Imagination and Association in Kant’s Theory of Cognition Corey W. Dyck (London, Canada) Befitting its importance for Kant, the faculty of imagination has been the subject of a number of dedicated studies, including but by no means limited to Hermann Mörchen’s Die Einbildungskraft bei Kant (1930, 2nd ed. 1970, Rudolf Makkreel’s Imagination and Interpretation in Kant (1990), and Jane Kneller’s Kant and the Power of Imagination (2007). As these scholars recognize, the imagination, while deeply mysterious, is also an essential contributor to our cognition and aesthetic experience and, as a result, any effort to frame Kant’s theory of imagination requires that we marshal all of the resources at our disposal. Of particular relevance are Kant’s discussions of the faculty in the context of his treatments, largely in his lectures, of empirical psychology and anthropology. Indeed, while the psychological tenor of Kant’s discussion of the imagination, and related doctrine of synthesis, in the first Critique has provoked the ire of a number of critics, most notoriously Strawson, scholars have shown that these sources can be productively used to flesh out Kant’s theory of imagination, particularly its range of functions in Kant’s aesthetics. So, in his now-classic study, Makkreel outlines Kant’s pre-Critical theory of the imagination as presented in the L1 (“Pölitz”) metaphysics lecture notes dating from the second-half of the 1770s, and draws attention in his discussion to the various functions that Kant assigns to the imaginative faculty in addition to those narrowly-cognitive functions he assigns to it in the first Critique (and which anticipate Kant’s later treatment of that faculty in the third Critique). Since Makkreel’s treatment, however, and apparently unattended to by more recent discussions, These include Wunsch 2007, Matherne 2015, and Horstmann 2018. a wealth of material relevant to Kant’s account of the imaginative faculty has been published in the form of the lectures on anthropology in volume 25 of the Akademie Ausgabe. These student lecture notes allow us to fill out many of the details of Kant’s account and to arrive at a fairly systematic classification of the various sub-faculties of the imagination; moreover, as I will argue, while the discussion recorded in the notes also sheds light on an important but under-appreciated change of emphasis in Kant’s account of the imagination from the pre-Critical to Critical period, it also serves to highlight important continuities, and shows how key arguments in the Critique are constructed, at least in part, in service of his previous anthropological claims regarding the imagination’s role in cognition. To this end, the following will be divided into three sections. In the first, I will consider Kant’s treatment of the imaginative faculty in the student notes to Kant’s lectures on anthropology in the 1770s, with the aim of working up a more-or-less comprehensive taxonomy of its various sub-faculties. In the second, I turn to Kant’s account of the activity of the imagination, particularly in accordance with the law of association, in the theory of cognition presented in the notes, and will show that Kant, apparently in spite of Hume, takes the result of this activity as the basis for causal cognition. In the third and final section, I contend that Kant’s treatment of affinity in the A edition Deduction is animated precisely by his concern to shore up his previous account of causal cognition against Hume’s sceptical challenge. In the end, we will see that in addition to their value in expanding our understanding of the functions of the imagination relating to aesthetic experience, Kant’s anthropological lectures provide an indispensable (if frequently overlooked) resource for appreciating the imagination’s complex role in his properly Critical account of cognition. 1. Towards the Einbildungskraft: Kant’s Taxonomy of the Imagination Throughout most of the pre-Critical period, Kant’s discussion of the imagination are couched in terms of an analysis of the bildende Kraft (formative power) or the Bildungsvermögen (formative faculty), In what follows, I have preferred to make use of the German original to more easily highlight the etymological connection between the various sub-faculties of the Bildungsvermögen (and for reasons relating to the use of the term ‘imagination’ which will become clear in what follows). Through his career, Kant uses ‘bildende Kraft’ designate a faculty of both epistemological and biological significance; Ina Goy (2012) tracks Kant’s various accounts (albeit without reference to the discussion in the Anthropologie lectures). terms which Kant appears to use indistinguishably (but I will use the latter). Makkreel (1990: 13) suggests a distinction between the terms (where the bildende Kraft designates the use of the Bildungsvermögen with respect to “given objects”; cf. Refl 330 15:130); however, Kant also uses the former to designate the faculty as a whole (see, for instance, Met-L1 28:230 and 235). The Bildungsvermögen is distinguished from the faculty of sensation (Empfindung) or the faculty of the senses (Vermögen der Sinne) (Anth-Coll 25:45, Anth-Fried 25:511 Met-L1 28:230) in terms of their products—with the senses yielding sensations or impressions and the Bildungsvermögen producing images (Bilder) (see especially Anth-Coll 25:45, Anth-Parow 25:303)—but also in terms of the activity of the Bildungsvermögen that is required for the generation of its product, in contrast with the passivity of the senses (see Anth-Fried 25:511, Met-L1 28:230, R 314 HN 15:124). In spite of this distinction, and crucially, Kant contends that both the faculty of sensation and the Bildungsvermögen belong to the faculty of sensibility (Sinnlichkeit); thus, the discussion of the imagination in the Anthropologie-Collins is presented in the context of the “Theorie der Sinnlichkeit” (Anth-Coll 25:44); in Anthropologie-Parow Kant is recorded as identifying the product of sensibility as a whole as “Bilder,” a description that obviously includes the Bildungsvermögen (Anth-Parow 25:303); and in the Met-L1 notes it is claimed that the facultas fingendi (which is identified with the bildende Kraft), among other faculties, proceeds from the division of sensibility (28:230). Andrew Stephenson has recently emphasized that Kant continues to classify the imagination as a faculty of sensibility in the Critical-period lectures and notes; see Stephenson (2015:496–8; and cf. Mörchen 1970:15, and Aquila 1987:80 and n4 for a similar emphasis). See also Met-L1 28:231, where the notes read “Diese bildende Kraft, die zur Sinnlichkeit gehört, ist unterschieden von der denkenden Kraft, die zum Verstande gehört.” However, Kant’s reasoning behind including the Bildungsvermögen within sensibility, at least in the pre-Critical notes and lectures, is not usually that the products of both count as intuitions See, however, Anth-Parow 25:305. but rather that in all of its functions the Bildungsvermögen is limited to working with the material supplied by the senses and in its activity it can only alter the form of this material (Anth-Coll 25:44, Anth-Fried 25:511), which is to say that according to the criterion introduced in the Inaugural Dissertation, the products of the Bildungsvermögen remain “sensible cognition” (cf. ID 2:394 and Met-L1 28:230, where Kant identifies sensibility as “das sinnliche Erkenntnißvermögen”). Focusing now on the divisions within the Bildungsvermögen itself, Kant sometimes (as in the Anthropologie-Parow or the Met-L1 notes) proceeds directly to a discussion of the various specific forms of Bildung or image-formation (enumerated by Makkreel in 1990:14). However, there are a couple of relevant distinctions that Kant elsewhere introduces before we arrive at this last set of (cognitively crucial) faculties. First, within the Bildungsvermögen itself, Kant distinguishes between two groups of faculties in terms of whether they involve the capacity to produce images from themselves “independent of the actuality of objects, where the images are not borrowed from experience” (Met-L1 28:237; cf. also Anth-Coll 25:77, Anth-Parow 25:305 and Refl 316 15:125). Yet, since Kant has just emphasized that the Bildungsvermögen is incapable of generating an entirely new sensation but is bound to the material provided by the senses, this must be qualified, as Kant later suggests, in terms of a capacity to bring forth images from itself that were not previously present in the same way to the senses (cf. Menschenkunde 25:945). This yields two groups of faculties within the Bildungsvermögen, which, following Kant’s later convention, we might refer to as the reproductive and productive, where the former are limited to reproducing representations, that is, bringing them forth as they were originally present to the senses, and the latter brings forth representations using the material of the senses but where these resulting representations are different (in form) than how they originally appeared to the senses. Here see again MK 25:945 (though contrast Anth-Bus 25:1456), and as in Met-L1 (28:237), this distinction is likewise behind Kant’s distinction between Nachbildung and Einbildung or Dichtung at Anth-Coll 25:87 and 95, and Anth-Parow 25:305. Kant sometimes refers to the entire “productive” faculty of the Bildungsvermögen as the Einbildungsvermögen (as at Anth-Coll 25:87, and Met-L1 28:237). A further distinction is applied to both the reproductive and the productive faculties of the Bildungsvermögen, namely, whether these capacities are exercised “voluntarily [willkührlich]” or “involuntarily [unwillkührlich].” This latter distinction, when applied to the reproductive Bildungsvermögen yields the faculties of memory (Gedächtniß), as the faculty of voluntary (or also conscious) reproduction of representations (cf. Anth-Fried 25:521 Menschenkunde 25:974, Anth-Mrong 25:1273), and its involuntary exercise, which Kant does not name at this point. It is this latter, involuntary reproductive capacity of the Bildungsvermögen that is further divided into the three capacities of Abbildung (direct image-formation), Nachbildung (imitative image-formation) also referred to simply as Imagination, and Vorbildung (anticipatory image-formation). The translations here are adapted from those of Makkreel (1990:13–14). As documented by Makkreel, Kant also distinguishes Einbildung, Gegenbildung, and Ausbildung, which I have not included here. Einbildung relates to the Einbildungskraft/Vermögen der Einbildung and so will be discussed later; Gegenbildung (or the Vermögen der Characteristik [Met-L1 28:237]or facultas signandi [Met-L1 28:230]) seems to be the Kantian equivalent of Baumgarten’s facultas characteristica (Baumgarten, Metaphysica §619) and not proper to the imaginative faculty; and Ausbildung (facultas componendi [Met-L1 28:230] or also perficiendi [Refl 313a 15:123,] and Anth-Fried 25:512) does not appear to find a clear precedent as a cognitive function in Baumgarten or elsewhere (though arguably it does belong among the reproductive power of the imaginative faculty—on this, see section 2 below). These three capacities are differentiated according to the temporal modalities of the representations they generate: Abbildungen are representations of (some object in) the present time, Nachbildungen are representations of the past, and Vorbildungen amount to representations of a future time (Anth-Coll 25:45, Anth-Parow 25:303–4, Met-L1 28:235). Given this, we might question the inclusion of Abbildung (and perhaps also Vorbildung) among the capacities of the reproductive Bildungsvermögen; however, Kant is clear that Abbildung does not merely involve the formation of a representation on the basis of a given sensation but requires that the sensible manifold be run-through, a process that involves the reproductive activity of the mind (Anth-Parow 25:303, Met-L1 28:235). It is for this reason that Kant indicates that the capacity for Abbildung is the “fundament” of the others, according to which the representations of Nachbildung and Vorbilden are made (Refl 315 15:125), namely, inasmuch as it is the Abbildungen resulting from this process that are reproduced or anticipated to recur later. I thus agree with Makkreel, albeit for different reasons, that these three capacities of Abbildung, Nachbildung, and Vorbildung are not intended to map on to Kant’s Critical doctrine of a three-fold synthesis (see Makkreel 1990:25). On this, compare Mörchen (1970:33). Kant similarly divides the productive faculty of the Bildungsvermögen into its voluntary and involuntary exercise. Kant identifies the involuntary capacity of bringing forth representations as Phantasie (Anth-Coll 25:87, Anth-Parow 25:314), which capacity differs from memory in its involuntary character and from Nachbildung or Imagination in that the representations differ in form from their original sensible presentations (Met-L1 28:237). As a creative but involuntary power, Phantasie can be symptomatic of a disordered mind, either insofar as it is without a rule (Regellos), and so in conflict with the understanding as in dreaming, or utterly unrestrained (Zügellos), a condition Kant associates with great genius but also hypochondria (Anth-Fried 25:514–15, Anth-Pillau 25:752–3, 764–5, Anth-Mrong 25:1261). By contrast, the voluntary productive exercise of the Bildungsvermögen is identified as the Einbildungskraft proper (cf. Anth-Coll 25:95, Anth-Parow 25:305). See also Refl 314 15:124, Refl 334 15:132, Refl 337 15:133, Anth-Pillau 25:758, Anth-Mrong 25:1258, and Anth-Bus 25:1462. Kant takes this faculty to be responsible for poetic creation (Dichten), on account of which he also frequently identifies it as the Dichtungsvermögen (Anth-Parow 25:321, Met-L1 28:237, Menschenkunde 25:981, Anth-Mrong 25:1277, Anth-Bus 25:1464), and also contends that it is the primary source of all invention (Erfindung), whether poetic, pertaining to new cognitions or even pragmatic, as in the Einbildung of a still-to-be constructed house (Anth-Coll 25:97, Anth-Fried 25:524, Met-L1 28:143, 237). Taking all this together, we can provide the following by way of a systematic division of the Bildungsvermögen according to Kant’s treatment of that faculty in his lectures on anthropology (and empirical psychology in the metaphysics notes): Significantly, Kant adheres to this division in its essentials in the writings and lecture notes of the Critical period. Thus, in addition to formally introducing the distinction between ‘reproductive’ and ‘productive’ faculties, he continues to distinguish between voluntary and involuntary functions of the reproductive faculty, referring to the former as memory (cf. Anth-Parow 25:314, Menschenkunde 25:974, Anth-Mrong 25:1273, Anth-Bus 25:1462, Anth 7:182) and the latter, at least at one point, as the “merely reproductive” faculty (Anth 7:182 See also Anth-Mrong 25:1257:“das reproductiv ist unwillkührlich,” and Refl 373 15:146-7.); similarly, Kant distinguishes between involuntary and voluntary productive faculties, typically identifying the former as Phantasie (Anth-Mrong 25:1258, Anth 7:167, 175, 180) and the latter as the Dichtungsvermögen or sometimes as the Imagination (Anth-Mrong 25:1258, 1277, Anth-Bus 25:1464). The most obvious difference, of course, is that Kant no longer presents this division in terms of an analysis of the Bildungsvermögen but rather of the Einbildungskraft, a change that makes its first appearance, in the materials we have available, with the lectures recorded in the Anthropologie-Pillau from the Winter Semester of 1777/78 (cf. Anth-Pillau 25:750f). Nor should this be taken merely as a change in terminology for Kant, as the Einbildungskraft is not defined narrowly (as the Bildungsvermögen was) in terms of a faculty for the spontaneous generation of images but rather more generally as the faculty for the “representation of things that are not present” (cf. Anth-Pillau 25:750), which likewise distinguishes it from Kant’s previous definition of the Einbildungskraft, that is, insofar as it was identified with the Dichtungsvermögen. There is, then, an important dis-continuity in Kant’s account of the Einbildungskraft between approximately 1775/76 (the time of the Anthropologie-Friedlander, where Kant’s presentation proceeds in accordance with the above—cf. Anth-Fried 25:511) and 1777/78 (when Kant introduces the Einbildungskraft in its now-familiar characterization). While the extent of this change has gone largely unnoticed in recent Kant scholarship, Matherne, for instance, claims that “[t]his formative power [in the Met-L1 notes] is what he will later call the ‘imagination’ [Einbildungskraft] in the first Critique” (2015:747). it is potentially rather significant for understanding Kant’s development in the 1770s. One case that bears mentioning concerns the contested and complicated question of the dating of the Met-L1 notes (for a discussion, see Ameriks and Naragon 1997:xxx-xxiii). Limiting ourselves to the consideration of the Bildunsvermögen presented at 28:235-8, Kant’s discussion bears a much closer comparison with the account presented in Anthropologie-Friedlander (and the earlier sets of lectures) rather than that given in Anthropologie-Pillau, with which it is taken (by for instance Wolfgang Carl—1989:117-18) to be contemporaneous or later. In any case, insofar as this does not amount to a mere change in terminology or reversion to a definition of imagination that reflects that given in Baumgarten’s source text, we might assume that the change has a philosophical motivation, and one reason that suggests itself is that it reflects a change in the scope of functions Kant assigns to the imaginative faculty. While the account of the Bildungsvermögen would have sufficed for the range of tasks relating to the generation, manipulation, and augmentation of images insofar as they figured as a mediating step between impression and conceptual cognition (cf. Anth-Coll 25:76–7), Kant would soon recognize that this performance with respect to images was itself conditioned by a prior activity on the part of the mind, in accordance with which it becomes possible to bring sensible impressions to images in the first place. One way in which this need is expressed is in terms of the generation of the pure representations of space and time, and the problem of accounting for precisely how this is possible would have been especially apparent to Kant (especially after Tetens’ critical discussion of this topic in his Philosophische Versuche of 1777). For Tetens’ identification of this problem, and his attempt to resolve it, on Kant’s behalf, by means of his account of the Dichtkraft, see Dyck (2016, and forthcoming [b]). Of course, that Tetens’ Versuche was read carefully by Kant and was published before the lectures in the Anthropologie-Pillau (a review of the text by J. G. H. Feder is in the August 16th issue of the Göttingische gelehrte Anzeigen in 1777), is consistent with the apparent changes to the theory of the imagination recorded in those lecture notes. As a result, Kant’s account of the imagination could no longer be limited to the mere reproduction of images and their use in poetic invention, but would need to be widened to include its activity in its novel transcendental function. So, while the above taxonomy can still be allowed to hold for (what will come to be called) the empirical function of the imagination, which is limited to the manipulation and production of images, On this see also Longuenesse (1999:206). this empirical use is now conditioned by a transcendental function that is not so limited. See also Makkreel (1990:25). This change is likewise reflected in the contrast between Kant’s early claims that the Bildungsvermögen is the “Grund” of the productive use of the imagination in its function as the Dichtungsvermögen (cf. Anth-Coll 25:58, Anth-Parow 25:305) and his later claim that the empirical use of the imagination is conditioned by its transcendental (and productive) use (cf. 23:18). Of course, Kant does at one point, though problematically, refer to “the pure image [das reine Bild]” of space and time (A142/B182). 2. The Association of Ideas in the Anthropology Lectures Whatever the significance of Kant’s later turn to the Einbildungskraft, the foregoing does draw attention to the ways in which Kant takes the imagination to contribute to cognition distinct from its image-producing capacity. Foremost among these, Kant’s lectures contain frequent discussion of the lex associationis, understood more broadly (and traditionally) as the law that governs the reproduction of ideas or representations, which Kant contends makes a key contribution to cognition, and especially (though this is frequently overlooked) to that of causal connection. For instance, while Watkins and Willaschek’s recent treatment (2017) of Kant’s theory of cognition does not pretend to be exhaustive, it nonetheless does not take into account the imagination’s associative activity, nor how it might be involved in (causal) cognition. Kant discerns this law at the ground of the activity of Nachbildung, according to which representations are reproduced as a result of being associated with a presently-occurring representation. As the Met-L1 notes read, “I reproduce representations of the past through association, according to which a representation draws forth another because it was accompanying it” (28:236), and in the Friedländer anthropology notes Kant contends that “the faculty of Nachbildung has its law [since] it follows the law of association of representations” (25:512). However, the law of association does not only govern the involuntary reproduction of ideas or representations, as Kant suggests that it is the law of the entire reproductive activity of the imaginative faculty, and so also of Gedächtniß. Thus, Kant also identifies it simply as “the law of reproduction [Das Gesetz der Reproduction]” (Anth-Pillau 25:752; cf. also Anth-Coll 25:87), and elsewhere he is explicit that the law governs the voluntary reproduction involved in memory (cf. Anth-Mrong 25:1272–3), and accordingly he repeatedly recommends repeating specific associations (of words with one another, for instance) as a way of improving memory (Anth-Coll 25:88, 89–90, Anth-Parow 315–17, Anth-Fried 25:522). As a number of commentators have noted, however, Kant takes the law of association to have wider importance for a variety of activities on the part of the imaginative faculty. So, Mörchen (1970:33–4) and Makkreel (1990:17) have pointed out that Kant also discerns this law at the basis of the operation of Vorbildung, insofar as associative connections among past representations yield, upon being given a similar representation or series of representations, an anticipation of such a representation in the future (Met-L1 28:236). For this reason, Kant sometimes refers to the “laws of the imagination” (as at Met-L1 28:236, but also Anth-Coll 25:87), and distinguishes specifically between the lex aßociationis idearum sociarum (or law of the association of accompanying ideas) and the lex expectationis casuum similium (or law of the expectation of similar cases) (Refl 225 15:86; cf. also Met-Mrong 28:883, Met-L2 28:585, Met-Dohna 28:674, and Anth 7:182), both of which are understood as falling under a generic law of association. On the latter law, see also Refl 336 15:150-1 and Anth-Fried 25:533. Further, Kant seems to conceive of the law of association as operating pervasively throughout the activities of the imaginative faculty, including influencing its productive functions. So, Kant’s explanation of the capacity of Ausbildung, insofar as it concerns the drive (Trieb) to represent a whole on the basis of the representation of the parts, relates directly to Baumgarten’s characterization of the law of association (in the context of his discussion of Phantasia): “percepta idea partiali recurrit eius totalis” (Metaphysica §561; cf. Met-L1 28:237). Arguably, the capacity of Gegenbildung, as a faculty of characteristic, also relies on association in order to establish a connection between the artificial signs that it produces and the item signified, a point likewise underscored by Baumgarten (Metaphysica §620; cf. Met-L1 28:238). Even the productive activities of the imaginative faculty, particularly phantasy, can involve the law of association, as Kant notes that in dream, the product of Phantasie, “the images hang together as they are associated in the waking state” (Anth-Coll 25:101), and in a Reflexion Kant suggests that when we play with ideas in accordance with the Einbildungskraft, our activity remains governed by the “law of association” (Refl 338 15:133). It is, in any case, insofar as it is the law of the reproductive function of the imaginative faculty that Kant holds that this law is involved in the generation of cognition, and particularly cognition of causal relations. The importance of the law of association in this respect is underscored by Kant in the Friedländer notes, where it is asserted that the law “is a considerable part of the human mind, and much depends upon the association of representations” (25:512). Kant is fairly consistent throughout the notes in distinguishing between three “grounds” or “conditions” that can underlie associations or on which associations “rest”, namely, similarity (Einträchtigkeit or also Ähnlichkeit), contiguity (Nachbarschaft), and affinity (Verwandtschaft ‘Verwandtschaft’ might be translated as ‘relatedness’ or even ‘relation’ (as in Jankowiak and Watkins 2014:70), but Kant himself suggests affinitas as the Latinate equivalent at Anth 7:174 and 177, for which reason I have opted to render it as ‘affinity’; this also has the advantage of making clear the connection between the presentation in the anthropology lectures and that in the A edition of the first Critique.). Similarity can ground associations insofar as there is some feature common to both representations, which similarity Kant notes is often reflected in the resembling names we give to similar things (Anth-Pillau 25:752). Kant sometimes considers similarity as a species of affinity, distinguished from causation, as at Anth-Fried 25:513 or Menschenkunde 25:947; however, I will suggest below that there is good reason to distinguish it from this (as indeed Kant more often does). Associations grounded on contiguity involve either the (repeated) succession of representations in time, which Kant also refers to as accompaniment (Begleitung or Beygesellung—cf. Anth-Fried 25:512, 15:129 [R 326]), or their occurrence in relation to a specific space, which Kant sometimes refers to as Nachbarschaft (Anth-Fried 25:513). The former, Kant notes, is the ground of its own sort of association, but inasmuch as it involves repeated temporal succession, it is in some way at the ground of all associations (Anth-Coll 25:89, Anth-Parow 25:315, Anth-Fried 25:512). Lastly, associations can also be grounded on affinity or a relation obtaining between two representations, which Kant characterizes primarily in terms of “stemming from” or (non-logical) “derivation” (Abstammung). Such a relation obtains, as Kant notes, “insofar as [representations] proceed from a single ground” (Anth-Fried 25:513) and the primary example of this is the relation between an effect, or effects, and a cause (Anth-Fried 25:513, Anth-Pillau 25:752). Significantly, in addition to distinguishing between these three grounds or conditions of association, Kant also distinguishes between the resulting associations in terms of their relevance for cognition. In the Anthropologie-Collins, for instance, we find a distinction between associatio bruta, which pertains to associations grounded on contiguity, associatio reflexa, which pertains to those grounded on similarity, and associatio intellectualis, which pertains to associations grounded on affinity (Anth-Coll 25:88, 90). The Anthropologie-Parow notes mentions “leges phantasiae brutae” (Anth-Parow 25:318), which is likely a reference to the first of these types of association. With the Anthropologie-Friedländer, Kant refines this somewhat, distinguishing only between “Associationen der Sinnlichkeit” and “Associationen des Verstandes,” identifying the former with associations grounded on (spatial or temporal) contiguity, and the latter with associations grounded on affinity (Anth-Fried 25:513). The latter version of this distinction effectively brings out an important difference between the resulting associations, namely, that associations grounded (solely) on spatial or temporal contiguity relate merely to the conditions of sensibility (space and time), whereas associations grounded on affinity (also) relate to the conditions of the understanding, a broad distinction between faculties (and types of cognition) familiar since the Inaugural Dissertation. Kant lists ‘cause’ among the concepts of the pure understanding at De mundi 2:395 However, Kant seems to invest this distinction with additional significance inasmuch as he also takes it to amount to a distinction between associations between representations that cannot be reliably taken to hold of objects (those of sensibility) and associations that can (those of the understanding). Such a distinction also seems to be implied by Kant’s contrast between “associations of sensations” (that is, representations pertaining merely to the subject) and “associations of intuitions” (or pertaining to objects) (Anth-Parow 25:319). Associations that are merely based on spatial or temporal contiguity can be misleading, or can interfere with or augment our ordinary cognition, as Kant illustrates through the example of a former student who returns to the place of a stressful examination and feels a renewed sense of dread (Anth-Fried 25:513). Accordingly, Kant claims that such associations are not simply different than those pertaining to the understanding but are even “counter to [wiedersinnig]” the understanding, which is presumably to say that there is no affinity or connection that can be discerned to obtain at the ground of the associated representations (Anth-Fried 25:513). It is only with respect to associations of the understanding that the connection among representations is such that we infer an actual connection obtaining the objects of the representations so connected in the mind. Among the examples of such connection that Kant provides are: “when it rains and the sun shines, one quickly looks around to see if there is a rainbow” (Anth-Fried 25:513) and “when someone sees a thunderstorm arise when there are dark clouds, one supposes the same effect when after he sees such clouds again” (Anth-Pillau 25:752). In both cases, the association is conditioned by a causal connection among the phenomena observed and, as evidence of its objective purport, it (as opposed to the other associative connections) prompts us to infer a similar occurrence when the previously observed conditions obtain. Indeed, associations of the understanding are distinguished from those of sensibility insofar as the former, but not the latter, involve or establish a connection (Verknüpfung) between the representations themselves on the basis of repeated observation (or “constant accompaniment”). Thus, the notes read that “not all accompanying representations [begleitende Vorstellungen] are associated [vergesellschaftet]”; rather, it is only “what we often see accompanying each other [that] we take for connected through the imagination” (Anth-Coll 25:28; cf. also Anth-Coll 25:89, Anth-Parow 25:315). Insofar as it is also comprehended under affinity as a ground of association, associations involving similarity can also be objective ,though Kant does warn that there is a potential for error when we infer from a similarity among signs of representations to similarities in the objects designated by them (Anth-Parow 25:318). There are a number of striking features of Kant’s discussion of association, but for now we might just focus on the most obvious, namely, its comparison with Hume’s discussion of the cognition of causal connection. As should be evident, Kant seems to help himself to elements of Hume’s empiricistic account of causal cognition as he takes, for instance, our cognitions of causal connection to be based on associations effected by the imagination on the basis of repeatedly apprehended successions. And while Kant was undoubtedly familiar with Hume’s discussion in the Enquiry concerning Human Understanding at the time of the lectures, it would appear that his discussion recorded in the notes does not attempt to engage with Hume’s own (sceptical) conclusions regarding the rational foundations of our cognition of causal connections. On the contrary, Kant shows little hesitation in presuming that the connections effected by the mind are grounded in an actual (causal) connection among objects, and in dubbing associations grounded upon affinity associations of the understanding , Kant is clearly signalling his presumption, contra Hume, that such associations are in fact in agreement with the demands of the understanding and, as such, suitable for serving as the basis for causal judgments. It thus seems peculiar that Kant would help himself to these elements of Hume’s account without attention to the sceptical consequences that they introduce. And while it might thus be tempting to think that Kant abandons this account, with all its attendant difficulties and potential incompatibilities, with the first Critique, This suggestion is floated by Jankowiak and Watkins (2014:70). I will argue in the following section that Kant does in fact retain this account (at least in its essentials) but also seeks to exorcise the spectre that it threatens to introduce into his account of causal cognition. 3. Association and Affinity in the Critique While there can be no denying that Kant was sensitive to Hume’s challenge during his earliest anthropology lectures in the 1770s, according to his own later confession of the fact (cf. Prol 4:260), it is unlikely that he would have thought an anthropology class the appropriate forum for engaging with this issue (especially when his own final response to Hume had likely still to be worked out). That said, the A edition of the first Critique would prove the very place to address Hume’s challenge and to vindicate his contention that the associative connections effected by the imagination are grounded on the actual causal connection or affinity of the manifold and, thus, can serve as the basis for causal cognition involving association. Before considering Kant’s argument, we should note that much of Kant’s account of causal cognition as presented in the early student lecture notes on anthropology remains in place in the KrV. So, and unsurprisingly, Kant continues to identify the law of association as the law of the imaginative faculty, affirming its status as a (merely) “empirical law” (A100) or an “empirical rule” (A112). Insofar as this empirical law remains a property of the mind, Kant also identifies the “association of representations” as the “subjective and empirical ground of reproduction” (A121—my emphasis). Moreover, Kant continues to hold that, insofar as the associative connection effected by the imagination is to be suitable as a basis for causal cognition, it will be grounded upon the affinity of the manifold; thus Kant identifies (empirical) affinity as the “objective ground” of associative connection (A121—my emphasis) and likewise refers to it as the “ground of the possibility of the association of the manifold, insofar as it lies in the object” (A113). Compare as well Kant’s brief discussions of affinity in lectures from the Critical period (cf. Menschenkunde 25:946-7), as well as in the published Anthropology itself (7:176ff) One important novelty in Kant’s account in the KrV is that he now provides more detail concerning how, precisely, associative connections—what he had called an association of the understanding but now refers to as a “unity of association” (A121)—find their (objective) ground in the affinity of the manifold. As Kant now explains, the imagination’s activity in reproducing representations in accordance with the law of association depends upon the fact that occurrences take place with sufficient regularity to trigger the associative mechanism on the part of the mind. As Kant writes, This law of reproduction, however, presupposes that the appearances are actually subject to such a rule, and that in the manifold of their representations an accompaniment or succession takes place according to certain rules; for without that our empirical imagination would never get to do anything suitable to its capacity, and would thus remain hidden in the interior of the mind like a dead, and to us unknown faculty. (A100) As Kant famously illustrates using the example of cinnabar, were it the case that there were no regularity in its appearance (if it were sometimes red, at other times black, light at one point, heavy another), there would be no occasion for the imagination to effect associations between “heavy cinnabar” and “the colour red” (A101). This allows us to fill in some of the details of Kant’s account of ordinary causal cognition outlined in the anthropology lectures. Setting out from the assumption of the affinity of the manifold of objects (appearances), which is to say, the assumption that “they stand under constant laws and must belong under them” (A113), this supplies the objective ground for the imagination’s activity. The apprehension of a regular accompaniment or succession among appearances triggers the imagination to act in accordance with the law of association (the so-called “subjective” ground of its activity) to effect a connection among the representations. When the connection among appearances is sufficiently constant, the bond effected by the imagination comes to constitute something like a rule as Kant himself explains: It is, to be sure, a merely empirical law in accordance with which representations that have often followed or accompanied one another are finally associated with each other and thereby placed in a connection in accordance with which, even without the presence of the object, one of these representations brings about a transition of the mind to the other in accordance with a constant rule (A100) The “rule” involved in the transition of the mind effected by the imagination, as Kant goes on to point out, is grounded on a rule “that the appearances themselves are actually subject to” (A100; cf also A101: “a certain rule to which the appearances are already subjected”). With this transition among representations in accordance with a constant rule established, there is nothing preventing the resulting associative connection from being subsumed under the schema of causality, which Kant defines in terms of “the real upon which, whenever it is posited, something else always follows” and which therefore “consists in the succession of the manifold insofar as it is subject to a rule” (A144/B183). This sketch, of course, leaves out a number of the crucial details of Kant’s account, such as (to name only one) how we can distinguish in practice between associations that are suitably grounded and those (merely subjective) associations that are not (since they are not likely to be distinguished in terms of the feeling of necessity involved). Even so, on the basis of this outline we can see the specific challenge Kant would have taken Hume’s scepticism regarding the basis of our knowledge of causation to pose. In Kantian terms, Hume’s challenge to our certainty concerning the uniformity of nature amounts to a challenge to the (empirical) affinity of the manifold. Of course, Hume does not dispute that we do in fact assume a regularity among natural occurrences, but that appearances are subject to laws that govern them would be, at best, a contingent matter of fact. Kant, however, sets the bar for affinity considerably higher than Humean regularity if that affinity is to serve as the objective ground for the sorts of associations that might count as genuinely causal cognitions; as he writes, “[f]or even though we had the faculty for associating perceptions [if we lacked an objective ground] it would still remain in itself entirely undetermined and contingent whether they were also associable” (A121–2). For Kant, and presumably for Hume, accepting that the regularity of appearances cannot be known to hold a priori has the consequence that the resulting associative connections formed by the mind on the basis of observed (contingent) regularities would fall well short of what is required for causal cognition. By contrast, Kant insists that in order for such connections to constitute an appropriate objective ground of association, it does not suffice to show that the manifold is as a matter of fact subject to connection in accordance with a rule, but that it must be so subject; thus, he emphasizes that associations must be in accordance with a “constant” or a “certain [gewisse]” rule (A100, 101—my emphases). Yet, insofar as we cannot know a priori that the manifold is subject to connection in accordance with rules, that is, we cannot know that the manifold must be subject to such connection, then we have no warrant for taking the associative connections framed by the mind to have their ground in anything more than a mere contingent regularity and, thus, to fall short of what is required for a cognition of causal connection. Ultimately, then, in order to defend his account of causal cognition, and his claim that the associative connection effected among representations has its ground in a rule-governed connection among the manifold of appearances, Kant turns to demonstrating the possibility of “the thoroughgoing affinity of the appearances (by means of which they stand under constant laws and must belong under them)” (A113). In the ensuing paragraph, Kant proceeds to summarize his argument which I will divide into three steps and briefly discuss. Note that I differ from Allison (2015:267ff) in preferring Kant’s exposition of the argument at A113-4, though I take it that both are concerned with the same argument (and so I draw on material from the later exposition to clarify the former). Kant begins, as might be expected, by noting that for every manifold of representations that might be something for me it is necessarily possible for us to be conscious of it; inasmuch as Kant takes this consciousness to involve thinking the identity of the subject with respect to a given manifold of representations: All possible appearances belong, as representations, to the whole possible self-consciousness. But from this, as a transcendental representation, numerical identity is inseparable, and certain a priori, because nothing can come into cognition except by means of this original apperception. The connection here is later clarified by Kant, since “only because I ascribe all perceptions to one consciousness (of original apperception) can I say of all perceptions that I am conscious of them” (A122). Kant then proceeds to consider what this ascription to a single, identical subject implies with respect to the manifold, and he contends that it necessitates a synthesis of that manifold (since we think the relation of the manifold to the same subject) but, insofar as this is just to introduce a unity into the manifold (the unity of the identical subject), it also implies that for any manifold of representations it must be possible for it to be brought into a synthetic unity (where, because this unity stems from the identity of apperception, this amounts to an a priori condition): Now since this identity must necessarily enter into the synthesis of all the manifold of appearances insofar as they are to become empirical cognition, the appearances are thus subject to a priori conditions with which their synthesis (of apprehension) must be in thoroughgoing accord. That Kant mentions “a priori conditions” here is, of course, a reference to the categories which, as functions of the understanding are just ways in which a given manifold of representation can be brought to the unity of apperception, that is, different ways of thinking the manifold in relation to the identity of consciousness (cf. A119). This is to say that the manifold of representations, insofar as it belongs to a possible self-consciousness, is necessarily subject to connection in light of the categories (including causality), a conclusion that Kant here glosses in terms of the manifold being subject to connection by laws: Now, however, the representation of a universal condition in accordance with which a certain manifold (of whatever kind) can be posited is called a rule, and, if it must be so posited, a law. The conclusion, then, is that the manifold is necessarily subject, in virtue of belonging to self-consciousness, to connection in accordance with laws, and it is this fact that Kant now dubs transcendental affinity (A114). Despite Kant’s insistence that the supposition of transcendental affinity has empirical affinity as its consequence (A114), it is not immediately evident how this provides anything more than a (further) subjective ground for the imagination’s activity rather than accounting for the possibility of its sought-for objective ground. Indeed, Kant seems to court this objection when he claims that this affinity cannot be encountered “anywhere except in the principle of the unity of apperception” (A122). What in any case enables transcendental affinity to fulfill this demand is Kant’s transcendental idealism, and specifically, his reconsideration of the notion of objectivity (cf. A104ff), in accordance with which what it means to think the relation of a manifold of representations to an object is just to show the necessity of unity for that manifold, “i.e., they must have that unity that constitutes the concept of the object” (A104). For more detail on this line of argument, see Dyck forthcoming [a] (especially section 2). Transcendental affinity, however amounts to the claim that for any manifold of representations of which we can become conscious, it must be possible for that manifold to be united in accordance with the categories, or as Kant elsewhere puts it, any apprehension of the manifold must be “in agreement with the unity of apperception, which would be impossible without synthetic unity in their connection” (A122). Transcendental affinity thus implies the necessity of (categorial) unity for any manifold of representations of which we are conscious, and this suffices, for the transcendental idealist, for the thought of the relation of those representations to objects. Accordingly, for Kant, the transcendental idealist, at least, is licensed in drawing the conclusion that “[a]ll appearances therefore stand in a thoroughgoing connection according to necessary laws” (A114—my emphasis), that is, that the objective ground of association, or empirical affinity of the manifold, is rendered conceivable through transcendental affinity. Having thus shown how the empirical affinity of the manifold of appearances is possible, Kant has headed-off the major concern relating to the prominent role he assigns to the imagination and its associative function in his account of causal cognition. In accordance with the doctrine of transcendental affinity (and transcendental idealism), Kant can now take associations on the part of the imagination to find their objective ground in the fact that appearances are (necessarily) subject to laws, and thus can presume that the connections among our representations established by the imagination will track genuine causal connections. This is not, of course, to say that Kant takes himself to have proven thereby that all of the associations on the part of the imagination are so grounded; far from it, Kant has only provided us with a license to assume that at least some of our associations are grounded on affinity, and while he does not proceed to offer us a way to distinguish between associations that do find such a ground and those that do not, one presumes that sifting out associations based on mere regular accompaniment from those that rest on a causal connection would (and does) involve rigorous empirical investigation and experimentation. Indeed, it is likely that Kant’s later account of the origin of (objectively valid) judgments of experience from judgments of perception, which express an at first merely subjectively valid connection of representations (cf. Prol 4:298), is relevant here, as the initial associative connections are refined into associations of the understanding (and distinguished from those associations expressed in “mere” judgments of perception) and thereby rendered into a form subsumable under the category of causality, though space does not remain to consider this proposal in detail here. For a similar comparison of associational connections to judgments of perception, see Aquila 1987:103ff, and for critical discussion of this proposal, see Wunsch 2007:227ff. Irrespective of the many details to be filled in, what cannot be denied in the end is that the role Kant assigns the role of the imagination and association in the lectures on anthropology is not a pre-Critical aberration or dogmatic vestige but constitutes an important part of his considered, Critical account of empirical cognition, a fact confirmed by his efforts to exorcise the spectre that has attended the involvement of the imagination and its associative activity since Hume. Indeed, the same point is underlined by the fact that Kant thinks that the foregoing also makes sense of the basis of Hume’s error: due to his transcendental realist perspective, Hume is incapable of accounting for how the connection of objects according to constant laws is possible. This means that Hume cannot account for the objective ground of the imagination’s activity, namely, in a unity ultimately supplied by the understanding, but can only admit a subjective ground, namely, the law of association, with the result being that genuine causal cognition is impossible since the result of the mere imaginative synthesis cannot be taken to have an objective ground in the affinity of the manifold. As Kant summarizes Hume’s error much later in the KrV: “he made a principle of affinity, which has its seat in the understanding and which asserts necessary connection, into a rule of association, which is found merely in the imitative imagination [nachbildenden Einbildungskraft] and which can present only contingent combinations, not objective ones at all” (A766–7/B794–5). Hume thus did not err in taking associations on the part of the imagination to figure prominently in the cognition of events, and to that extent, it is no surprise that Kant’s account parallels Hume in the lectures on anthropology; but insofar as this relies for its possibility on a transcendental ground that is (inevitably) overlooked by Hume, it is important for Kant to supplement this account with a demonstration of the possibility of affinity in the Critique. Contrast Allison (1972:204 and 2015:267), who takes Kant to offer a “reductio ad absurdum of the Humean position” (specifically, the denial of affinity). On the view presented here, Kant’s argument for affinity does not show Hume’s position to be impossible (or to entail a denial of the bare associability of representations), but rather is limited to explaining Hume’s erroneous denial of causal cognition as the inevitable result of his transcendental realistic perspective; I take it that Hume’s (and indeed the empiricistic) account of the mechanism of association is nonetheless left in tact. We can see, then, that even in Kant’s Critique the positive account of the imagination’s role in our cognition of causal connection as it was elaborated in its essentials in the lectures on anthropology remains in place. Kant holds to this account, well aware of the core that it shares with empiricistic accounts of such cognition generally, but actively seeks to dispel the spectre which has haunted it since Hume. Significantly, this reserves an important role in Kant’s mature theory of cognition for the imagination, and its activity in accordance with the law of the association of ideas, which is not reducible to its image-generating capacity. More broadly, we can also see that Kant’s anthropology lectures remain a vital supplement to his mature account of cognition in general and to his discussion of the imagination in particular. Bibliography Allison, Henry E., 2015, Kant’s Transcendental Deduction: An Analytical-Historical Commentary, Oxford: Oxford University Press. ———, 1974, “Transcendental Affinity—Kant's Answer to Hume,” in: L.W. Beck (ed.) Proceedings of the Third International Kant Congress (1970), Boston: Reidel, 303–11. Ameriks Karl and Naragon, Steve, “Translators Introduction” in: Ameriks and Naragon (trans. and eds.), Lectures on Metaphysics, The Cambridge Edition of the Works of Immanuel Kant, Cambridge, Cambridge University Press, xiii–xliv. Aquila, Richard E., 1987, “Matter, Form, and Imaginative Association in Sensory Intuition,” in: Bernard den Ouden and Marcia Moen (eds.), New Essays on Kant, New York: Peter Lang, 73109. Baumgarten, Alexander Gottlieb, 1757, Metaphysica, reprinted in: Kants gesammelte Schriften, vols. XV & XVII. Carl, Wolfgang, 1989, Der schweigende Kant: Die Entwurfe zu einer Deduktion der Kategorien von 1781, Abhandlungen der Akademie der Wissenschaften zu Gottingen, Philologisch-Historische Klasse, Dritte Folge Series 182, Göttingen: Vandenhoeck and Ruprecht. Dyck, Corey W., forthcoming [b], “Tetens as a Reader of Kant’s Inaugural Dissertation,” for Proceedings of the XII International Kant Kongress, Berlin: De Gruyter. ———, forthcoming [a], “The Proof Structure of Kant’s A-Edition Objective Deduction” in: Kants transzendentale Deduktion der Kategorien: Neue Interpretationen / Kant's Transcendental Deduction of the Categories: New Interpretations, G. Motta and D. Schulting (eds.), Berlin: DeGruyter. ———, 2016, “Spontaneity before the Critical Turn: Crusius, the Pre-Critical Kant, and Tetens on the Spontaneity of the Mind,” Journal of the History of Philosophy, vol. 54.4, 625–48. Goy, Ina, 2012, “Kant on Formative Power” Lebenswelt: Aesthetics and Philosophy of Experience, 26–49. Horstmann, Rolf-Peter, 2018, Kant’s Power of Imagination, Cambridge: Cambridge University Press. Jankowiak, Tim and Watkins, Eric, 2014, “Meat on the Bones: Kant's Account of Cognition in the Anthropology Lectures,” in Alix Cohen (ed.), Kant's Lectures on Anthropology: A Critical Guide. Cambridge University Press, 57–75. Kneller, Jane, 2007, Kant and the Power of Imagination, Cambridge: Cambridge University Press. Longuenesse, Béatrice, 2001, Kant and the Capacity to JudgeSensibility and Discursivity in the Transcendental Analytic of the Critique of Pure Reason, trans. C. Wolfe, Princeton: Princeton University Press. Makkreel, Rudolf A., 1990, Imagination and Interpretation in Kant: The Hermeneutical Import of the Critique of Judgment, Chicago: University of Chicago Press. Matherne, Samantha, 2015, “Images and Kant’s Theory of Perception,” Ergo: An Open Access Journal of Philosophy (2) 737–77. Mörchen, Hermann, 1970 Die Einbildungskraft bei Kant, 2nd ed., Tübingen: Max Niemeyer Verlag. Tetens, Johann Nicolas, 1777, Philosophische Versuche über die menschliche Natur und ihre Entwicklung, 2 vols., Leipzig: M.G. Weidmanns Erben und Reich; reprinted, Hildesheim: Olms, 1979. Watkins, Eric and Willaschek, Marcus, 2017, “Kant’s Account of Cognition,” Journal of the History of Philosophy (55), 83–112. Wunsch, Matthias, 2007, Einbildungskraft und Erfahrung bei Kant, Berlin: DeGruyter.