It should perhaps come as no surprise that Kant’s account of imagination, ‘a blind though indispensable function of the soul … of which we are seldom even conscious’, should involve some mystery (A78/B103). Here’s one mystery: what happens to imagination between the first and second editions of the first Critique? In the first edition, Kant treats imagination as a ‘fundamental faculty’ distinct from sensibility and understanding (A124). However, in the second edition, Kant seems to collapse imagination into understanding: ‘It is one and the same spontaneity, that, there under the name of imagination and here under the name of understanding, brings combination into the manifold of intuition’ (B162 n.). To compound the mystery, in the third Critique imagination emerges once again as an independent faculty, capable of freely engaging with understanding in judgements of the beautiful and in artistic genius.

Yet, even if one concedes that Kant ultimately regards imagination as an independent faculty, one might wonder whether Kant needs a theory of imagination. This question has been pressed particularly within the context of Kant’s theory of cognition, since it can seem as if Kant has the resources to account for cognition by means of sensibility and understanding alone. Here, then, is another mystery: why does Kant appeal to imagination at all in his theory of cognition?

In Kant’s Power of Imagination, Rolf-Peter Horstmann offers a reconstruction of Kant’s account of imagination that promises to solve both mysteries. Horstmann contends that the reversals in Kant’s account of imagination are merely apparent and that a careful look at the first and third Critiques reveals imagination to be an autonomous faculty that plays a persistent and necessary role in Kant’s theory of cognition.

In section 1, Horstmann focuses on the first Critique and argues that in the Transcendental Deduction Kant attributes to imagination an essential role in cognition. More specifically, Horstmann maintains that imagination has a ‘transcendental function’ in cognition, namely, it contributes to the ‘constitution’ of the ‘object of cognition’ (pp. 8–9). In order to make his case, Horstmann offers a detailed reconstruction of what he calls Kant’s ‘two-stage’ model of the constitution of objects of cognition, and he seeks to show that imagination has a crucial role to play at both stages (p. 64).

Horstmann presents the two-stage model as follows. The first stage occurs at the ‘sensory’ level and results in an intuition, while the second stage occurs at the ‘conceptual’ level and results in a representation of an object of cognition (p. 27). In more detail, Horstmann claims that the first stage breaks down into two ‘phases’ (p. 32). In the first phase, ‘sensations’, which he glosses as ‘physiological events’ that are ‘not conscious’, are transformed into ‘perceptions’, which he interprets as ‘conscious states’ that ‘represent qualities like colors, shapes, sounds, etc.’ (p. 31). Then in the second phase, perceptions are transformed into ‘intuitions’, which Horstmann describes as ‘complex representations of as yet undetermined objects’, i.e. ‘appearances’ (pp. 29, 101). Meanwhile, in the second stage, those intuitions are transformed into conceptual representations of the objects of cognition, through which those objects are constituted (p. 8, n. 6).

Horstmann contends that imagination has an active role to play at both stages, and that in the first stage it plays its role in an autonomous fashion. In the first phase of the first stage (sensations to perceptions), imagination engages in ‘non-synthetic activities’ through which it ‘discerns’ sensations that can be individuated into perceptions (pp. 35, 33). Here, Horstmann maintains that imagination has ‘absolute’ freedom: ‘it is free to follow either no rules or rules that are exclusively its own’ (p. 43). Then in the second phase (perceptions to intuitions), he argues that imagination has ‘relative’ freedom: although it is free in the sense that it is not restricted by concepts, categorial or otherwise, it must produce intuitions through synthetic activities that are consistent with the demands of apperception and the ability ‘to think of [oneself] as an identical subject’ (pp. 42–3). By contrast, in the second stage, insofar as the goal is to produce a conceptual representation of an object of cognition, Horstmann claims that imagination’s synthesizing activities are constrained by the categories and it thus lacks freedom from understanding.

According to Horstmann, attributing the two-stage model to Kant has two advantages. First, by clarifying the imagination’s contribution to the two-stage model, Horstmann claims his reconstruction ‘can make plausible Kant’s view that the power of imagination has a genuine and original function to fulfill’ (p. 30). Second, he submits that his view ‘can explain the differences between the two versions of the Deduction’ (p. 30). On Horstmann’s interpretation, Kant came to regard the A version as ‘flawed’ because he gave an inadequate account of how imagination operates at the first stage (p. 47). Horstmann suggests that, ‘in a tacit confession of defeat’, Kant rewrites the B Deduction in such a way that ‘omits’ an analysis of the first stage, concentrating instead on the second (p. 48). However, he insists that this omission represents ‘no substantial change’ in Kant’s view, rather a shift in focus (p. 49).

In section 2, Horstmann turns to the role imagination plays in Kant’s theory of cognition in the third Critique. According to Horstmann, in the third Critique Kant remains committed to the two-stage theory of the constitution of objects of cognition. Horstmann adduces evidence for Kant’s commitment to the first stage from his remarks on imagination’s aesthetic ‘comprehension’ and ‘composition’ of intuition in §§26 and 35 (pp. 62–3). Meanwhile, with regard to the second stage, Horstmann argues that Kant refines his view from the first Critique by carving out a space for imagination to function with freedom even at the second stage. In particular, Horstmann takes Kant’s account of free play and of ‘schematiz[ing] without a concept’ to point in this direction (CPJ, 5: 287). For in these activities, Horstmann claims that even though imagination operates freely, without any conceptual constraint, it nevertheless ‘awakens understanding’ by producing intuitions or images that are ‘suitable for conceptualization’ and ‘for which a concept introduced by understanding can be found’ (pp. 85, 97–8). On Horstmann’s interpretation, this concept-free schematizing is one that imagination is not compelled to by understanding, but rather engages in ‘voluntarily’ (p. 102). In this way, Horstmann takes Kant to extend his analysis of imagination as an autonomous faculty in significant ways in the third Critique.

While much more could be said about Horstmann’s rich analysis, in what remains I want to focus on three sets of concerns that pertain to his characterization of the relationship between imagination and intuition, the shift between the Deductions, and aesthetic experience.

Beginning with imagination and intuition, one of Horstmann’s central moves is to deny what many interpreters take for granted, namely, that sensibility alone is what delivers intuitions. Insofar as sensibility is ‘the capacity to receive sense impressions passively’, Horstmann claims it cannot actively transform sensations into intuitions the way the first stage requires (pp. 11, 33). Instead it is imagination, on Horstmann’s interpretation, that is responsible for intuitions.

Although I am sympathetic to reading Kant as committed to imagination producing intuitions in the cognitive context, Kant distinguishes between different types of intuitions:

Given this distinction between intuitions of the senses and of imagination, even if we concede that imagination produces intuitions in cognition, this does not necessarily mean that it alone is responsible for intuitions. To the contrary, one could read the progression from the Transcendental Aesthetic to the Analytic of Concepts as follows. In the Transcendental Aesthetic, Kant offers an account of intuitions of the senses, which give us objects, independently from imagination. Then in the Analytic, he offers an analysis of what is done to those intuitions through the synthesizing activities of imagination and understanding. Now, even if one endorses this view, it is still possible to allow for imagination to produce intuitions in the cognitive process, e.g. the ‘images’ Kant describes in the A Deduction (A120–1). However, this progressive reading treats those intuitions of imagination as parasitic on intuitions already delivered through the senses. Such a reading would allow one to accord intuitions to imagination in the cognitive process without denying them to the senses, as a straightforward reading of the Aesthetic and the transition to the Analytic seems to suggest.

My second concern pertains to Horstmann’s characterization of what motivates the shift between the A and B Deductions. As I understand Horstmann’s reading, once Kant recognizes that he gives a flawed view of the first stage in the A Deduction, his strategy is one of avoidance: he simply sets the topic aside. I have two worries about attributing this strategy to Kant. In the first place, it is not clear to me that Kant does set the first stage aside in the B Deduction. Although Horstmann appears to dismiss the remarks on imaginative synthesis in §26 as an aside (p. 48), another way of reading this section is as culminating, in part, in an analysis of how we arrive at ‘perception’ through imaginative synthesis. So interpreted, one of Kant’s main goals in §26 is to offer an account of what enables imagination to engage in ‘the synthesis of apprehension [i.e.] the composition of the manifold in an empirical intuition, through which perception, i.e., empirical consciousness of it (as appearance), becomes possible’ (B160, my emphasis). And this goal, you might think, is what motivates him in picking as one of his summary examples how we ‘make the empirical intuition of a house into perception through apprehension of its manifold’ (B162, my emphasis).

Horstmann might respond that this is a case of the ambiguity of the term ‘perception’ (pp. 50–1); however, I have a further worry. Suppose that Kant does omit an analysis of the first stage in the B Deduction. Per Horstmann, Kant neither regards the account of sensibility in the Transcendental Aesthetic as a self-standing account of how we arrive at intuitions, nor is he satisfied with the ‘flawed’ version of this stage in the A Deduction. If this is the case, then Kant’s omission seems dire: he needs an account of the formation of intuitions for his theory of cognition. Yet if we appreciate that the Deduction builds on the prior account of intuitions of the senses that give us objects in the Aesthetic, then we do not need to attribute this strategy to Kant.

My third set of concerns pertains to Horstmann’s analysis of aesthetic experience. Although giving a full account of aesthetic experience is not Horstmann’s goal, my worry is that the picture that he does paint is overly cognitive. Here, I am going to focus on Horstmann’s discussion of aesthetic experience in which we lack the relevant concept for the object. Horstmann addresses such experience in two contexts: first, in his discussion of aesthetic experiences that occur in a ‘reflective scenario’ in which we have an intuition but lack the relevant concept (p. 68), and, second, in aesthetic cases of schematizing without a concept (p. 100). Horstmann suggests that, even though we lack the relevant concept, imagination relates to understanding because we feel that the intuition ‘might be open to some conceptual interpretation’ (p. 75) or that the image is one ‘for which a concept could be found’ (p. 100). The implication is that there is, indeed, a concept that fits my intuition, but I simply have not found it yet. This seems to be borne out by Horstmann’s description of experiencing a Turner painting as beautiful: you experience the painting as ‘the most fitting realization of a concept that is on the tip of [your] tongue’ and when you are told that the painting is ‘meant to depict a sunrise’, you ‘react by exclaiming: Yes! That’s It!’ (p. 76, n. 56).

My unease about this characterization of aesthetic experience is that it is only contingently without a concept: if we but search hard enough we will find the right one. However, as I read Kant’s view, aesthetic experiences are ‘without a concept’ in a stronger sense: there is simply no concept that could do full justice to what we intuit. I believe that this comes out most clearly in Kant’s account of aesthetic ideas, a topic Horstmann sets aside as not relevant to Kant’s account of cognition (p. 102, n. 72). According to Kant, ‘Beauty (whether it be nature or of art) can in general be called the expression of aesthetic ideas’ (CPJ, 5: 320). And Kant defines an aesthetic idea as ‘that representation of imagination that occasions much thinking though without it being possible for any determinate thought, i.e., concept, to be adequate to it’ (CPJ, 5: 314). As I understand Kant, an aesthetic idea is thus something that is so imaginatively rich that no single concept can do justice to it. So even if I learned that the Turner painting is of a sunset, if the painting embodies an aesthetic idea, then the concept of a ‘sunset’ will be inadequate to it. By my lights, Kant’s appreciation of this conceptual open-endedness of aesthetic experience is one of the deep insights of the third Critique and I thus worry that Horstmann’s effort to use the third Critique as a tool to teach us about cognition, valuable as this effort is, threatens an assimilation of aesthetic experience to cognition.

Possible concerns aside, in this incisive volume, Horstmann offers an interpretation that not only promises to dispel certain mysteries about Kant’s account of imagination, but also makes a valuable case for why anyone interested in Kant’s theory of cognition needs to take his analysis of imagination seriously.