Anil Gomes and Andrew Stephenson have compiled a fantastic collection of essays from some of the brightest scholars actively working on issues pertaining to Kant’s theory of the mind. While almost all areas of the philosophy of mind are addressed to some degree, the fourteen papers in this volume focus on three issues: intuitions and their objects, the possibility of knowing one’s own mind and the relations between judgements. The essays are generally lucid and their arguments informed by the wealth of secondary literature on Kant and often also by contemporary work in the philosophy of mind and related scientific fields. One especially nice feature of the book is that it often has the feel of a symposium, as many essays explicitly cross-reference each other and engage with each other’s arguments. In what follows, I’ll attempt to give a sense of the main questions, arguments and debates discussed in these essays.
Anil Gomes’s introductory essay (‘Kant, the Philosophy of Mind, and the Twentieth-Century’) does some nice stage-setting. Among other things, it includes a brief history of the reception of Kant’s thought in the twentieth century. He catalogues first the rejection of Kant early in the century by British thinkers (especially Moore, Russell, Cook Wilson and Prichard), and then Kant’s rehabilitation in the middle of the century (through the work of Lewis, Strawson, Sellars, and then their many students in the next generation).
In the next two essays, Lucy Allais (‘Synthesis and Binding’) and Katherine Dunlop (‘Understanding Non-Conceptual Representation of Objects’) present competing interpretations of Kantian intuitions. While both defend non-conceptualism, they offer different takes on the representational content of intuitions. Many conceptualists assume that intuitions are formed through acts of synthesis (and are thereby conceptual), but Allais denies this and claims instead that intuitional content should be understood in terms of what contemporary psychologists and cognitive scientists call sensory ‘binding’, i.e. the combining of diverse sensory inputs into representations of singular objects. Binding generates intuitions, and synthesis (which she concedes is always conceptual) is something that happens to intuitions that have already been formed.
While Dunlop is sympathetic to Allais’s non-conceptualist motivations, she proposes an alternate interpretation. She rejects Allais’s proposal about sensory binding on the grounds that most contemporary research on binding locates the process at a neural level rather than a functional level (intuitions need to be explained functionally, not neuro-mechanically, she thinks), and because many ‘objects’ of sensory binding (e.g. dots on screens or shadows) will not be ‘objects’ in the properly Kantian sense. Her alternative proposal is that intuitional content should be understood in terms of what contemporary psychologists call ‘primitive object representation’ (p. 56). This has to do with the capacity (present already in infants) to represent the continuity of an object’s motion (even when part of the motion is occluded) and the cohesion of an object’s contours. She defends her view by showing that these abilities are functional (not neuro-mechanical) and that they will generally pick out proper Kantian objects. While I think Dunlop makes a strong case for her interpretation, part of me wonders whether she and Allais could not both be right. Kant may well have wanted to ascribe both sensory binding and primitive object representation to intuitions. After all, writing in the eighteenth century, he would not necessarily have been in a position to know that these are cognitively distinct processes.
The next three papers (Stefanie Grüne’s ‘Are Kantian Intuitions Object-Dependent?’, Colin McLear’s ‘Intuition and Presence’ and Andrew Stephenson’s ‘Imagination and Inner Intuition’) are also about intuitions, though now the debate switches to the question of ‘object-dependence’. When Kant says that intuitions ‘present’ objects to the mind, is he saying that the intuition’s object must literally exist at the same time as the intuition, in some real relation to the subject? Or must the object have existed at some time in the past? Or must the intuition only seem (phenomenally) to present an existing object? This question is important because it bears directly on currently popular ‘relationalist’ interpretations of Kant’s theory of cognition. According to such interpretations, intuition is not a mere mental state in the mind, but rather a relation between the mind and the intuited object. But this would mean that intuitions, qua relations, are partially constituted by their objects, which means that there could not be an intuition without the presence of an object. Hence, the relationalist has a great deal at stake in this debate.
Grüne argues against the strong kind of object-dependence that would be needed to support the relationalist reading (‘strong’ object-dependence requires that the object exist at the same time as the intuition, while a ‘weak’ object-dependence would require only that the object must have existed at some earlier time; see pp. 67–70). Her strongest grounds for rejecting strong object-dependence are primarily textual. For one, Kant describes an a priori intuition as one that precedes the presence of its own object (P, 4: 282). However, the smoking gun is Kant’s claim that imagination (especially as manifest in hallucinations and dreams) involves intuitions of objects that do not exist, which seems to be in direct conflict with the object-dependence claim (even in its weaker form, since dreams and hallucinations are often of objects that have never been experienced at all; McLear will make this point too, see p. 92). Grüne concludes that the strong view is definitely untenable, and the weak view is probably false too.
McLear is the only one of the three willing to defend a strong form of object-dependence, but even his support is somewhat tentative. For on the one hand, he agrees with Grüne that the text points away from a strong reading, and even from the weak reading. He instead proposes an even weaker form of object-dependence – the intuited object’s properties must have been present at some time now or in the past – which is textually supported, but does not do much to help the relationalist case (p. 93). Yet on the other hand, he offers an interesting take on the hallucination problem that might allow for the strong reading to come out on top after all. His suggestion is that hallucinations are intuitions and do involve the immediate presence of an existing object, but they are inner intuitions of inner objects (for example, the object presented in a hallucination of a dagger is an inner image of a dagger). In this way, the relationalist can retain strong object-dependence, while maintaining a strict distinction in kind between a veridical perception of an (outer) object and a hallucination of one.
Stephenson picks up where McLear leaves off, devoting his paper to the suggestion that relationalism’s strong object-dependence can be saved from the hallucination objection by appeal to this ‘inner intuition proposal’ (as Stephenson dubs it). Aside from a lack of adequate textual support, he thinks there are significant philosophical problems with the view. Most notably, it would entail that memories are about inner images rather than the (outer) things and events recalled (because memories are imaginings, and on this proposal imaginings are intuitions of inner, non-spatiotemporal objects). For what it is worth, this reviewer began these chapters leaning towards object-independence and is firmly in that camp afterwards.
The next three essays all address questions surrounding our awareness of our own mental states, and of the empirical self itself. Ralf Bader (‘Inner Sense and Time’) delves into some aspects of Kant’s transcendental psychology to answer the question of how we come to represent outer objects in time. Time is said to be the form only of inner sense, yet outer objects end up both spatial and temporal. It would be tempting to say that spatial objects gain temporality by way of the temporality of their representations, but since all representations are temporal, it would follow that even objects like God or noumena would be temporal as well, which can’t be correct. His solution is to distinguish between representation in general and the specific type of representation that involves awareness of an object. The claim is that, since time is the form of awareness, this form is transferred to outer objects through an act of reflection on the representations of spatial objects.
Andrew Chignell’s paper (‘Can’t Kant Cognize Himself?’) questions a common reading according to which Kant thinks it impossible to have the empirical self as an object of cognition. Chignell agrees that the empirical self cannot be an object of inner sense, but that this is compatible with it being an object of cognition, and, moreover, with it being cognized as a substance. After showing that the textual evidence weighs in favour of his reading, he rests his philosophical case on a point about parity: we manage to cognize outer objects as substances even though we perceive only their qualitative determinations; therefore, there is no obstacle to cognizing the self as a substance, even though we only perceive (through inner sense) its determinations (i.e. individual mental states).
Patricia Kitcher (‘A Kantian Critique of Transparency’) addresses our awareness of the contents of the mind in the context of the ‘transparency thesis’. This is the idea (popularized by Gareth Evans) that my rational justifications for asserting ‘I think that P’ are no different from my justifications for asserting simply that ‘P’. The idea is that I can report the contents of my belief without any sort of ‘inward glance’ to see what beliefs are floating around in my head. Kitcher offers a Kantian critique of this idea that appeals to the unity of apperception and the requirements for being a rational knower. While she agrees that self-ascriptions of belief do not involve an inner sensing of the contents of our minds, she claims that these self-ascriptions do require that we represent necessary connections between our representations (specifically, inferential relations). That is, self-ascriptions of belief involve a recognition of the logical dependence of some beliefs on others, and so in this sense these self-ascriptions are not perfectly ‘transparent’.
The next two papers both address Kant’s theory of judgement. Jessica Leech’s paper (‘Judging for Reasons’) concerns a puzzle regarding the modalities of judgements. Kant suggests that every judgement involves each of the four functions of judgement (quantity, quality, relation and modality). But modality does not concern the content of a judgement, so in what way does every judgement have a modality? Leech argues that the modality of a judgement consists in the fact that judgements must be ‘part of a connected course of reasoning’ (p. 186), and that the necessary connections represented between judgements give rise to the transcendental unity of apperception. What is especially noteworthy about this interpretation is that modality in judgement is hardly a sideshow, and instead takes centre-stage in Kant’s account of the conditions on experience: the modal relations between judgements are what first give rise to the transcendental unity of apperception.
Jill Vance Buroker (‘Kant on Judging and the Will’) addresses a related question about the force behind judgements (rather than the contents of the judgements themselves). Buroker considers the ‘illocutionary force’ behind judgements and asks whether assent is voluntary according to Kant. According to direct doxastic voluntarism, we have a real choice about whether to assent to a certain proposition. According to indirect doxastic voluntarism, we have a choice in whether (and how) to employ various epistemic maxims and procedures (e.g. whether to doubt, or whether to seek confirming or disconfirming evidence). Following recent work by Chignell, Buroker argues that only in the case of Kantian Glaube is assent directly voluntary. However, she concludes by arguing (persuasively) that Kant subscribes to an indirect doxastic voluntarism in all kinds of empirical cognition, insofar as the implementation of the regulative principles of systematic unity (described in the Appendix to the Transcendental Dialectic) are ‘performed voluntarily to satisfy the interests of theoretical reason’ (p. 203).
The final three papers all address questions relating to our knowledge of the self. Ralph Walker (‘Self and Selves’) considers a cluster of issues along these lines. Kant famously argues against any kind of substantive knowledge of the self as object of experience in the Paralogisms. However, according to Walker, he is too coy about some respects in which we can have knowledge of the self as subject of experience. Specifically, he considers the questions whether the noumenal self can coherently be conceived as existing atemporally (it can), whether we can know the self as identical over time (we can, sort of) and whether we are justified in believing in other minds (we are). This last point is especially interesting, because Walker offers on Kant’s behalf a novel argument based on moral considerations. Even if we have no solid theoretical reason for attributing rationality to other humans, the categorical imperative gives us an obligation to regard others as rational ends in themselves.
Tobias Rosefeldt (‘Subjects of Kant’s First Paralogism’) criticizes a standard reading of the First Paralogism. Kant’s target in the First Paralogism is the following sort of syllogism. If something can be thought only as subject, it is a substance; the thinking I can be thought only as subject; therefore, the I is a substance. Kant claims that this apparently valid argument contains a hidden equivocation. Where most commentators take ‘subject’ as the equivocated term (it is subject-rather-than-predicate in the major premise, and subject-rather-than-object in the minor premise), Rosefeldt makes the case that ‘can be thought’ is the equivocated term. The major premise has to do with objects that can be given in intuition, but the I mentioned in the second premise cannot be intuited at all. Rosefeldt concludes with some interesting remarks about how his interpretation can connect the claims of the Paralogisms with Kant’s general claim that all transcendental illusions arise from the attempt to seek the unconditioned behind conditions.
Finally, Paul Snowden (‘The Lessons of Kant’s Paralogisms’) offers some critical remarks about the Paralogisms (in the spirit of Strawson’s take on Kant). He is concerned to discern whether there are any upshots of Kant’s critique of rational psychology that could be relevant to contemporary philosophy. However, he thinks this project can get started only by discarding central elements of Kant’s theory, including his transcendental idealism and his claim that we cannot have an intuition of the self. He suggests that ‘once we abandon this whole structure Kant’s assumption that materialism cannot be right is lost, and there is no longer any deep epistemological problem with pursuing a theoretical investigation into our fundamental nature’ (p. 261). While the attempt to keep Kant relevant to contemporary philosophy is surely a laudable project, I am sceptical of these types of Strawsonian critiques, for the simple reason that it is not clear how much of Kant is really left once so many core aspects of his theory have been amputated.
Overall, this is a consistently excellent volume. Many of the papers are sure to inform and even inspire future work on these topics, and I recommend it highly for anyone working on related issues in Kant’s theory of the mind.
