Right at the beginning of Kant's Thinker, Patricia Kitcher describes Kant's theory of the cognitive subject as constructed around the notion of apperception as ‘complex, subtle, and immensely fruitful’ (Kitcher Reference Kitcher2011: 3). She indeed informs us about many complexities and subtleties – the historical background, the meaning of the theory, and the arguments surrounding it. Her account clearly raises the state of the art to a new level. But does she establish the fruitfulness of Kant's theory?
Kitcher does not tell us what she means by ‘fruitfulness’ but two kinds of analyses, known from current work in the history of philosophy, come to mind. The first begins with the reconstruction of an author's premises and conclusions, relating them more or less to their historical background, and then defends the author's views against certain charges raised against him or her. Let us call this ‘internal fruitfulness’. Distinct from this is the attempt to prove an ‘external’ fruitfulness. It originates from problems we are currently faced with, and attempts to show that the classical author's theory can help us to solve them. Of course, the boundary between these kinds of fruitfulness is not a sharp one, but there are clear cases of both. It appears that chapters 12–14 of Kant's Thinker deal with issues of internal fruitfulness; and Kant scholars often regard a clearer reconstruction of a Kantian argument, or a clever defence of it against objections, as fruitful enough. But Kitcher thinks there is more to be learned from Kant's theory of the subject. In chapter 15, she straightforwardly claims that we can treat Kant as ‘our contemporary’, and that his theory ‘offers “new” and plausible perspectives on issues of considerable recent interest’ (2011: 249).
I think that Kitcher fails to support this claim. Of the four topics she singles out to prove her point I shall address two. First, are transcendental arguments able to convince the sceptic? Second, how can we explain that (a certain kind of) self-knowledge possesses a special authority? These are closely related, and the second topic possesses a special importance. Not only does Kitcher devote the longest section in chapter 15 to it but the issue (explained in relation to views put forward by Peter Strawson and Gareth Evans) is also used to motivate her entire study (Kitcher Reference Kitcher2011: 3–5). The other two discussions – which try to make Kant's theory fruitful for current debates over consciousness and the problem of other minds – raise worries as well, but I cannot address them here.
1. Transcendental Arguments Against Scepticism about Self-Consciousness
Kitcher first discusses Barry Stroud's well-known objection against Strawson. Stroud rejects the usefulness of transcendental arguments as tools against scepticism (about the external world, other minds, the necessity of basic presuppositions of knowledge as presented by Kant, etc.). Such arguments, he argues, can only be successful if they do not merely show that we must accept certain beliefs but also that these are true independently of what we believe. Thus, a transcendental argument in favour of the principle of causation does not merely have to show that we must use that principle. It must also show that the principle is true (Stroud Reference Stroud1994: 231–2). Kitcher assumes that a similar objection might be raised against Kant's claim about apperceptive self-consciousness – and argues that her justification of the claim has the ‘ability to withstand the classic objection to transcendental arguments’ (Kitcher Reference Kitcher2011: 249).
Very briefly put, her argument runs like this (Kitcher supplies many details and background assumptions, especially in chapter 9). For the kind of cognition that applies concepts to items of perception, conscious synthesis is crucial because without it, thinkers could not know the reasons for their cognitions. Faced with a series of four stroke symbols, an ox may have a visual representation of them. By contrast, judging that ‘these are four strokes’ means subsuming them under the concept of ‘4’. In order to do this, one has to use a counting rule and then start counting: ‘1 stroke, 2 strokes …’ until one sums up the different items and makes one's judgement (e.g. 2011: 128, 174). But that is possible only if one keeps in mind those different representations and if one knows, at least implicitly, that it is oneself who is combining them in order to come to a judgement about the number of strokes. ‘Conscious synthesis is crucial for rational cognition. Without it, … rational cognition is impossible, because cognizers would not know the bases of their cognitions. With that consciousness, however, the cognizer creates a relation of rational dependence across his states in part by being at least implicitly cognizant of that relation’ (2011: 252).
Can this argument ‘withstand the classic objection to transcendental arguments’? Kitcher agrees with Stroud that self-verifying propositions are exempt from sceptical doubts, propositions such as ‘I am speaking’ or ‘I am alive’, ‘I am thinking’ or ‘there is thought’ (Stroud Reference Stroud1994: 233). These propositions might be false – for instance, there might be no thinking beings – but whenever I say ‘I am speaking’ that statement is self-verifying and, mutatis mutandis, the same holds for ‘I am thinking’. But Kitcher claims that ‘I think’, understood as the expression of apperceptive self-consciousness, is not self-verifying at all. Rather, the ‘key point’ that saves it from radical doubt is that
it (partially) creates its referent: Given the presence of appropriate representations and mental act awareness, this representation creates the reality to which it refers. … the ability of cognizers to deploy ‘I-think’ is part of the complex mental act that brings about the relation of rational dependence across mental states that constitutes their existence as thinkers. (Kitcher Reference Kitcher2011: 252; emphases added)
I doubt that Stroud's sceptic will be impressed by this. Here is why. First, what is the ‘reality’ that is ‘created’ in an act of thinking? Apperceptive self-consciousness or the ‘I think’ is, in itself, not a complete thought or knowledge-claim. As Kant says, this ‘poorest of all representations’ (CPR, B408) describes only the form, not the content of a knowledge-claim (e.g. CPR, A346/B404; A382). It only indicates a specific mental process – namely, a spontaneous act of judging. But to make a knowledge-claim proper we have to speak of ‘I think that p’, where p stands for some judgemental content. Kitcher (e.g. 2011: 260) notes this herself, and it helps to make sense of her reluctance to accept Stroud's claim that ‘I am thinking’ is self-verifying: while such a thought is possible, it remains elliptical. Thus, the ‘reality’ that we are talking about is whatever makes true (or false) a complete statement such as ‘I think that there are 4 strokes’. But then Kitcher faces a dilemma, depending on where one puts the emphasis. Either such a statement is about there being four strokes; but then surely the sceptic will object that I may have miscounted. Or Kitcher means – more consonant with her discussion – that what I cannot doubt is that I think that there are four strokes. Then her point fails for a different reason. Even assuming that such a belief is unavoidable, what Kitcher would have to show is that it is true independently of what we believe, or must believe. But her claim that the referent of ‘I think’ or even ‘I think that p’ is created in an act of cognition undermines exactly this.
There is a further, related problem. As Kitcher notes, Kant's claim that apperceptive self-consciousness is necessary for knowledge is, strictly speaking, a claim about a necessary possibility: the ‘I think’ does not always have to accompany all of one's representations, but it is necessary that it is possible that it does so (CPR, B132). As Kitcher (e.g. 2011: 252, 261) says, we must only be ‘implicitly’ conscious of our synthetic activities. There are compelling reasons for this. We do not always explicitly think about our cognitive activities, let alone in all their details. If we did, that would rapidly lead to cognitive overload. While Kitcher by no means disputes this, she occasionally ignores the qualification resulting from it. Her contention that the ‘I think’ ‘creates its referent’, and is therefore exempt from doubt, can only work if the cognitive activity is explicit. But it does not have to be that way.
It has been noted that scepticism about the external world was ‘not high up on Kant's philosophical priority list’ (Patzig Reference Patzig1979: 71). This is correct to an even greater degree when it comes to his views about apperception. Neither was he eager to refute scepticism about self-knowledge (his aim was the validation of the categories, after all), nor do I see a good reason to use his theory as a tool against such scepticism. This is not a case where ‘our’ problems and Kant's are related in a fruitful way.
2. The Special Authority of Self-Knowledge
Next, Kitcher claims that Kant's theory can be made fruitful for current discussions about the special authority of self-knowledge. This issue is not about the sceptic's gap between having to take something to be true versus its being true, but about how to explain that (some) self-knowledge is known in a first-person rather than a third-person way or, to use Strawson's terminology, that (some) mental states can be self-ascribed without any criteria. As the story goes, I do not have to make any observations to know that I am in pain, experience a noise or make a judgement. Why?
Kitcher discusses the views of a number of authors who, following Sidney Shoemaker, try to explain the (alleged) impossibility of ‘self-blindness’ about such cases. Just to be clear, such an explanation is a complex affair. There are at least two things to consider. As ‘I am feeling happy’ is a special case of a singular statement of the form ‘a is F’, one should first try to answer why we cannot be mistaken about the reference of the subject term. Second, why can we not be mistaken in applying the predicate term to the subject term? Many current debates under the heading of the ‘immunity from error through misidentification’ focus on the former question only. But what Kitcher wants to explain is not this, but why one cannot be mistaken about ascribing a certain property to oneself – namely, the property of thinking that p.
She groups the positions into two main camps: those who give a too ‘personal’ account of the impossibility of self-blindness, and those who provide an altogether ‘impersonal’ account. Richard Moran, for instance, belongs in the former camp. He explains the impossibility of self-blindness by the assumption that our epistemic judgements are ‘up to the subject’. In the latter camp belongs the ‘transparency’ account of Evans. He points out that when I ask myself whether I believe that milk is liquid I use the same procedures as when I ask whether milk is liquid. According to Kitcher, Kant's theory of self-knowledge belongs in neither of these camps. Instead, the privileged self-ascription of a belief takes place just in case (i) the belief is a rational cognition, i.e. is a belief for which the cognitive subject knows the reason, and (ii) the cognitive subject herself has consciously judged that belief and reason appropriately hang together. The details of Kitcher's discussion are very interesting; unfortunately, they are also too complex to be recounted here. I must confine myself to indicating why I fail to see the alleged fruitfulness of Kant's theory for the contemporary debate.
First, Kant is not concerned with the subject-matter of the debate between Evans, Moran and others. As Kitcher (Reference Kitcher2011: 5, 265) herself emphasizes, what Kant intends to explain is not the ‘mineness’ but the ‘togetherness’ of representations. If his account did also explain the mineness of representations, this would be a rather surprising by-product. We would like to understand how that is possible.
Second, Kitcher's statement that ‘if I am implicitly conscious in making a judgment then I am entitled to self-ascribe it’ (2011: 261) leads to several difficulties. One might agree that if one is actually consciously making a judgement, then one can indeed self-ascribe the judgement – but if the apperceptive activity is merely ‘implicit’, then it is not obvious that the judgement is self-ascribed. Once again, Kant only asserts that it is necessary that it is possible that the ‘I think’ accompanies my representations – rather than arguing that it always accompanies them – and so his claim is not about actual (and criterionless) self-ascription. Conversely, the current debate is not about whether we are ‘entitled’ to self-ascribe a judgement, but how we actually do so in the absence of criteria.
Thus, there are two gaps that would have to be bridged to make Kant fruitful for the present debate: between mineness and togetherness, and between actually self-ascribing a representation and being entitled to do so. Perhaps a closer historical and philosophical reflection would have been helpful here. For instance, how are the problems of the mineness and togetherness of representations related? Against what background assumptions did they emerge, and how did they develop further? Are the historical trajectories of the problems related such that we can understand how an account that was not made for one of them can nonetheless be applied to it?
Relating Kant to debates of one's own times is always a thorny affair. In an article published under the heading ‘Was soll uns Kant nicht sein?’, the philosopher-psychologist Wilhelm Wundt (Reference Wundt1892) attacked Kantians for being uncritical about how certain of Kant's doctrines were the result of time-bound problems. At the same time, Wundt claimed that it is possible to investigate apperception psychologically by an experimental analysis of complex reaction times and the steering of attention in perception. We may find it hard to recognize Kant in this. I wonder whether later generations may not find it equally hard to see Kant as responding to the problems Kitcher deals with in chapter 15 of her otherwise highly insightful study.Footnote 1
