduction of the Categories (Walter de Gruyter, 2018) and Reflexivity and Representation. Essays on Kant and German Idealism (Walter de Gruyter, 2019). A book on Hegel’s “Science of Logic” is in the works.