Introduction

After initial enthusiasm sparked by Strawson’s The Bounds of Sense (1966), Kant’s transcendental arguments have been sharply criticized by analytic commentators. As Stroud (1977b, 105) observed, “it is not easy to incorporate the depth and power of Kant’s transcendental deduction into present-day philosophical attitudes and preconceptions.” However, rather than trimming Kant’s views to conform to contemporary predilections, philosophically it is much more illuminating to reconsider some of our present-day attitudes and preconceptions in order to understand and benefit from Kant’s transcendental proofs.¹

This book aims to contribute to contemporary epistemology as well as to Kant scholarship. Central to this study are important yet unappreciated resources of Kant’s Critique of Pure Reason, both methodological and substantive, that provide a genuinely transcendental proof of realism sans phrase. Kant has several projects in the first Critique. Kant’s main project is to establish the possibility of a priori knowledge, and thus the possibility of rigorous (“scientific”) metaphysics.² Another is to explain how mathematics and physical science are possible (B20, Prol. §§5, 6, 15).³ However, Kant has a third key aim in the first Critique. While examining the possibility of rigorous metaphysical science, Kant provides a sound transcendental response to global perceptual skepticism. Kant’s concern with such skepticism is reflected, for example, in his famous remark on the philosophical scandal that no one had yet proven the existence of the external world (Bxxxix note). Despite his confidence in mathematics and Newtonian physics as paradigm examples of knowledge, including synthetic a priori knowledge (B128), when introducing the vital importance of the B Deduction Kant recognizes that the key philosophical issue concerns whether the subjective conditions of thought are objectively valid, that is, whether the subjective conditions of thought are conditions of the very possibility of any and all knowledge of objects (B122–3, especially 3:102.30–2; cf. B127, 3:105.15–17).

Kant’s “Refutation of Empirical Idealism” has an anti-Cartesian conclusion: “inner experience in general is only possible through outer experience in general” (B278–9). Due to widespread preoccupation with Cartesian skepticism, and to the antinaturalism of early analytic philosophy (reflected in its basic division between “conceptual” and “empirical” issues), most of Kant’s recent Anglophone commentators have sought a purely conceptual, “analytic” argument in Kant’s Refutation of Idealism – and then criticized Kant when no such plausible argument can be reconstructed from his text. They charge that Kant’s transcendental arguments must argue by elimination, though they fail to eliminate the possibility of Descartes’s evil deceiver, or alternative forms of cognition, or the possibility that the mere (individually subjective) appearances of things would suffice for the possibility of self-consciousness.⁴

In chapter 1 I argue that these disappointments overlook three key features of Kant’s response to skepticism: the decidedly non-Cartesian philosophy of mind involved in Kant’s epistemology, Kant’s semantics of cognitive reference, and Kant’s decidedly non-Cartesian philosophical method.⁵ Scholarly attention has focused so exclusively on Kant’s transcendental proofs and transcendental idealism that Kant’s key methodological innovations have been neglected. Kant developed a new philosophical method for conducting his critique of pure reason, and for devising and assessing his transcendental proofs and his defense of transcendental idealism, called “transcendental reflection.” Kant’s account of transcendental reflection, like his name for it, are conspicuously rare, almost absent, from Kant scholarship. Yet Kant insists that

transcendental reflection is a duty from which no one can escape if he would judge anything about things a priori.

(A263/B319)

Kant fostered some of this neglect by providing no comprehensive account of transcendental reflection. Chapter 1 highlights its key features. Kant uses transcendental reflection to identify several of our key cognitive capacities by identifying several of our key cognitive incapacities. These cognitive capacities are logically contingent, though transcendentally necessary conditions for the very possibility of human knowledge, indeed for the very possibility of self-conscious human experience. Kant identifies some of our cognitive capacities by using wildly counterfactual thought experiments. Four of these thought experiments are considered briefly to elucidate their role in transcendental reflection and to explicate how these thought experiments are to be properly appreciated. Each of these four thought experiments is examined extensively in later chapters.

Kant’s epistemology highlights four integrated ways in which we are cognitively dependent on a commonsense spatiotemporal world; his semantics entails that the skeptical hypotheses which alone call such dependencies into question are themselves cognitively transcendent, idle speculations. Both of these aspects are required to prove, as Kant puts it, the reality of outer sense (Bxl–xli note, B276–7 note), namely that we sense and do not merely imagine perceptible objects distinct from ourselves. Both of these aspects are required to understand Kant’s transcendental response to global perceptual skepticism, which concludes this book. I contend that this approach to understanding and reconstructing Kant’s transcendental proofs yields a sound, genuinely transcendental proof of realism sans phrase regarding our empirical knowledge of molar objects and events in space and time.

The transcendental proof developed here differs markedly from those familiar in the literature. However, it is squarely based on a key transcendental proof of realism – and, surprisingly, for mental content externalism – that Kant himself provides, though without pursuing it to its surprising logical conclusion. I argue that this key transcendental proof ultimately shows that transcendental idealism is groundless, because Kant’s arguments for transcendental idealism are unsound.⁶ Kant’s own transcendental analysis of the necessary a priori conditions for the possibility of unified self-conscious human experience ultimately provides a sound version of the standard objection to Kant’s arguments for transcendental idealism, the so-called “neglected alternative.” Hence I use Kant’s own transcendental analyses to show that Kant’s own transcendental idealism is untenable. Guyer’s (1987, 417) criticisms of Kant’s transcendental idealism are stated in terms “Kant would have understood.” This is an important point of critical charity. My arguments take this charity a significant step further, by criticizing transcendental idealism squarely on the basis of Kant’s own analyses in the Transcendental Analytic.

The significance of this result depends on the character of Kant’s transcendental idealism, discussed in chapter 2. Kant’s discussions of transcendental idealism frequently suggest that it is a highly metaphysical view, including the occurrence of causal events in a “noumenal” realm that transcends space and time. Historically, the standard interpretation of Kant’s idealism was metaphysical. However, the idea that noumena or things in themselves causally affect our sensibility, and thus provide us with sensations, has been rejected on two basic grounds, forcefully advanced recently by Bird and Strawson: it is unintelligible because it distinguishes between appearance and reality in such a way that things cannot in principle appear as they really are, and it requires applying the concept of causality transphenomenally, contra Kant’s Schematism of the Categories. In response to these and related objections, some recent scholars (e.g., Prauss, Allison, and Buchdahl) have argued that such objections do not pertain to Kant’s views, because despite some suggestions to the contrary, Kant’s transcendental idealism is not metaphysical, it is only methodological or transcendental or only rests on two points of view or two kinds of description. Though prominent, such nonmetaphysical interpretations are now subjected to sustained criticism from scholars who once again defend metaphysical interpretations of Kant’s transcendental idealism.⁷

In chapter 2 I defend a metaphysical dual-aspect interpretation of Kant’s transcendental idealism. I focus on the test case of noumenal causality, which I argue is intelligible and is required out of fidelity to Kant’s texts and doctrines. In general, transcendental proofs aim to establish a priori conditions necessary for our having self-conscious experience at all. Transcendental idealism holds that such conditions do not hold independently of human subjects; those conditions are satisfied because they are generated or fulfilled by the structure or functioning of our cognitive capacities. Kant argued repeatedly that transcendental idealism is the only possible explanation of the transcendental conditions of possible experience. Kant’s analysis of human agency also shows that his practical philosophy is committed to noumenal causality, both from a first- and from a third-person perspective. The standard objection from Jacobi to Strawson is that Kant’s transcendental idealism is incoherent. In reply, I argue that Kant’s theory of meaning and his transcendental reflection on sensibility show how Kant legitimately can speak about, and determine that, our passive sensibility must be causally affected by nonspatiotemporal noumena. These points ground my criticisms of Allison’s view of affection, and Strawson’s view of meaning.

Showing that Kant’s transcendental idealism is metaphysical underscores the significance of the conclusion to my reconstructed Kantian proof of realism (sans phrase). This proof entails realism, broadly construed, about molar objects and events in our environs. Showing why Kant’s idealism is coherent also reveals some very important, though widely neglected features of Kant’s semantics. Kant’s semantics are important for his reply to global perceptual skepticism. Fortunately, my criticisms of transcendental idealism do not undermine Kant’s semantics. Instead they indicate that Kant’s semantics is separable from his transcendental idealism, and so is available for the revised transcendental proof of realism developed here.

In chapter 3 I argue that Kant was mistaken that transcendental conditions of possible experience require transcendental idealism. I further argue that Kant can be shown to be mistaken on the basis of his own transcendental proofs. I defend these claims by analyzing a widely neglected doctrine of Kant’s, “the transcendental affinity of the sensory manifold.” I argue for six claims: (1) This doctrine remains vital to the second edition of the Critique, even though many passages on the topic were omitted from that edition; (2) Kant’s link between transcendental idealism and transcendental arguments is substantive, not methodological; (3) Kant’s views on transcendental affinity show that there are nonsubjective, transcendental material conditions for the possibility of unified self-conscious experience; (4) These conditions and Kant’s arguments for them directly undermine Kant’s own arguments for transcendental idealism; (5) These points reveal some serious flaws in Allison’s defense of Kant’s idealism; (6) Realists of any stripe have much to learn from Kant’s transcendental analysis of the conditions of unified self-conscious experience, because Kant’s doctrine of the transcendental affinity of the sensory manifold provides a sound argument supporting the conclusion of his Refutation of Empirical Idealism, that inner experience in general is only possible through outer experience in general. Indeed, Kant’s analysis of transcendental affinity provides a transcendental proof of (not “from”) mental content externalism.

Chapters 4 through 6 develop a second criticism of transcendental idealism internal to Kant’s first Critique. Famously, one of Kant’s central aims is to justify our causal judgments about spatiotemporal objects and events, and thus to answer Hume’s skepticism about our knowledge of or beliefs about such relations. The standard view among Anglophone Kant scholars is that Kant’s Transcendental Deduction of the category of causality fails, and that a sound argument to justify our causal judgments can be found, if at all, in the Analogies of Experience, especially the Second Analogy, which is almost universally supposed to contain Kant’s “answer” to Hume’s skepticism about causality. However, Strawson condemned Kant’s argument as “a non-sequitur of numbing grossness”; a charge to which Beck (among others) responded vigorously.

Kant’s justification of causal judgments in the Analogies is examined in chapter 4. This examination requires pursuing Kant’s analysis of causal judgments much further into his Metaphysical Foundations of Natural Science (MAdN, or Foundations for short). These further issues are considered in chapters 5 and 6. One main point is that Kant was far more subtle about the issues pertaining to causal judgment than either his commentators, or other philosophers addressing these issues. On this topic, philosophers have much to learn from some careful Kant scholarship. Kant’s “answer” to Hume is not provided by the Second Analogy! Kant’s three Analogies of Experience form an integrated set; no one of the principles of causal judgment defended in the Analogies can be used without conjoint use of the other two. This important fact has been widely disregarded, although Guyer (1987) clearly identified it.⁸ In chapter 4 I further develop Guyer’s point, arguing that the integration of the three Analogies is even deeper, more thorough, and more important than he recognized. Ultimately, the integrity of the principles of the three Analogies entails that we can only make legitimate causal judgments about spatiotemporal objects and events.

Reexamining these issues about causal judgment reveals absolutely fundamental points about Kant’s justification of causal judgments, including a fundamental flaw in the justification provided in the first Critique, a flaw Kant himself recognized. In 1787 and 1792 Kant noticed two basic problems with his Foundations. Their only solution is to divorce metaphysics from mathematics. When Kant did this in 1798, it opened a crucial gap, not only in the Foundations, but in the Critical system as a whole. Why is the Foundations so important to Kant’s Critical philosophy? The Critique of Pure Reason defends the general causal principle that every event has a cause. However, the Analogies of Experience require the specific principle that every physical event has an external cause. This principle is not defended, indeed it is not even formulated, in the first Critique – nor is it often identified or defended (rather than assumed) in other philosophical analyses of causality. Only in the Foundations does Kant first distinguish these two causal principles, and only there does he attempt to justify the second, specific principle. Kant’s defense of this specific causal principle in the Foundations is coupled with an important shift in Kant’s view of the metaphysical basis of his transcendental philosophy, and with an ineluctably empirical basis of metaphysics. These two results, derived from central principles of Kant’s Critical philosophy, subvert the fundamental structure of Kant’s system of transcendental philosophy (below, §§35, 41, 48): transcendental philosophy cannot have its intended priority over Kant’s Critical metaphysics.⁹

Kant’s difficulties do not end there. Careful examination shows that Kant’s justification of the specific causal thesis, that all physical events have an external cause, is irreparably flawed. Chapter 5 examines Kant’s attempt to justify the application of mathematics to objects in natural science by metaphysically constructing the concept of matter. Kant develops these constructions in the Foundations. Kant’s specific aim is to develop a dynamic theory of matter to replace corpuscular theory. In the preface to the Foundations Kant claims to completely exhaust the metaphysical doctrine of body. However, in the general remark to “Dynamics” (MAdN, ch. 2) Kant admits that once matter is reconceived as basic forces, it is impossible to construct the concept of matter.

I argue that Kant’s admission is only the tip of the problem. I show that Kant’s proof that matter consists of forces is fallacious. I then reanalyze and substantiate the circularity in Kant’s definition of density. These two fundamental problems demonstrate the untenability of Kant’s metaphysical method, and they require the radical revision of the relation between mathematics and metaphysics Kant undertakes in his opus postumum. I show that some of Kant’s most surprising and critical later claims about the Critical philosophy are correct, and that they require the sorts of remedies Kant contemplates in the opus postumum.

These are very significant findings, at least within Kant’s Critical philosophy. However, there are further difficulties. Chapter 6 examines Kant’s key aim of justifying Newtonian mechanics by showing how physics as a rational science is possible. According to Kant’s Foundations, a proper science is organized according to rational principles and has a pure a priori rational part, its metaphysical foundation. In the preface to the B edition, Kant claims that his account of time explains the a priori possibility of Newton’s Laws of Motion. I argue that Kant’s proof of the Law of Inertia fails, and that this casts grave doubt on Kant’s enterprise of providing a priori foundations for Newton’s physics. Hence even if Kant’s transcendental and metaphysical analyses of causal judgment were sound, they would fail to achieve another of Kant’s key aims. More importantly, the failure of Kant’s proof of the Law of Inertia also marks the failure of Kant’s transcendental idealist proof of the specific causal thesis, that every physical event has an external cause. Thus Kant’s transcendental idealism fails to deliver Kant’s promised “answer” to Hume. Viewed systematically, Kant’s Critical metaphysics also cannot have its intended priority over physical science and empirical fact.

The irreparable flaw in Kant’s metaphysical proof of the Law of Inertia underscores the failure of Kant’s attempt to underwrite physics by philosophy, and strongly suggests the impossibility of providing such philosophical foundations of physics, whether transcendental or metaphysical (in Kant’s Critical sense of the term). These findings strongly reinforce the philosophical turn, away from Kant’s foundational program, based in his untenable and systematically inadequate transcendental idealism, towards a broadly realist approach to epistemology. Unique to my development of this theme is that this turn towards realism can and ought to make central use of Kant’s transcendental reflection and can and ought to learn much from Kant’s own transcendental proofs of the necessary conditions of the very possibility of our self-conscious experience. In this way, the critical findings of chapters 4 through 6 reinforce both the critical and the constructive findings of chapter 3.

Examining Kant’s analyses in such detail also reveals some important philosophical insights. Chapter 7 develops four of these, regarding freedom of action, causal judgment, Kant’s anti-Cartesianism, and global perceptual skepticism. Kant contends that our freedom of thought and agency can only be defended by transcendental idealism. If that were correct, my criticisms of Kant’s arguments for transcendental idealism would have drastic implications for Kant’s account of rational agency. I argue that these implications do not hold. Instead, the debate about the relation between the (phenomenal) psychological realm and our (noumenal) rational freedom is moot because, although he only notes it once, Kant in fact argues that psychological determinism is in principle unknowable, even in the phenomenal realm. This is the joint conclusion of Kant’s Analogies and Paralogisms. Kant contends that causality is strictly related to substance. The three Analogies form a mutually integrated set of principles. Kant’s Paralogisms show we have no knowledge of a substantial self. If we have no evidence of a substantial self, then we cannot apply any of the principles of the Analogies to the self. Consequently, we cannot justify any determinate causal judgments in psychology (as Kant understood it). Hence determinism is in principle unjustifiable even within empirical psychology. This result suffices, on Kant’s views, to permit us to appeal to moral considerations to settle the question, as well as we can, whether we are free. Kant’s Incorporation Thesis and his account of moral responsibility provide sufficient grounds to answer affirmatively.

This book provides ample evidence of Kant’s fallibilism and externalism. “Externalism” is the view that factors of which a subject is not or perhaps cannot be aware have a significant bearing on that subject’s “mental” contents, semantic meaning, or the justificatory status of his or her beliefs. The term coins recent developments within analytic philosophy of mind, philosophy of language, and epistemology. The idea, however, is not new. Kant was the first great non-Cartesian. His transcendental analyses of the necessary conditions for the very possibility of self-conscious human experience aim to uncover a host of factors of which people ordinarily are quite unaware (by Kant’s account, no one was aware of them prior to 1781) that alone provide the necessary framework within which human beings can so much as have beliefs or consider questions about their justificatory status. Kant was a staunch justificatory externalist avant la lettre. Kant’s externalism, however, does not beg the question against global perceptual skepticism: Kant’s externalism need only be true, it need not be known to be true, to serve the role Kant assigns it in his transcendental proofs, in which Kant does not, and need not, appeal to his externalism as a premise.

I contend that Kant’s attempts to eliminate the possibility of alternatives to the specific causal principle, that every physical event has an external cause, are excessively preoccupied with the infallibilist notion of justification that is central to Cartesian-Humean skepticism, and to Kant’s demonstrative, “apodictic” ideals of transcendental, metaphysical, and systematic knowledge. The proper solution to Kant’s problems about causal judgment lies instead in developing further our understanding and use of Kant’s transcendental reflection and his semantics of cognitive reference. In principle the alternatives to the specific causal principle are cognitively transcendent, idle metaphysical speculations. The solution to these worries lies in appreciating both the strengths and the limits of our cognitive situation as human beings. Identifying and appreciating these is precisely the aim of transcendental reflection. If we appreciate these, we can understand, assess, and accept a revised, genuinely Kantian transcendental proof of transeunt causality in the form of the metaphysical causal principle. This provides a genuinely transcendental proof of realism sans phrase. This is a second, genuinely transcendental argument for mental content externalism.

Recent devotees of analytic transcendental arguments have found Kant’s response to perceptual skepticism wanting, a view prominently advocated by Stroud. I argue that Kant’s semantics of cognitive reference suffice to show that global perceptual skepticism is a prime instance of transcendental illusion. Demonstrating this reveals several key assumptions and oversights in Stroud’s presentation of global perceptual skepticism. This shows that such skepticism is not at all the innocent, commonsensical phenomenon Stroud claims. Global perceptual skepticism rests on deceit and petitio principii, in ways revealed by Kant’s transcendental proof of realism.

Chapter 7 thus contends that, pace Kant, transcendental idealism is not necessary for responding by transcendental proof to global perceptual skepticism, nor is it necessary for defending the theoretical possibility of free rational action; instead, it is a desperate gasp of a misleading and dispensable infallibilism, a view already undermined and replaced by Kant’s fallibilist “new way of thinking,” transcendental reflection. In sum, transcendental affinity provides a genuine transcendental proof of (not “from”) mental content externalism. This proof entails that transcendental idealism is false, and identifies the key fallacy in Kant’s arguments for transcendental idealism. Transcendental idealism also fails to underwrite Kant’s analysis of causal judgments, because Kant’s Foundations fails to fill the “gap” Kant identified in the first Critique. Four integrated “externalist” aspects of Kant’s epistemology are identified, concerning the source of sensations, mental content, the objects of causal judgment, and justification: Kant is a fallibilist about empirical knowledge and about epistemology. This book defends a transcendental proof of realism sans