1. Reply to Brook

I am pleased to hear that Andrew Brook agrees with (most of) the main claims in my book. Before briefly commenting on his prima facie challenging notes on the so-called ‘localization problem’, the place of the Refutation of Idealism and the issue of matter, let me first quickly address just one other point that Brook raises. I agree that Kant’s radical subjectivism already flows from his claims regarding space and time as the ‘forms of organization that the mind imposes on sensory input’, as Brook puts it. As he notes, space (and to a much lesser extent time) does not make a real appearance until Part II of the book. Though the forms of intuition are certainly subjective, they are however not more radically subjective, as Brooks suggests, than the principle of apperception that lies at the root of my thesis of radical subjectivism. With this book, I wanted to look at Kant’s radical subjectivism from the perspective of the Deduction, whereby the B-Deduction’s ‘second step’ in particular fleshes out the claim, made in the Aesthetic, that space is ‘the condition of the possibility of appearances … and is an a priori representation that necessarily grounds outer appearances’ (A24/B39; emphasis added). The radical subjectivism I focused on concerns the idea that the principle of apperception, by determining space, is ‘solely constitutive of the possibility of perceiving objects as determinate spaces’ (2017a: 22). It is then transcendental apperception by means of a determination of pure intuition that first grounds appearances as objectively real outer appearances.

As to the changes between the A- and B-edition of the Deduction, I took them not to be so radical as to affect the main arguments, so I do not believe there is a real chasm between the two editions. Brook claims that the second edition Refutation constitutes even more radical a break with the first edition (i.e. the Fourth Paralogism), which would also be a problem for my thesis of radical subjectivism. I take here the position of P. F. Strawson (Reference Strawson1968: 246, 258), namely that differences between them are not due to the fact that Kant changed position, but because he thought the first formulation was misunderstood. But Brook broaches here what might strike one as a central problem for Kant: how can we take empirical intuitions to provide the real possibility that cannot be ‘conjured out of’ concepts (cf. Schulting Reference Schulting2017a: 362) if even spatial and temporal structure is ‘actively’ contributed by the mind and not already in some way in the manifold, and indeed if the intensive magnitude of an intuition, which corresponds to matter, is first determined by the mind?

Whatever the case may be, I do not think that Kant changed his mind in the Refutation, nor do I think that the ‘localization problem’, as Brook contends and which he links to the issue about the Refutation, is a real problem for Kant’s transcendental philosophy, because the question of which particular location is to be determined, and why this one (and not another), seems an outright empirical question, which already assumes that a priori spatiotemporal structures and the a priori ability to synthesize manifolds must be in place before we can locate elements of the manifold empirically. This is not to say empirical ‘knowledge acquisition’ is not an interesting or important problem, but it just seems to me to be irrelevant to Kant’s transcendental project.

With regard to the Refutation, obviously I cannot go into any detail here, but there are at least two reasons for seeing the Refutation as consistent, rather than as a break, with the claims made in the Deduction (in both editions). The first thing to note is that the Refutation is not a proof of transcendental idealism, rather it is a refutation of an idealism that takes objects in space as (possibly) merely imaginary. Kant thereby takes the view, presented in the Aesthetic, that space and time are not properties of things in themselves as granted. Although arguably, by here granting this view he does in some way presuppose its truth in the background of the refutation, Kant cannot and does not need to refute idealism from within the transcendental-idealist framework. So when he talks about Dinge outside me, this concerns objects outside me in space, not things in themselves outside possible experience. It is striking that the very terms ‘appearance’ and, apart from its use in the introductory section to characterize Berkeley’s idealism, ‘things in themselves’ do not appear in the Refutation itself; instead he uses the neutral Ding/Dinge or Gegenstand/Gegenstände.

Secondly, Kant never espoused the idea, as Brook seems to suggest, that matter is a feature of our mental selves (which is at any rate the sense of self meant in the Refutation); rather, it is always a feature of possible experience (substantia phaenomenon; A277/B333), and is that which fills space, which together with all the appearances in it (A374–5) is only transcendentally in us. I do not think that in the Refutation Kant abandons this view of matter in favour of a supposedly more realist view, but rather confirms it in that the matter of the objects (Dinge) outside us (empirically) and we ourselves as determinate existences in time (B277) are regarded as on an equal footing. There is no sense in which ‘now [i.e. in the Refutation] matter is not part of the mind and actually has temporal, spatial, and conceptual order’, whereas in Kant’s account prior to the Refutation matter was part of the mind.

The realism defended in the Refutation is thus nothing ‘new’, but was there all along. Precisely because matter (and not just form) is part of possible experience, we are able to a priori determine matter as that which fills space, in virtue of the categories of quality (cf. Schulting Reference Schulting2017a: 335, n. 28). I do not underestimate the complexity of Kant’s views on matter – not least of the question of what metaphysically grounds it – which he continued to reflect on right until the end of his career (cf. the Opus Postumum), but the issues that Brook brings up do not seem to me to pose a real problem to the general framework of transcendental philosophy as such.Footnote ²

2. Reply to Gomes

In his excellent account of my argument in chapter 4 of the book, Anil Gomes counters my criticisms of his own position and raises further key issues relating to the B-Deduction. He also says that there may be more agreement between us than I made it appear. I think he may be right, although I also believe there is still some important variation in our approaches to the Deduction. Unfortunately, I cannot treat here all the themes that Gomes broaches (such as his important point about categorial illusion; but see below in reply to Newton).Footnote ³

My thesis of Kant’s radical subjectivism implies that the subjective conditions of thought are not only necessary but also formally sufficient conditions of the objects of thought (see e.g. Schulting Reference Schulting2017a: 22). Basically, I take Kant to argue in the Deduction that there is no gap between the transcendental unity of apperception which governs self-consciousness and the transcendental conditions for the objects of thought. Or as Robert Pippin formulates it: ‘The object just is “that in the concept of which the manifold is united”; representation of an object just is rule-governed unity of consciousness’ (Pippin Reference Pippin2015: 71).

As Gomes nicely summarizes, in chapter 4 I argued, in line with the main thesis of radical subjectivism, that the gap that might be thought to exist between the subject’s categorial representation of the object in virtue of the unity of consciousness and the object so represented is not really there. Or, in Gomes’s words, there is no gap between ‘(S): Necessarily, subjects ϕ objects in accordance with the categories’ and ‘(O): Necessarily, the objects of subjects’ ϕing exemplify the categories’. So prima facie Gomes is right to say that from the four gap-denying options that he names I hold option (3), namely, ‘There is no gap to be bridged because in establishing (S) Kant ipso facto establishes (O)’.

Now Gomes held the view, the one I critiqued, that the gap between (S) and (O) is first bridged in the ‘second step’ of the B-Deduction. I criticized this view, for, most importantly, if there were a fundamental gap in the ‘first step’ it could not be remedied by the ‘second step’ because the latter structurally depends on the success of the former. I still believe this. However, I now think Gomes and I are, in one respect at least, closer than I made it out to be.

First, the ‘second step’ is of course supposed to show something about the relation between (S) and (O) that was not already shown in the ‘first step’. So the argument that establishing (S) ipso facto establishes (O), the argument which I indeed take Kant to conclude in the ‘first step’, cannot mean that the ‘second step’ is redundant or at best draws trivial conclusions from the ‘first step’ argument. It would appear that some gap-bridging needs to be done. For one might object that my denial of a gap between (S) and (O) applies only to the relation between subject and an object in general, namely between the unity of consciousness and the synthesized manifold of representations that constitute the concept of an object, but still leaves wide open the gap between subject and a spatiotemporal object (cf. 2017a: 77ff.). The ‘second step’ is then a necessary step to close this last gap. But here the radicality of my thesis of Kant’s radical subjectivism shows: for here too I claim it is the same category-applying subject of the understanding (S) that first establishes the object as ‘a determinate space’ (B138), if by virtue of its effect as a figurative synthesis on the sensible manifold (B152; see chapter 7).

Secondly, the ‘first step’ argument is an argument from the unity of apperception to the concept of an object (which establishes S) and then to (O), and in this sense I concur with Gomes saying that ‘in this minimal sense … the existence of a gap … seems to me non-negotiable’. The intimacy I claim exists between subject and object must indeed first be shown by way of the progressively structured proof that starts with the familiar ‘I think’ proposition in §16 subsequently moving to the definition of an object in §17. As I argued and as Gomes rightly notes, this move is a ‘non-ampliative’ one and can be carried out through a proper understanding of the constituents of apperception, something I do in sections 4.8 and 4.9, but which I have done in more detail in my previous book (2012).Footnote ⁴

To be sure, that analysis is not dependent on the appeal to idealism – more precisely a non-reductive form of phenomenalism – that I made in section 4.10 (2017a: 174ff.). Rather, the idealism is implied by that analysis and it helps us understand the way in which the identity relation between (S) and (O) should be interpreted: that is, the identity is not such that the object qua a thing in itself be seen as existentially dependent on (S).

Let me briefly turn to Gomes’s interesting exposition of the ‘second step’ as a way of responding to my two other criticisms of his earlier proposal for closing the gap. Gomes suggests that the synthesis responsible for the unity of space is ‘undertaken by the understanding, but without any use of concepts’, i.e. involving apperception but not the categories or ‘any concept-governed synthesis’. In this way, if I read this correctly, the ‘second step’ is to be taken to argue that objects in space and time are unified/synthesized in virtue of the unity of apperception, and as such amenable to being unified/synthesized under the categories. The gap between (S) and (O) is bridged because the synthesis that establishes the unity of spatiotemporal objects ‘has its origins in the same source as the categorial synthesis’. As Gomes rightly says, this removes the ground for my criticism that his earlier proposal risked claiming that the objects themselves necessitated the instantiation of the categories, for on his currently proposed interpretation, it is the unity of apperception that carries out the combination necessary for categorial determination.

However, I think – but can only hint here – that among other problems of a more interpretative nature this solution faces the following problem: it creates a dualism within the apperceiving subject, namely between (S) and (s), where small ‘s’ denotes the subject that ‘does not establish the exemplification in virtue of a discursive act of categorial synthesis’, and as a result there arises a regress problem, given that the synthesis that is at issue in the Deduction is an original synthesis, more original than which there is none (B132). What is the regress? Assume that the non-category-governed synthesis by (s) that establishes the unity of space is a priori and original, and that the categorial synthesis carried out by (S) is also a priori and original. Then which a priori and original synthesis a priori synthesizes (s) and (S) such that we have a discursive synthesis, i.e. a judgement about spatiotemporally unified objects? Is it an a priori synthesis*? But this leads to a regress, which a priori synthesis was in fact designed to block (see my discussion in 2017a: 222–3, 314; 2017b: section 4.4). As I think Gomes would agree, there is just one a priori original act of synthesis, so we are owed an explanation – which avoids this regress problem – of the way in which (S) and (s) are one and the same function of the understanding that, in one act, unites the spatiotemporal manifold and applies the categories in a judgement.Footnote ⁵

3. Reply to Howell

Robert Howell comments on one claim in my earlier book (Schulting Reference Schulting2012) about the standard view that the argument in the B-Deduction is either regressive or progressive. As I explained over several pages (2012: 61–75), my view is that the B-Deduction is both progressive and regressive. But Howell thinks I ‘misread the Deduction’s structure’ in claiming, supposedly, that the Deduction’s ‘main argument is regressive’, and he says that apparently on my view the Deduction ‘does not contain Kant’s main argument for the objective validity of the categories at all’. I am frankly rather surprised by this last remark: I make it plainly clear, throughout the book, that the Deduction’s main argument is about showing the objective validity of the categories. How could it not be? I also explain repeatedly that the B-Deduction argument, while it is globally regressive, must be and is a non-regressive ostensive (e.g. 2012: 22, 26–7, 29–30), internalist (72), Cartesian-like (71–2, 74), ‘dogmatic’ genetic proof of sorts (23–7), in other words: progressive. I even call it the ‘P-argument’, whereby P stands for progressive.

To the extent that I am able to follow Howell’s reconstruction of my views (he says contradicting things), I do not recognize in it my own understanding of the B-Deduction as a complex integrated multi-level argument (2012: 14), namely, that it is explanatory, validatory and directly demonstrative (cf. 2012: 225, n. 1; see also 2017: 296ff.). I quoted Vaihinger as a precursor of this integrative view (2012: 64, 242, n.71). The fact that one can find both regressively and progressively structured parts in the text of the Deduction, as Howell indicates, is irrelevant to the point of my claim: the argument of the B-Deduction as a whole operates on two levels – regressively construed, as giving a global explanation of the possibility of knowledge in line with the Prolegomena, given the factum of knowledge; and progressively construed, what we see in the text of the ‘main argument’ of the B-Deduction as the actual proof for showing the possibility of knowledge from apperception (from §16).Footnote ⁶ I made it quite clear that Kant must ‘provide a “dogmatic” proof of sorts of the epistemological claims that he is making in TD, something that a mere regressive argument cannot accomplish’ (2017: 57–8; emphasis added), and that that proof is given from §16 onwards, which is delineated in great detail in all of chapters 6 to 9 of my first book.

My progressive construal is, to be sure, unlike mainstream progressive readings, such as Howell’s (see Schulting Reference Schulting2012: 73ff.). Mainstream progressive readings are incompatible with regressive readings such as Ameriks’s. Mine is compatible with such regressive readings, and I consider this an advantage. It is not compatible though with seeing Kant as primarily preoccupied with anti-scepticism, as I acknowledged, but those (like Howell) who insist that Kant is preoccupied with anti-Humean reasoning might want to reconsider their assumptions. At any rate, Howell’s anti-sceptical construal of the progressive argument begs the question against my construal of that argument, which contrary to Howell allows ‘merely associatively-organized sequences’ that are not, even in principle, category-subsumed.

Turning to the book that is under discussion here, Howell criticizes me for failing to provide reasons in the Deduction that are ‘independent of the Transcendental Aesthetic for accepting transcendental idealism’ (emphasis added), and he sees this as leaving a gap in my argument for Kant’s radical subjectivism. I find this criticism puzzling. If we take the argument for idealism in the Aesthetic to concern the argument that space and time are not properties of things in themselves, but are mere representations (however interpreted), and that therefore neither are objects represented in space things in themselves, then my reading of the Deduction’s claim concerning the concept of an object as involving idealism does not depend on that argument, since my reading of that claim simply does not involve space and time. Space and time, and the vexed relation to the Aesthetic, do not make an appearance until the ‘second step’ of the B-Deduction (which I discuss in chapter 7). Contrary to what Howell claims, then, I do provide an Aesthetic-independent argument for idealism from the concept of an object in general, which says, roughly, that the concept of an object in general including its possible instantiation in empirical appearances is wholly constituted by and thus dependent on the subject’s determining agency and consists merely in the a priori synthesis of representations in virtue of that agency, and therefore does not exist outside the subject’s synthesized representations, hence it does not exist in itself, or, as thing in itself.Footnote ⁷ This is my (here much shortened) argument for idealism from the concept of an object in general without in the least appealing to the Aesthetic argument for idealism from space and time. One can of course disagree with this interpretation, but one cannot dismiss it as a failure on the ground that it is not independent of the argument from space and time.

Howell presents a further criticism that he regards as damning for my argument for idealism from discursivity, but which I think is based on an interpretative mistake on his part. So let me dispatch this one before proceeding. Howell says that we can make true judgements about things in themselves, e.g. we can judge that ‘every thing in itself is non-spatiotemporal’. I agree insofar as this is a true analytic statement that follows directly from Kant’s thesis of idealism. But, as I argued similarly in a critique of Van Cleve (Schulting 2017a: 408–9), the predicate of non-spatiotemporality is affirmed here merely of the concept <thing in itself>, so of things in themselves in general and, just as Howell agrees, not of any particular thing x, underlying the subject concept, which is picked out ‘as the particular, fully determined, infinitely propertied’ thing that it is in itself (de re). In other words, I have a mere intellectual grasp of a particular thing in itself qua a thoroughly determined individual as little as I would have an objectively valid cognition in virtue of an empirical intuition of such an individual as appearance; I only know, in virtue of Kant’s thesis of idealism, how to use the concept <thing in itself> and what it means to affirm concepts such as <nonspatiotemporal> of it. And this was my point: already discursivity, even apart from the constraints of our sensibility, implies idealism about things in themselves. We only have a concept of a thing in itself and its necessary properties (or necessary lack of certain properties), but we do not thereby properly think the thing in itself as a thoroughly determined given individual in any theoretically determinate sense at all (cf. Prol., 4: 359, quoted in 2017a: 422, n. 45). Howell appears to think we do have such an intelligible grasp, in virtue of de dicto true analytic statements, of the thing in itself de re, but Kant abandoned this view after the Inaugural Dissertation.

4. Reply to Newton

I am grateful to Alexandra Newton for recognizing the fact that with the thesis of radical subjectivism I rehabilitate, in some way at least, the centrality of the Copernican turn, which, as Newton rightly says, has been downplayed in recent more metaphysical readings of Kant. However, as becomes clear in her criticisms, my Kant is probably much more of a metaphysical realist, and also more ‘modest’, more subjectivist, than she deems warranted. Newton raises four worries; I shall only be able to consider three of them. Her last point concerning non-conceptualism – interesting though it is – must await another opportunity.Footnote ⁸

Regarding her first point, concerning spontaneity: I am unsure why she takes me to make a distinction between ‘the principle of the understanding’s activity’ and ‘its exercises’. Following Pippin, I took spontaneity to be an irreducible principle insofar as it cannot be reduced to the causal order of the content, or to the external ground (phenomenal or noumenal), of my ϕ-ing; for every ϕ-ing that I do, even if an external ground of my ϕ-ing can or could be determined, I take myself as the one that spontaneously, wholly out of my own accord as the subject of so ϕ-ing, does the ϕ-ing. There is an identity of sorts between the principle of spontaneity and its exercise (see also Schulting Reference Schulting2017c). If one wants to call this irreducible principle of apperceiving, judging or understanding absolute spontaneity, I am fine with that as such, since the principle is absolute in the sense of it being irreducible to ‘an external source’ and because there is no other principle governing my ϕ-ing (qua ϕ-ing). However, I prefer not to call it ‘absolute’ inasmuch as Kant reserves absolute spontaneity for the idea of transcendental freedom, whereas the spontaneity of the understanding, as Kant himself says, ‘cannot produce from its activity … other concepts than those which serve merely to bring sensible representations under rules’ (Groundwork, 4: 452) (see the discussion at Schulting Reference Schulting2017a: 128ff.). Of course, the non-absolute spontaneity at issue here is not a Sellarsian relative spontaneity that is merely a posteriori responsive to causal impingements but it is one that is ‘concurrent’ with sensibility, as I put it, so relative to the necessary sensible input that it unifies. Incidentally, the Pölitz note Newton quotes is pre-Critical, as I note at 2017a: 137, n. 59, and thus cannot without qualification be used as support for an account of Kant’s mature theory of spontaneity.

As to her second worry, I grant that on my reading, as with the ‘I’ of apperception, ‘the object of judgement must likewise be indexed to me, as the object of my experience’, and is thus indeed an object for me (B138). But this is not problematic, since Kant’s ‘my’ here (B138) refers merely to the necessary subject of judgement, not to any particular subject. So if I judge ‘This chair is a Gispen’, say, the relevant point about apperception being involved is that I implicitly know that I as judger so judge, not that Dennis Schulting so judges (cf. Schulting Reference Schulting2017c). There is nothing subjectivistic about this, as Newton seems to suggest. On my account, it does involve phenomenalism, as I explained in the book, but this is not a phenomenalism of the ‘pernicious’ kind, which holds that all experiences are just the experiences of particular individuals locked up in their own solipsistic universes. The judgement ‘This chair is a Gispen’ is objectively valid, because it is about the object seen in front of me, which, provided the empirical evidence backs up my claim, is indeed a Gispen, objectively, not because I just happen to believe that. One should not mistake Kant’s talk about ‘my representations’ and ‘an object for me’ for any relativistic talk about the objectivity of the objects of judgement. Precisely because in a judgement I take my empirical representations to be about an object vouches for the latter’s identity as a veridical object, not as being just my perspective on it, for the transcendental rules of the necessary unity of one’s own consciousness are the very categorial rules that establish the objective validity or reality of my representations, as thereby first constituting the object of my judgement – given sensory input of course. Transcendental apperception thus is identical to, because constitutive of, the objectivity of the object of my judgement. There is no separation, in the strict transcendental sense of these terms, between the subject of thought and its object(s). This is why Kant identifies a judgement as the objective unity of apperception. Newton’s own take on the relation between the ‘I think’ and judgement cannot account, it seems to me, for why it is that ‘transcendental I-thoughts articulate the logical form of any judgment’.

Newton further wonders whether my claim about transcendental apperception not being subject to illusion or deception (more precisely, that I cannot misapply the categories) is not obviously at odds with the fact that so many metaphysicians seem to be mistaken about the claims they make. So ‘What sort of errors do they commit, if they are not errors in the misapplication of categories such as “substance” or “cause” to things in themselves?’ Kant writes at the start of the A-Deduction:

Thus metaphysicians are not deceived because they misapply the categories to things that reach further than experience. But they are mistaken in believing that while employing the categories a veridical cognition of such things is possible without the help of possible experience, i.e. without empirical intuition. The lack of criticism on their part is not due to a failure ‘to grasp [their] own acts of synthesis’; rather, they fail to see that a legitimate act of intellectual synthesis is objectively real only if it applies to or has an effect on our sensibility (cf. B152). That a particular claim turns out to be illusory, i.e. not based on an empirical intuition of a real object, does not imply that the transcendental apperception necessarily involved in the claim is illusory.

I am not suggesting, as Newton seems to think I am, that transcendental truth and empirical truth come apart in any actual empirical true judgement about a given object, or that in such a judgement the ‘relation (between substance and accident) could exist without the empirical relata existing’. Far from it. However, while it is of course trivial that a false empirical judgement about an object is not empirically true, nonetheless transcendental truth does necessarily apply to it as it does to a true empirical judgement about an object. To take Newton’s example of a brown table which I falsely judge to be green: I do not misapply the categories ‘substance’ and ‘accident’ when I falsely judge the brown table to be green, but I falsely attribute a certain accident, namely <green>, to the substance that is a brown table. A false judgement about a given object x is still a judgement about x, and so necessarily involves, as in the above case, the categories ‘accident’ and ‘substance’. As I argued in the book, I do not see why in a false empirical judgement about some given object categories would be misapplied because the judgement is false. (One could of course misapply, say, the category of substance and accident in that what one thought was a substance turns out to be an accident of a more fundamental substance.)

Lastly, Newton contends that the limitations that I impose on our intellectual grasp of things in themselves,Footnote ⁹ that is, that we have no cognitive access to things in themselves whatsoever, result in ‘Kant’s entire project of grounding the objectivity of cognition through a priori concepts of the understanding … fall[ing] to the ground’. She says: ‘For our understanding will not, under this assumption [i.e. of the limitations I suggest], know the objects as they really are, but will only have access to the ways they seem to be.’ This is a familiar Hegelian objection against Kant’s idealism,Footnote ¹⁰ but it is based on a conflation of things in themselves and objects; objects are appearances, mere representations, they are not things in themselves. The objective reality of the objects of my cognition is not in the least compromised by my claim that our discursivity implies that we are complete ignoramuses about things in themselves. Things in themselves are not objects ‘as they really are’, they are in fact not possible objects at all.Footnote ¹¹ There is nothing more objectively real than the objective reality established by the categories, given sensory input (see chapter 7). There is thus no sense in which our so-called perspective on objects limits our view of how objects really are; we are rather intimately connected with the objectively real world, with nature, in virtue of our transcendental capacities cum empirical intuition (and given the assumption of the systematicity of nature). By contrast, things in themselves concern the intrinsic nature of things, which of course is metaphysically more real than, and grounds, the objective reality of the objects (appearances) that we are able to know. But being ignorant about the intrinsic (metaphysically more basic) nature of things, of how they are in themselves, does not imply that we are lacking in knowledge, in principle at least, about the objective reality of the objects (appearances), of how they really are.Footnote ¹²