![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
Kant is said to be the first to distinguish between consciousness of oneself as the subject of one’s experiences and consciousness of oneself as an object, which he calls transcendental and empirical apperception, respectively. Of these, it is empirical apperception that is meant to enable consciousness of any empirical features of oneself; what this amounts to, however, continues to puzzle interpreters. I argue that a key to understanding what empirical apperception consists in is Kant’s claim that each type of apperception corresponds to a distinct type of unity of apperception—that is, a distinct way in which representations can be related for a subject. Whereas transcendental unity of apperception requires that representations be actively combined by the understanding, empirical unity of apperception obtains when representations are passively combined by the reproductive imagination. In light of this, I develop a novel account of Kant’s two types of apperception, according to which they correspond to a cognitive subject’s consciousness of two essential aspects of herself—namely, her spontaneity and receptivity.
Disclosure Statement
No potential conflict of interest was reported by the author.
Notes
1 Kant characterizes apperception in general as a subject’s consciousness of itself (B68, A107, B132).
2 Versions of this view are suggested by Smith [Citation1923: 293–4], Mohr [Citation1991: 104–5, 157–8], Brook [Citation1994: 56–7, 66–7, 78], Ameriks [Citation2000: 252–4], Sturm [Citation2001: 174], Allison [Citation2004: 277–80], Rosefeldt [Citation2006: 287], Sturm and Wunderlich [Citation2010: 55], Kitcher [Citation2011: 25–6, 129, 159], Schmitz [Citation2013: 1058], Renz [Citation2015: esp. 593–4], Bader [Citation2017: 132], Longuenesse [Citation2017: 91, 110], Kraus [Citation2019: 172, 190], and Khurana [Citation2019: 964]. A variant treats empirical apperception through inner sense as a subject’s awareness of any of her representings as acts of awareness [Weldon Citation1958: 261–2; Wolff Citation1963: 198–9]. Longuenesse [2006: 302] briefly lists possible readings of EA, one of which is similar to the view that I defend. While the correspondence between EA and the empirical unity of apperception is central to my view, however, Longuenesse’s alternative takes the representations of which the subject is conscious in EA to stand in the transcendental unity of apperception [ibid.: 305]. Keller [Citation1998] and Valaris [Citation2008] argue that EA, through inner sense, makes a subject aware of her ‘point of view’ on outer experience. As Valaris notes, this reading is an ‘inference to the best explanation’ [ibid.: 3] and not directly supported by Kant’s text. I will argue that the text directly supports a different reading.
3 Kant also calls TUA the original/pure/synthetic/objective unity of apperception/consciousness/self-consciousness (A107, B139–40, B135–6) and EUA the subjective unity of apperception/consciousness (B139–40).
4 Arguably, Kant allows that representations can be united in one consciousness—namely, when a subject has representations that bear some relation to each other—without being united for that consciousness—that is, without the subject being aware of her representations as so related. See note 8.
5 It might be objected that Kant is not claiming here that EUA obtains prior to combination by the understanding, but only that the manifold of intuition is given for combination through inner sense. While the sentence in question is ambiguous, I believe that my reading is preferred when read together with the subsequent sentence that I highlight below, in which Kant claims that EUA obtains through association of representations. As I will discuss, he argues that association by the imagination in the synthesis of reproduction is necessary for and precedes the synthesis of the understanding by which the manifold of intuition is united in a concept of the object. This reading is also supported by Kant’s claim in §8 that ‘consciousness of itself (apperception)’ through inner sense—i.e. EA—’requires inner perception of the manifold that is antecedently [vorher] given in the subject, and the manner in which this is given in the mind without spontaneity’ (B68, my emphases). This suggests that EA—like EUA, which it accompanies—requires consciousness of intuitions as they are given prior to combination by the understanding. I thank two anonymous referees for urging me to defend my reading of this sentence.
6 Schmitz does argue that what one is aware of through inner sense is temporally ordered outer intuitions [Citation2013: 1052]. She does not, however, discuss EUA, or its relation to association.
7 As I explain in section 3.1, Kant does argue that association by the reproductive imagination plays an essential role in awareness of a temporal series of intuitions as a series. But the Standard View does not recognize this connection between awareness of temporal relations and association. Kitcher notes that EUA is said to occur through association. She does not, however, connect it to EA, but treats it as a ‘special case’ of TUA that is ‘irrelevant to cognition’ [Citation2011: 158–9]. I disagree with both claims. First, as I will discuss, Kant identifies TUA with the objective unity of consciousness, which requires that representations be fully determined by the understanding. In contrast, EUA is a merely subjective unity of consciousness that obtains in merely associated representations that are not fully determined by the understanding. Thus, EUA cannot be an instance of TUA. Second, I will argue that, far from being irrelevant to cognition, EUA underlies cognition of both outer objects and the self.
8 To satisfy this condition, S need not be aware that intuitions are associated. It is sufficient that (1) S is aware of the temporal relations between a set of intuitions; (2) these temporal relations obtain merely because of corresponding associations; and (3) S does not judge whether these temporal relations are required in light of the objects that the intuitions are of. I explain (3) in sections 3.3 and 3.4.
9 Since I do not have space to discuss the difference between ‘intuition’ and ‘perception’, I treat these terms interchangeably. See Tolley [Citation2020] for discussion.
10 See Strawson [Citation1970] and Matherne [Citation2015].
11 This is compatible with the claim, defended, e.g., by Allais, that intuition alone suffices to present particulars in some sense to a subject. As Allais allows, this is consistent with its being the case that perceiving something requires the subject to be aware that she is doing so [Citation2017: 26].
12 Of course, it is possible to be aware of smoke after being aware of fire, without coming to be aware that one is aware of smoke after being aware of fire. Animals have the former kind of awareness without the latter. See McLear [Citation2011].
13 Thus, I disagree with Valaris’s claim that consciousness of a ‘temporal series of outer perceptions in me’ just is consciousness of ‘temporal goings on outside me’ [Citation2008: 15]. I can be aware of a representation as of fire followed by a representation as of smoke, without judging that these representations correspond to an actual succession of fire and smoke outside me.
14 See Longuenesse [Citation1998: 212–25], Allison [Citation2004: 189–92], Longuenesse [Citation2005: 64–78], and Messina [Citation2014] for discussions of the role that the understanding plays in determining pure intuitions of space and time.
15 Henceforth, I typically avoid confusion by using the generic term ‘representation’ rather than ‘intuition’ to mark that—as discussed above—the awareness in question involves the understanding.
16 Recall Kant’s claim that apprehension is inseparable from reproduction.
17 That is, although P was only perceived at t1, since Q was perceived at t2, and Q inheres in P, P must exist at t2.
18 §18 is entitled ‘What objective unity of self-consciousness is’ and begins thus (B139, see also B137, A105–8):
The transcendental unity of apperception is that unity through which all of the manifold given in an intuition is united in a concept of the object. It is called objective on that account, and must be distinguished from the subjective unity of consciousness …
19 Since my main focus in this paper is on EUA, I set aside questions about how pure intuitions are determined by the understanding.
20 More specifically, this requires that no element of the set is left undetermined by a category. Kant argues that the objective unity of consciousness (i.e. TUA) requires that intuitions be determined ‘with regard not only to one, but rather to all the logical functions of judgment’ (R5932, 18: 391). This cannot mean that every intuition must be subsumed under all twelve categories, since some are incompatible. Rather, it (and its relations) must be determined by at least one category from each of the four titles. My thanks to an anonymous referee for raising this question.
21 One might object that this conflicts with Kant’s claim in §15 that application of the categories presupposes a still ‘higher’ unity (B131). While I do not have space for a full discussion, I take Kant’s main point to be that the unity of TUA ultimately depends on the identity of the act of uniting representations in one consciousness: that is, it depends on the same subject being conscious of each representation and of uniting them. That identity is presupposed by, and is in that sense ‘higher’ than, any particular act of categorial unification performed by a subject (see A107, B413). However, Kant is clear that the subject only comes to be conscious of the identity of her acts in virtue of performing them——i.e. through combining representations in accordance with the categories (A108; see also B133–5). Each such act generates representations that are instances of TUA. My thanks to an anonymous referee for raising these issues.
22 The distinction that I am drawing here appears to cut across the two types of unity discussed by McLear [Citation2015]. On the one hand, EUA is a unity brought about by sensibility (broadly construed to include the reproductive imagination), independently of the understanding (although, as I have noted, application of the mathematical categories is necessary in order for the subject to think of EUA as a determinate unity). Moreover, in being conscious of EUA, the subject does not think of the combination of representations as required by the objects that the representations are of. In both respects, EUA is closer to what McLear calls ‘aesthetic unity’. However, it is not a unity for which the whole is prior to its parts; rather, as with what McLear calls ‘discursive unity’, the whole is composed out of the intuitions that are its parts.
23 One exception is Tolley, who carefully distinguishes between perception and cognition. He does not, however, identify the former with EUA, and appears to ascribe unity of apperception as such to the understanding [Citation2020: 3227n27].
24 Thiel notes that, in addition to a ‘substantial unity’ of the self and a unity that is ‘a necessary condition of mental activity’, Tetens—whose views on self-consciousness are widely believed to have influenced Kant’s—also identifies a unity of the self that is ‘psychological’ and ‘observed’ [Citation2015: 157–8, Citation2018: 68–71]. In contrast, Thiel claims that, for Kant, there is no ‘observed unity’ of the psychological self based on inner sense alone [Citation2015: 159, Citation2018: 73]. But this at least partially overlooks Kant’s distinction between TUA and EUA. Although EUA does not itself afford awareness of an identical self, Kant explicitly characterizes it as a type of unity of apperception that is both passively observed and governed by psychological laws. On my view, there is an important similarity here between Kant and Tetens.
25 I have modified the Guyer/Wood translation to read ‘when’ rather than ‘if’. ‘Wenn’ is ambiguous between the two; however, the former better conveys that this judgment reports that a sequence of representations is ‘found together in perception’.
26 I take the former to be expressed by what Kant calls a ‘judgment of perception’ and the latter by a ‘judgment of experience’ (Pr., 4:299–301). See Sethi [Citation2020] for a full defence.
27 A related question asks whether the reproductive imagination can generate a genuine unity of consciousness. I believe that it can: Kant speaks specifically of the ‘unity of association’ (A121) and the ‘empirical unity of consciousness’ that results when perceptions are ‘combined by the imagination’ (R5933, 18: 392; see also B139–40), claiming neither that such unities are impossible nor that they are simply instances of the objective unity of the understanding. Rather, he describes them as subjective unities, albeit ones of which one can be conscious as unities only on the basis of an ‘objective ground’—namely, the unity of time in which they are represented (which itself depends on TA [B140]). See also note 39.
28 E.g. A194/B239, A156/B195, A201–2/B247, A368. Representations lack ‘relation to an object’ when they are not judged to correspond to an object that the representations are of. See Tolley [Citation2020].
29 E.g. A53/B77, A201–2/B247, A376; Anthropology 7:133–7, 7:180, 7:208, 7:240n21, 7:241; CJ, e.g., 5:230, 5:243–5, 5:256, 5:317, 5:321, 5:323, 5:350, R6315, 18:621. Cf. A239/B298.
30 Kant draws a related—but, I believe, different—distinction between the active and passive self in his discussion of self-affection (B153). There, his main point seems to be that, in EA, the self is both the active subject of consciousness and the passive object that is affected and given to consciousness (see B68, 20:270). The contrast in the Anthropology goes further and concerns whether the subject is active or passive with respect to the manifold of intuition in general. I discuss self-affection in note 47.
31 Boyle also claims that Kant distinguishes between knowledge of the self as active and as passive. Since Boyle’s goals are not exegetical [Citation2009: 134], however, he does not fully defend this reading. Moreover, he takes knowledge of the self as passive to consist merely in knowledge of one’s sensations and appetites, and he does not mention EUA or the imagination’s role in producing it. On my view, both are central to EA as awareness of a unity of representations that is passively brought about.
32 Note Kant’s description of a ‘play of impressions’ as those that occur ‘from nature’ rather than through the subject’s activity.
33 Or, less ambiguously, ‘When I represent a body, I represent weight.’
34 Although Kant paradigmatically describes TA as accompanying acts of unifying intuitions, it also accompanies acts of combining mere concepts (A79/B105, R6311,18:610–11).
35 One might worry that this definition is not general enough, since it ties TA to the understanding in particular, rather than to theoretical reasoning in general. I believe that the text supports this: Kant consistently links TA with the understanding, even substituting the former for the latter (A94/B127, 18:272, 23:18). Although I cannot fully defend this, I don’t take this to mean that there are uses of reason not accompanied by TA. For one, the principles of theoretical reason serve precisely to regulate uses of the understanding (A665–6/B693–4). As for thoughts involving ideas of reason, these are akin to thoughts involving concepts without intuitions. I am indebted to an anonymous referee for urging me to consider the applicability of TA in contexts other than the synthesis of intuitions.
36 Here, one might point out that it is equally true on my view, as on the Standard View, that consciousness of EA is a consciousness of representations qua representations (although we disagree about which representations). The phrase ‘qua representations’ is employed by the Standard View to flag that, in EA, the subject turns her attention away from the objects represented by fully synthesized representations to the very same fully synthesized representations qua representations in time. In contrast, my view does not hold representations fixed across TA and EA: I have argued that TUA requires a further act of subsuming representations with EUA under the relational categories. Thus, the phrase ‘qua representations’ is not needed to capture the difference between EA and TA, on my view, although neither is it inaccurate. I thank an anonymous referee for pressing this point.
37 On Hoppe’s reading, in contrast, Kant denies that there can be any non-objective unity of consciousness [Citation1983: 129–40, 221–2] and considers consciousness of non-objective states to be entirely separate and scattered [ibid.: 132ff.].
38 See Corr. 11:52 and McLear [Citation2011].
39 The fact that our capacities for TA and EA depend on each other in these ways does not entail that there is no in-principle difference between them. I have argued that TA and EA, respectively, accompany awareness of two types of unity that, Kant makes clear, are in-principle different: TUA is an objective unity fully determined by the understanding, and EUA is a merely subjective unity due to association by the imagination. Now, as discussed, in order for S to be aware of a merely associated series of representations in time as a unity (i.e. in EUA), she must represent time itself as a unity (i.e. in TUA). Moreover, in order to ascribe a merely associated series of representations to herself, she must be able to think of herself as the identical subject who can actively bring those representations into TUA. But the fact that her awareness of EUA depends in these ways on her capacity to bring representations into TUA does not entail that merely associated representations themselves are already in TUA, precisely because they are merely associated. Nor should the difference between TUA and EUA be thought to be one of degree. No increase in the degree of associative unity between representations can bring them into objective unity; rather, what is required is an act of a fundamentally different kind—an act of the understanding, rather than of the reproductive imagination.
40 Thus, I agree with Dyck that consciousness of myself as I appear (i.e. EA) is a necessary component of every perception. For Dyck, however, this is because the former consists in awareness of ‘the unity of time as an object’ [Citation2006: 43], which he claims to obtain through the active unification of successive acts of attention that accompany the synthesis of apprehension. On my view, whereas EA depends on awareness of time as a unity, it is not identical with it.
41 Of course, experience does not require that the subject actually self-ascribe either transcendental or empirical unities. Rather, it presupposes the kind of awareness that makes such self-ascription possible.
42 I believe that my account also aligns better with Kant’s conception of the ethical subject. In the Groundwork [4:452], he says that
a rational being … has two standpoints from which he can regard himself and cognize laws for the use of his powers … first, insofar as he belongs to the world of sense, under laws of nature (heteronomy); second, as belonging to the intelligible world, under laws which … [are] grounded merely in reason.
43 Nayak and Sotnak [Citation1995: 149] and Emundts [Citation2007: 197] claim that the category of causality is not applicable to inner sense. I disagree. First, Kant repeatedly describes representations as the effects of causes, both external and internal (e.g. A98, B276, A368). Second, he makes clear that inner perceptions are governed by laws of nature, such as the psychological laws of association. (e.g. B141–2, B152; Pr. 4:295; Anthr. 7:140–1). Third, contra Nayak and Sotnak, the Second Analogy does not argue that it is necessary to distinguish between a subjective and objective succession in order to apply the category of cause; it claims only that the latter is necessary for the former. Finally, despite initially suggesting otherwise, the discussion at B291–2 concludes (in line with the Refutation of Idealism) that we can think of the succession of inner states as alterations (and therefore as caused) if something persistent is given in outer intuition (see also 18: 611). See Hatfield [Citation1992] and Frierson [Citation2014] for discussion of Kant’s empirical psychology, and Sturm [Citation2001] for further criticisms of Nayak and Sotnak’s claim. See also Chignell [Citation2017].
44 Frierson argues that Kant allows empirical application of the category of substance to the soul [Citation2014: 21–6), whereas Kraus [Citation2019] argues that such application is merely regulative.
45 Representations bear causal relations not only to other representations, but also to objects. Plausibly, the latter type of relation also falls under the scope of empirical psychology, although I cannot explore this further here.
46 My emphasis on the natural laws of psychology might seem to draw my view close to ‘empiricist’ accounts of Kant’s theory of mind, defended, e.g., by Kitcher [Citation1990] and Brook [Citation1994]. There are some crucial differences, however. Most importantly, to the extent that Kitcher and Brook seek to give empiricist accounts, it is of the operation of all of Kant’s mental faculties, especially the activity of the understanding by which representations are brought into TUA, which they take to be governed by causal psychological laws [Kitcher Citation1990: 83, 122–3; Brook Citation1994: 5–7, 35]. Second, both discuss only briefly EA and inner experience, and appear to endorse versions of what I call the Standard View (see note 2). Finally, neither takes Kant’s empirical psychology and the laws of association to be relevant to cognition [Kitcher Citation1990: 77–9, 153ff.; Brook Citation1994: 8–9, 106–7, 135]. In contrast, I have argued that psychological laws govern only the syntheses of the reproductive imagination by which representations come to have EUA, whereas the subject brings representations into TUA in accordance with rational laws of the understanding (see note 47). Indeed, I take the difference between the kind of law that governs them to be crucial to the distinction between TUA and EUA. Moreover, I argue that it is precisely the laws of empirical psychology that underlie self-cognition.
47 One might object that this does not go far enough: surely it must also be possible to situate the active subject and her judgments in time? On my view, however, a judgment is essentially an act of the understanding for Kant, rather than an item that can be grasped through inner sense, and I do not think that he regards these acts as events in time. For one, everything that occurs in time is governed by laws of nature, whereas acts of the understanding are governed by rational laws (see McLear [Citation2020]). Second, in a Reflexion that asks ‘Is it an experience that we think?’, Kant argues that, since the acts of thinking that constitute experience involve determining the objective position of events in time, if they were themselves in time then it would have to be in a different, higher-order time in which the acts of first-order time-determination occurred. He concludes that this is absurd, and so that thinking itself is not experienced in time (R5661, 18:318–19; see R6311, where he indicates that mere thoughts without intuitions do not affect inner sense [18:611]). Third, he makes clear that it is only through her empirical character that the subject belongs to the phenomenal world, whereas the ‘actions and inner determinations’ that the subject is conscious of through ‘pure apperception … cannot be accounted at all among impressions of sense’ (A546–7/B574–5). How does this square with Kant’s doctrine of self-affection? In the above Reflexion, he says that, whereas consciousness of a mere thought of a square is not an experience, ‘this thought brings forth an object of experience or a determination of the mind that can be observed, insofar, namely, as it is affected through the faculty of thinking’ (18:319). Although it might seem like Kant is claiming here that the mere thought of a square can be observed, I believe that his point is rather that this thought only becomes an experience when an a priori intuition of a square is added to it through the imagination (see also B154). This is why Kant continues that, through this experience, one can ‘demonstrate [the square’s] properties’, since geometrical demonstration always requires construction in intuition for him. Space restrictions prevent me from further exploring his vexed doctrine of self-affection, but I do not believe that it amounts to the claim that thinking in general can be observed through inner sense. Rather, Kant ties self-affection in particular to the figurative synthesis (B153–4) and the successive acts of apprehension through which the manifold of intuition is grasped in time through inner sense. The latter enable consciousness of the subjective series of perceptions. As Kant says ([Leningrad Fragment on Inner Sense, tr. in Notes and Fragments, 366; see also B155, 20:270, and Dyck [Citation2006]):
That we can affect ourselves … is possible only through our apprehending the representations of things that affect us, i.e., of outer things, for thereby do we affect ourselves, and time is properly the form of the apprehension of representations which are related to something outside us.
48 I first formulated the main ideas in this paper as part of the core argument of my dissertation. I am indebted to my advisors, Hannah Ginsborg and Daniel Warren, for their invaluable guidance and insight. For discussion, advice, or comments on various stages of the paper’s development, I am very grateful to Gordon Belot, Richard Booth, Matt Boyle, Sarah Buss, Victor Caston, Peter Epstein, Patrick Frierson, Katharina Kraus, Olga Lenczewska, Béatrice Longuenesse, Ishani Maitra, Colin Marshall, Laura Ruetsche, Tad Schmaltz, Umrao Sethi, Clinton Tolley, Jessica Williams, the NAKS Workshop for Junior Women Scholars, audiences at the 2021 Pacific APA, University of Michigan, University of North Carolina, Chapel Hill, Clark University, as well as two anonymous referees for this journal.