<i>Kant on Freedom and Rational Agency</i>

Markus Kohl’s Kant on Freedom and Rational Agency is a thorough reconstruction and qualified philosophical defense of Kant’s views concerning the metaphysics of freedom, the nature of rational agency, and their connection to one another and to Kant’s broader philosophical commitments. As Kohl puts it, a “central goal of this book is to explicate the Kantian link between freedom as autonomy and normative concepts like rationality, objectivity, cognition, or knowledge” (1). That there are such links is difficult to deny, but Kohl goes on to make a stronger and more controversial claim—namely, that “there is no self-standing core of this account [of freedom and rational agency] that can be profitably separated from Kant’s idealist claim that the sensible-natural world depends (in some way) on the human mind” (2). This is a welcome thesis concerning a surprisingly anemic area of Kant studies. There have been significant investigations of Kant’s conception of freedom and autonomy, as well as his conception of reason and rationality. But very little critical philosophical investigation has been done concerning their philosophical connections, and still less concerning their metaphysical underpinnings. The best explanation on offer for this is unfortunately sociological. Since (primarily Anglophone) philosophers interested in reading Kant by and large have thought the doctrine of transcendental idealism, at least as a kind of metaphysical theory, to be entirely hopeless (though arguments for why this is the case are rather thin on the ground), every attempt has been made to dissociate his philosophical views concerning reason and rationality from their idealist trappings. Thus, Kohl’s book is a welcome departure from this lackluster historical reception of Kant’s metaphysics, and the book itself is an admirable and ambitious explication of Kantian philosophical commitments and their contemporary relevance and significance.The book has five parts with ten chapters total. The core argument of the first four parts of the book is framed in terms of an attempt to show that Kant’s metaphysics of freedom is defensible by virtue of arguments it provides against what Kohl calls “naturalistic compatibilism.” It must provide such arguments because, on Kohl’s reading, Kant’s metaphysics of freedom raises the worry that he offers us “an inscrutable model of free agency without explaining why we should reject alternative naturalistic models that allow us to understand and explain free agency” (5). Kohl’s aim is to show that acceptance of such inscrutability is worth the price, because the overall Kantian theory is more adequate to our self-conception as free persons than that which is provided by forms of naturalistic compatibilism. Given that the overall assessment of the arguments of the first four parts hinges on adequacy to our self-conception as free persons, Kohl’s reconstruction is best thought of in terms of what Strawson (1959: 9–10) called “descriptive metaphysics”—the project of describing the actual structure of our thought about the world, the content and commitments of our “conceptual scheme.”The final part of the book aims at integrating the position articulated through the first four parts, which primarily concerns themes related to Kant’s theoretical and practical philosophy, with his conception of the freedom of the imagination. Kohl argues that “aesthetic freedom” of the imagination is “a further (third) distinctive species of transcendental freedom” (337).The overall ambit of the book is impressively wide-ranging and covers a large cross section of both Kant scholarship and contemporary Kantian (and Kantian-adjacent) philosophical theory. It is dense but clearly written, and Kohl does an admirable job of trying to present Kant’s position in a manner that is not obfuscated by Kant’s jargon. As is typical of a short review concerning such rich material, I am able to discuss only a few parts of the book in detail. I focus on the framing of the arguments of the first four parts, and thus on the putative opposition between Kant’s views and those of what Kohl terms “naturalistic compatibilism.” I then raise concerns about two central arguments Kohl makes concerning the defensibility of Kant’s putative incompatibilist agent-causal libertarianism against its naturalistic compatibilist competitor.Kohl says that,Kohl concedes that his conception of naturalism is “stipulative” but contends that “it tracks a systematic position that Kant was concerned to address and refute” (9). The naturalistic position, so construed, when combined with the position that freedom and determinism are compatible, yields the naturalistic compatibilism that Kohl takes as the central foil to Kant’s view, and thus as the measure of whether there is a plausibly compelling version of Kant’s view on offer. The central conflict of the first four parts of the book is thus whether Kant’s “incompatibilist, anti-naturalistic view of free will is more adequate to our moral self-conception than compatibilist-naturalistic views” (10).I found this framing device rather problematic. Granting Kohl the fact that his is a rational reconstruction of Kant, rather than a primarily historically driven form of inquiry, nevertheless we want to be sure that we are engaging with Kant’s views themselves, rather than some contemporary facsimile thereof. But I find it difficult to recognize in the characterization of naturalistic compatibilism any plausible historical position with which Kant would have been concerned. For example, the metaphysical agnosticism of Locke and especially Newton, who refuses to feign hypotheses, would not obviously accept (3), and more generally would concede that there is no demonstration available of (2) or (4). No one convinced of Leibnizian or Wolffian principles would accept (1), and no one sympathetic to Hume would endorse (2)–(4). Indeed, Hume for much of his career rejects the existence of powers! It may well be true that naturalistic compatibilism is, as Kohl contends, “the dominant view in Anglo-American philosophy,” but he gives no evidence that it is a view held by “Kant’s most important historical predecessors including Hobbes, Hume, Wolff, and Leibniz” (4), and from the above we can surmise that this is because they did not adhere to any such view. Probably the best we can say is that Kant indeed opposes any position that accepts both a thoroughgoing temporal determinism (i.e., that a preceding temporal event necessitates a subsequent event) about events and says that all human action is to be understood fundamentally in terms of the occurrence of events. Kant contends that the combination of these two positions would undermine the possibility of any sort of free action capable of grounding attributions of responsibility and attitudes of praise and blame.There is, however, a distinctive upside to Kohl’s approach, as it allows him to frame nearly all of Kant’s arguments in terms of positions in contemporary Anglophone philosophy, and thus compare those arguments, favorably or not, to such positions. Indeed, this seems to be Kohl’s primary aim. For those interested in ‘the literature’ that constitutes the contemporary philosophical corpus, this is obviously useful. And Kohl does indeed articulate clear connections between his reading of Kant and contemporary philosophical views of responsibility, freedom, and agency. In the remainder of this review I take up two interesting arguments Kohl makes on behalf of Kant.Kohl contends that there are two central arguments Kant provides for thinking that naturalistic compatibilism cannot satisfy our moral self-conception. First, acts of judgment or reason understood naturalistically supply only conditional causes, a position that Kohl contends is inconsistent with our moral self-conception because “we presuppose that we are governed by objectively necessary norms” (129). Conditional causes are inconsistent with objectively necessary norms because naturalism fails to explain how such absolutist moral claims could ever be justified. Why might this be the case? I found the argument for this position somewhat elusive. The argument appears to consist of two claims, the second purportedly entailed by the first. The first claim says that (A) the “naturalistic mechanisms that process the intake of sensations into moral judgments are contingent” in the following sense: “the occurrence, character, and representational output of these processes depends on whether certain empirical (environmental, genetic, neurological, psychological) conditions obtain; whether these conditions obtain in turn depends on further empirical conditions; etc.” (111; see also 97–98). The second claim says that (B) naturalistic mechanisms so understood are “mutable as a result of being contingent: their character and representational output ‘may change subsequently’ (RezS, 8:14) due to changes in their natural conditions” (111).1I do not see how claim B follows from claim A. The fact that all events (as effects) occur as members of an infinitely regressive nonterminating chain of causes does not show that members of this chain are genuinely (i.e., absolutely) contingent. To do so we would also have to show that the causal chain as a whole could be otherwise than it is. But that is an additional claim, and it is not part of the original four theses that constitute “naturalism” as Kohl understands it.Indeed, the supposed connection between A and B flies in the face of Kant’s own famous criticism of the compatibilisms of Wolff and Leibniz, which depend on a distinction between hypothetical (i.e., Kohl’s “conditional”) and absolute necessity. Kant argues that “the distinction, which everyone recites parrot-fashion, does little to diminish the force of the necessity or the certainty of the determination. For just as nothing can be conceived which is more true than true, and nothing more certain than certain, so nothing can be conceived which is more determined than determined” (NE 1:400).Hence, so far as naturalism (as Kohl defines it) is concerned, there is nothing about the conditional existence of an effect based on its cause that entails the position that effects are “mutable” in the manner indicated, nor does this seem to be a central hook on which Kant hangs his own argument.Kohl’s second argument against naturalistic compatibilism contends that naturalistic causation is incompatible with the contingency of our choices. Here the argument is that (C) normative obligations (of permission and requirement) presuppose the ability to do otherwise; (D) causal determinism entails the inability to do otherwise. From (C) and (D) and the acceptance that causal determinism is true, Kohl’s Kant concludes that naturalism is incompatible with the position that moral requirements apply to us.Kohl’s interpretation here advocates that we see Kant as holding “Strong Imperfectionism,” which is the claim that “we can never be inevitably determined by reason to act rightly” (139). Alternatively, on Kohl’s reading freedom of choice requires “a causally unconditioned, absolutely contingent spontaneity” (157), for only then are “metaphysically real alternative possibilities” available to the agent.Perhaps Kohl is correct about the philosophical point concerning what free choice under moral imperatives requires. What strikes me as implausible about his reading here is the position that applicability of norms requires absolute contingency—a complete indeterminacy with regard to whether one chooses to act for or against the moral law. There are at least two problems with ascribing this view to Kant.First, there is the well-known issue that indeterminacy with respect to the grounds of action makes free action indistinguishable from chance. To the extent that Kohl addresses this objection at all (neither the word luck nor chance appears in the book in a relevant context), it is to say that while the moral law is taken as a causal law (83), the notion of the concept cause (or cause-effect) here is understood in a purely practical manner, such that “the practical transformation of the theoretical category transcends the strict logical-inferential link between ground and consequence: it ejects the thought that given the ground the consequence cannot but follow” (83). Kohl’s Kant thus construes the causality of the moral law as one where a ground suffices, but never necessitates, its consequent.Readers familiar with the historical development of Kant’s views will notice that the position that causes can be either determining or merely sufficient is largely identical with that developed by Christian August Crusius, a well-known critic of German school rationalism and an important influence on Kant. Crusius is especially concerned to construct an ontology according to which it is coherent to think that although nothing happens without a ground, not all grounds are “determining” in a sense that would eradicate the possibility of an agent doing otherwise (i.e., not ϕ-ing) in the very same context in which they ϕ. Thus Crusius distinguishes between a principle of determining ground and a (weak) principle of sufficient reason in such a way that, he contends, allows that an agent may be the sufficient cause of his action to ϕ, while still allowing that the agent could have not ϕ’d under precisely the same circumstances. Perhaps unsurprisingly, Crusius also construes God’s activity voluntaristically (a position that Kant rejects throughout his career), which is both in line with his conception of free action and importantly limits the scope of what we can hope to know or find intelligible.What is surprising about this ascription of the Crusian view to Kant is that Kant ardently rejects Crusius’s position concerning free will. He says, “Suppose that … the occasion for acting arrives. You … immediately slide in the direction of what is less good, for the [moral] grounds which solicit you do not determine you…. Of what use are [moral] precepts for performing the work of virtue? Actions are the product of chance; they are not determined by grounds” (NE 1:402). Kant wrote this in 1755, but to my knowledge there is no published work of his in which he repudiates this objection. There are occasional unpublished handwritten notes where Kant seems to flirt more seriously with the Crusian position, but they are just that, notes. In contrast, Kant’s published work seems to emphasize freedom, spontaneity, and autonomy as relying on an agent’s being the appropriate causal source of their activity, rather than it being absolutely contingent as to whether the agent acts or not.I want to emphasize that the problem I’m indicating here isn’t the one that Kohl addresses—namely, that Kant introduces a different form of causality to apply to practical activity (though I think one can raise objections here as well). Instead, the problem is the sense in which groundless activity could ever be imputable, since (1) such activity is indistinguishable from chance, and (2) what is the result of chance is not itself imputable.Second, the interpretive plausibility of Kohl’s position is further eroded by the fact that he fails to appreciate Kant’s Critical idealist distinction between determination, as characterizing the nature of the relation existing between a ground and its consequent, and a particular species of such determination, what Kant calls “predetermination.” The latter, but not the former, is essentially temporal in nature. As Kant puts it in the Religion, “[spontaneity, and thus freedom] is at risk only with predeterminism, where the determining ground of an action lies in antecedent time, so that the action is no longer in my control [Gewalt] but in the hands of nature, which determines me irresistibly” (Rel 6:49–50n; see also A533/B561, MM 6:280n).Kohl presents Kant’s position as a libertarian incompatibilist as so obvious as to be not worth questioning. Indeed, he treats it as a desideratum on a plausible interpretation of Kant’s Critical philosophy—“We must read Kant as endorsing an incompatibilist conception of freedom” even if this freedom, construed as “causally undetermined non-temporal action” is “somehow compatible with natural determinism” (16). In my view, failing to differentiate between forms of determination, and subsequently the relevant sense of compatibility in play, gets Kohl on the wrong track from the outset. To the extent that Kant endorses any form of ‘incompatibilism’ as it is thought of in contemporary philosophy, it is an endorsement of the view that the necessitation of events by their temporal antecedents is incompatible with the kind of free action that grounds imputation and attitudes of praise and blame. But this does not at all entail that he thinks of freedom as incompatible with any form of necessitation (e.g., necessitation by one’s essence or nature). To contend that Kant endorses the latter requires significantly more argument than Kohl (or most other scholars) provides. The general tendency of Kant scholarship to ignore this distinction between forms of determination has resulted in an otherwise avoidable lack of clarity in the reception of Kant’s views on freedom and agency.Finally, it is worth making one other general remark concerning the argument of the book as a whole. Since it is an argument concerning the defense of a particular “self-conception,” we need to ask who holds this self-conception, and whether opponents to philosophical defenses of it are interested in offering genuine alternative positions as opposed to showing why such a self-conception is fundamentally mistaken rather than defensible on other grounds. I would have thought that Kant’s greatest opponent is not the advocate of naturalistic compatibilism but rather the advocate of jettisoning such a self-conception altogether (whether on naturalistic, compatibilist, or other grounds). By emphasizing so heavily the reading of Kant as a defender of a particular self-conception rather than of what is true or knowable überhaupt, Kohl’s reading leaves Kant particularly vulnerable to such skeptical objections.I want to close by emphasizing that Kohl’s book has many virtues and laudably addresses the connections between Kant’s metaphysics and his conceptions of rationality and freedom. This allows for a much better sense of Kant’s philosophy than is typical of the Anglophone reception of Kant. Nevertheless, I think Kohl’s attempt to read Kant as but one more interlocutor in our contemporary philosophical conversation about freedom pushes his arguments in the wrong direction and fails to take account of some of the more radical and intriguing aspects of Kant’s views on freedom and rationality.