Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise (review)

After a lifetime of distinguished work on Kant (with important books on Lessing and Spinoza along the way), notable for its meticulous scholarship and deep sympathy with its subject, Henry Allison has now produced a detailed and impressive study of the first book of Hume's A Treatise of Human Nature. Having subtitled his chief work on Kant's Transcendental Idealism "an interpretation and defense" (Yale University Press, 1983, revised edition 2004), he could not very well use the same subtitle for his work on Hume, for on some central matters no one could plausibly defend both Kant and Hume. Instead, he subtitles this work "a Kantian reading," but he gives a more accurate account of his intentions in the Introduction, where he says that his work "provides a critical analysis of Hume's views from an avowedly Kantian perspective. The result is both a defense and a critique of Hume" (1). Allison does indeed both criticize and defend Hume. What he criticizes is Hume's perceptual model of cognition, which he does in two steps. First he argues that Kant's conception of cognition as consisting in judgments that apply concepts to intuitions is superior to Hume's model of cognition as consisting in either current perceptual impressions or forceful, lively, and vivacious ideas thereof. And second he argues that the defects of the purely perceptual model of cognition can be recognized from internal difficulties in Hume's own attempts to explain our knowledge of space, of causal connections, and of external objects even without explicit appeal to Kant although in agreement with Kant's own criticisms of those aspects of Hume's theory of knowledge that he knew (basically, just Hume's treatment of causation). What Allison defends is Hume's "metascepticism." This is the view that rather than simply being saved from logically sound skeptical arguments by our natural dispositions to belief, on which such arguments have no effect, we should be just as skeptical of skeptical arguments as those arguments would have us be about the particular sorts of claims to knowledge they claim to cast into doubt, and for that reason adopt an attitude of philosophical "insouciance" toward skeptical arguments. The key difference between these two positions is that while the former is entirely descriptive, simply pointing out that our basic mechanisms of belief are insulated from argumentation altogether, the latter position is normative, telling us that we should hold skeptical arguments up to the same standards to which skepticism would hold up other arguments, and that if we do so we shall actually find skeptical arguments less than compelling. If I have any criticisms of Allison's work, they are that he spends more time than is necessary on the critique of Hume's perceptual model of cognition, which after all has long had no serious defenders, that he does not present this critique as perspicuously as he might have, and that he does not develop his interesting conception of Hume's metaskepticism as fully as he might have.

Allison's critique of Hume's perceptual rather than judgmental (or in Allison's Kantian term "discursive") model of cognition is not as perspicuous as it might have been because Allison structures his book as a commentary, simply following Hume's order of discussion section by section, when he might better have ordered his book into two parts, the critique of the perceptual model of cognition followed by the defense of metaskepticism, and then brought Hume's various topics into this framework as relevant. As it is, Allison follows Hume's order of exposition in offering us chapters on (1) "Hume's Elements" (impressions and ideas, the Copy Principle and the Separability Principle); (2) "Hume's Theory of Space and Time"; (3) "Hume's Epistemological Divide" (Hume's fork, which Allison argues is not identical to Kant's distinction...