![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
This paper relates Kant’s account of pure apperception to the agential approach to self-knowledge. It argues that his famous claim ‘The I think must be able to accompany all of my representations’ (B131) does not concern the possibility of self-ascribing beliefs. Kant does advance this claim in the service of identifying an a priori warrant we have as psychological persons, that is, subjects of acts of thinking that are imputable to us. But this warrant is not one to self-knowledge that we have as critical reasoners. It is, rather, an a priori warrant we have, as thinkers, to prescribe to given representations their conformity to principles of thinking inherent in our capacity of understanding itself.
Notes
1. A proponent of the agential approach to self-knowledge need not hold that the only warrant we have, or could have, for self-knowledge derives from our beliefs being embedded in our critical reasoning.
2. The most important such proponent is Tyler Burge, who in a series of influential papers has developed detailed accounts of a priori entitlements that we have as persons, entitlements that include one to self-knowledge. These accounts, as he puts it, ‘have a Kantian flavor’ (Burge Citation2003, p. 335 n50). In particular, in ‘Our Entitlement to Self-Knowledge,’ Burge develops an account of an epistemic entitlement to self-knowledge that turns crucially on the claims that being a critical reasoner requires that one be the subject of ‘mental acts and states that are knowledgeably reviewable’ (Burge Citation1996, 98) and that this reviewability requires that one have a certain non-observational entitlement to first-person judgments regarding those acts and states (ibid). Burge’s emphasis on the possibility of a subject’s applying ‘the I think’ to her thoughts echoes Kant’s famous discussion at B132f. I hasten to add that Burge is careful to distinguish his position from Kant’s (see, e.g. Burge Citation1996, p. 99, n5), and his aim in these papers (Burge Citation1988, Citation1993, Citation1996, Citation2003, Citation2011) is not Kant exegesis (although he does offer several penetrating observations about Kant along the way). And I need especially to stress that the reading of B132f., and more generally of Kant’s account of the unity of apperception, that I will be rejecting is not one that Burge endorses.
3. See Evans (Citation1982), Boyle (Citation2009), and Setiya (Citation2011).
4. Here ‘understanding’ is to be taken in the sense in which understanding contrasts with judgment and reason. Kant also occasionally uses this term in a broader sense to refer to our higher capacity of cognition, which encompasses judgment and reason, as well as understanding in the narrower sense of the term. To look ahead, to understand something, in turn, is to represent its essence positively with a warrant sufficient for that representation to serve as a principle of our cognition. This does not require that the representation constitute cognition of the essence.
5. The term ‘transparent’ is one coined by Roy Edgley (Citation1969, p. 90).
6. As Patricia Kitcher points out, in reading Kant this way, Evans is following Strawson (see her Kitcher Citation2017, discussed below).
7. Moran and Boyle are happy to allow that a critical reasoner may have other modes of access to her beliefs in addition to the agential one. All they claim is that agential access is privileged, in that it tracks the proper functioning of the capacity for belief. So the thesis of the transparency of belief to which they subscribe is weaker than that which Evans champions: they claim only that there are some cases in which a critical reasoned can determine what she believes by addressing world-directed questions, and not that the only way in which she can determine what she believes is by addressing such questions.
8. Neither Burge nor Moran reads Kant as advancing, at B131-2 or elsewhere, the claim that our beliefs are transparent.
9. One might worry here that what Kant says in this passage is inconsistent with his famous claim that concepts and intuitions constitute cognitions only in conjunction with one another: this claim seems to entail that concepts and intuitions cannot be species of cognition. But this inconsistency is only apparent. On Kant’s account, a concept can be considered either of itself, or as it constitutes a cognition in conjunction with intuition, and the same goes for intuition. At A320/B372, Kant is using ‘concept’ and ‘intuition’ to refer to these representations as they constitute cognitions.
10. Burge and Moran, as well as other proponents of the agential approach, maintain that our knowledge of our sensations is of a different character than our knowledge of our thoughts and attitudes. And they exclude sensations from the nonobservational self-knowledge that a critical reasoner can, as such a reasoner, have of her mental acts and states, on the grounds that sensations are not essentially the content of her thoughts and attitudes. See Burge (Citation1996, 107), Moran (Citation2001, 9–10), as well as Aaron Zimmerman (Citation2008) and Matthew Boyle (Citation2009). For a helpful discussion of this point, as well as of what, following Burge, has come to be referred to as the rationalist approach to self-knowledge, see Brie Gertler (Citation2011, especially Chapter 6).
11. A related problem is worth mentioning: Kant distinguishes between two sorts of cognitions – concept and intuition – and it isn’t at all clear that to have a perceptio that is an intuition is, itself, to think it, or indeed to take any attitude toward it. Kitcher addresses this concern in the course of developing a reading of Kant on which his position is amenable to the agential approach to self-knowledge (Kitcher Citation2011, 151f).
12. This line of thought seems to raise difficulties only for Evans’s transparency thesis, on which the question of what one believes always can, and is to be, settled solely by answering the world-directed question. And, even so, it isn’t entirely clear what the difficulty is supposed to be. Kitcher seems to be assuming that the prior de se consciousness one has of oneself as the subject of rational thought that Kant’s account posits is either itself a belief that cannot be determined in the world-directed fashion, or a source of such beliefs. Thanks to Thomas Land for raising this worry.
13. This is the central thesis of her (2017). See also Chapter 15 of (2011).
14. ‘Aktus’ and ‘actus’, in Kant’s terminology, signifies, not just any action (Handlung), but an actuation, or inner action, i.e., a subject’s activity (Activität) insofar as it suffices to realize its capacity (Vermögen) as a power (Kraft). See here the final section of my (2009).
15. Where Kant parts company with Descartes, and others, is in denying that <I think> is cognition of the thing that thinks: it is a mere representation of the thing that thinks, because we cannot prove that any such thing is really possible. He maintains that the representation <I think> has the function of making it possible for one to determine one’s own thinking out of one’s de se consciousness of how one is, in bringing them under intellectual principles, to relate representation to an object. And he holds that the <I think> can serve this function without being cognition of the thing that thinks. It can, and does, serve this function in virtue of constituting cognition that a human being can have, through pure apperception, of herself in respect of ‘actions and inner determinations’ of which she must, timelessly, be the subject, if she is, as a human being, to be a person at all (A546/B574).
16. Consider here how, in his 4 December 1792 letter to J.S. Beck, Kant reacts to Beck’s attempt to gloss the Claim. Kant offers the following emendation to Beck’s attempt: ‘Instead of … “The I think must accompany all the representations in the synthesis,” “must be able to accompany [begleiten können]”’ (11: 395). Notice that Kant does not correct Beck’s gloss of the Claim in any way other than to add the modal qualifier ‘können.’
17. One might object that the objective validity of the principle of contradiction cannot depend on our thought’s relation to things because pure general logic abstracts ‘from all content of our cognition, i.e. from any relation [Beziehung] of it to the object [Object]’ (A55/B79). To be sure, the principle of contradiction does, on Kant’s account, abstract from all relation of our cognition to the object. But this is entirely consistent, on his view, with its owing its standing as cognition to its being applicable to objects of our experience, much as pure mathematics owes its objective validity to there being ‘things that can be presented to us only in accordance with the form of our sensible intuition’ (B147), despite its abstracting from the dynamical character of these things.
18. For a helpful discussion of this point, see Keller (Citation1998).
19. For a reading of Kant’s conception of the spontaneity of cognition along these lines, see my (2009).
20. Here it is useful to contrast the present reading with a similar one that Dieter Henrich considers, and rejects, in (Henrich Citation1976), p. 79f. On this reading, Kant conceives of the subject as having strict Leibnizian identity (one that requires that identicals be indiscernible). Henrich rejects this reading as incompatible with Kant’s characterization of the capacity that is essential to the subject as one for altering its condition in and through reflection on how it is to think. The present reading provides the resources for distinguishing between pure apperception itself, which constitutes the thinker as transcendental subject, and the alterations it determines in itself as empirical subject.
21. This provides some indication of how the present reading is not subject to a second worry that leads Henrich to reject his reading of the subject as enjoying strict Leibnizian identity (Henrich Citation1976, p. 79f.). The worry is, in effect, that Kant cannot provide a case for such an account of the subject because we are not given to ourselves, in consciousness, as transcendental subjects. The response is that Kant does not purport to arrive at his account from what is given to him, or us, in a clear consciousness.
22. For details, see ‘Essence, Nature, and the Possibility of Metaphysics,’ forthcoming. Here I can, given the constraints of this paper, only gesture at the approach to Kant’s critical philosophy that I develop and defend here, and elsewhere.
23. This paper was written for the conference Transparency and Apperception held at Ryerson University in May 2018. I thank the organizers of this conference – David Hunter, Thomas Land, and Boris Hennig – and other conference participants for helpful discussion. Part of this paper was presented at the UCLA Kant conference held in February 2019. I thank John Carriero for organizing this conference, and to all its participants for their helpful comments. I also thank Robert M. Adams, Tyler Burge, Tom Christiano, Suzanne Dovi, Frode Kjosavik, Thomas Land, Shaun Nichols, Santi Sanchez, and Mark Timmons for helpful comments and discussion.
Additional information
Notes on contributors
Houston Smit
Houston Smit is Associate Professor of Philosophy at the University of Arizona.