Kant’s Transcendental Psychology. – book reviews

Kant’s Transcendental Psychology. – book reviews

Paul Guyer

In this work Patricia Kitcher proposes to revise the underlying assumptions that have dominated Anglo-American (and a good part of German) interpretation of Kant’s theory of knowledge at least since the publication of Strawson’s The Bounds of Sense and Bennett’s Kant’s Analytic (as well as D.P. Dryer’s Kant’s Solution for Verification in Metaphysics) in the annus mirabilis of 1966. She aims to do this by rejecting the exclusive emphasis on transcendental arguments with anti-sceptical import that has characterized so much of this work and instead restoring the respectability of Kant’s transcendental psychology, which was last taken seriously, although not very well understood, by early nineteenth-century Kantians such as J.F. Fries; the reward for this effort, she argues, will be not only an historically more accurate as well as philosophically more rewarding interpretation of Kant, but a suggestive approach to research programs in contemporary cognitive science (particularly cognitive psychology) as well. In the end, however, Kitcher does not prove that there is any fundamental philosophy and transcendental psychology; on the contrary, her exposition of Kant’s transcendental psychology offers a model for anti-sceptical transcendental argumentation which shows both how it works but also what its reasonable and inevitable limitations are. So there is some confusion here, but, I think, it is fruitful confusion. (That I think this will not be a surprise: in spite of differences in detail and terminology, including perhaps our understanding of the term “psychology” itself, the basic assessment of the strengths and weaknesses of Kant’s argument for the objective validity of the categories which Kitcher reaches is not very different from that which I have offered in my own works; see Kant and the Claims of Knowledge, (Cambridge: Cambridge University Press, 1987, pp. 417-28], especially the Afterword.)

Both the confusion and the strength of Kitcher’s approach lie in her basic model of transcendental pshchology. On Kitcher’s account, Kant never commits the error of “psychologism”, that is, assuming that normative principles can be directly derived from factual premises, but he does offer an epistemology which is “weakly psychologistic” in the positive sense of recognizing that we can only “hope to understand the nature of thought or the limits of knowledge–or to prescribe methods for improving our reasoning practices–[on the basis of] some understanding of the capacities that make cognition possible” (p. 9). Transcendental psychology is then the abstract investigation by essentially reflective means of the capacities necessary for cognitive accomplishments we take ourselves to be capable of, leaving more detailed investigation into the particular manner of the realization of such capacities for empirical disciplines such as cognitive psychology. Thus, transcendental psychology consists of “analyses of cognitive tasks.” In more detail, it is comprised of two different types of arguments: first, what Kitcher calls “epistemic analyses,” “Arguments showing that we must have a certain type of faculty (i.e. something in the constitution of the mind that supplies a [priori.sub.omicron] elements) if knowledge, or a particular kind of knowledge, is to be possible”, and “analyses of empirical capacities”, which are “Arguments showing that, given an account of the cognitive tasks that we perform that would be accepted by all parties to [a] dispute, it follows that we possess one or several of the types of faculties noted in” an epistemic analysis (p. 19). Kitcher distinguishes three senses of “a priori” in Kant: a [priori.sub.L], or the logical a priori, which is simply the claim to universality and necessity; a [priori.sub.omicron], meaning that a proposition or concept contains elements that do not originate from experience: and a [priori.sub.Kappa], implying the validity of a knowledge-claim independently of empirical confirmation (pp. 15-16). I find the distinction between the latter two senses obscure, though I’m not sure how much this obscures any of Kitcher’s argument. In other words, an “epistemic analysis” is an argument yielding a conditional conclusion of the form that if we are capable of a certain sort of knowledge-claim or cognitive performance then we must be possessed of certain a priori forms of thought (e.g. forms of intuition, concepts, or principles), and an “analysis of empirical capacities” is an argument that shows to the satisfaction of all concerned that we do indeed have the cognitive ability which entails the priori element argued for in the “epistemic analysis”. But, although Kitcher does not make this explicit, this model can apply to two different types of case. On the one hand, it can simply offer an explanation of a particular cognitive ability when it suggests that a certain cognitive accomplishment C which an “analysis of empirical capacities” shows us to have presupposes a certain a priori element O; on the other hand, it may show us more than that in a case in which the a priori element O, demonstrated to exist by the “analysis of empirical capacities” which proves to us that C obtains and the “epistemic analysis” which shows that O is the condition of C, also implies the validity of some further cognitive capacity or accomplishment D which might not have been expected by someone or even anyone accepting C and investigating its condition O. The first sort of case, which Kitcher initially seems to have in mind, suggests the usefulness of transcendental psychology for cognitive science: the abstract characterization of the a priori condition O of some empirical capacity C provides a target for more concrete empirical research into its psychological or even ultimately physiological basis. But the second sort of case suggests that transcendental psychology as Kitcher understands it can provide a model for anti-sceptical transcendental argumentation: one argues with a sceptic about cognitive capacity D by showing him that some other cognitive capacity C which he does not doubt presupposes a priori element O which, to his surprise, also suffices to ground cognitive capacity D. At the same time, of course, this model also shows the limits of transcendental argumentation: if the sceptic is prepared to doubt C as well as D, that is, to reject the “analysis of empirical capacities” which is supposed to demonstrate C, then the argument that O also implies D will not move him. Of course, if C seems sufficiently plausible, the logical possibility that someone might doubt it will not be particularly worrisome; thus transcendental arguments may be powerful without promising to refute every conceivable sceptic. This is how it stands, I would suggest, with Kant’s own most promising transcendental arguments, such as the second Analogy of Experience and the Refutation of Idealism.

Kitcher’s work is interesting, if not as clear as it might have been, precisely because she exploits both of these models for transcendental psychology. The first model is most evident in her discussion of the “science of sensibility” (Ch. 2), where she is primarily concerned to argue that Kant’s epistemic analysis of the a priori forms of intuition shows that an abstract “pure process form” or capacity for construction must underlie the “pure product form” or a priori components necessary to explain how we generate a three-dimensional model of space from two-dimensional retinal stimulations, which “pure process form” would then require further empirical investigation. But the second model takes hold in her lengthy and often quite rewarding discussion of the transcendental deduction of the categories, which is spread out over chapters 3, 4, and 6. At first, it might seem as if her strategy will be confined to the “regressive” method which has been stressed by Karl Ameriks, in which the point of a transcendental argument is taken to be only that of showing, against an empiricist rather than a sceptic, that certain cognitive abilities presuppose certain a priori concepts. But as she proceeds and approaches the “long argument” which constitutes her interpretation of the final form of the deduction, Kitcher makes it clear that she takes the deduction to use the progressive as well as the regressive method; in her words,

Kant’s best argument for the categories has subjective and objective elements; it is both regressive or analytic (because it presupposes some cognitive capacities) and progressive or synthetic (because it argues from the presupposition of fairly minimal capacities to very strong claims about the categories). (p. 169)

This is to say that an “epistemic analysis” in transcendental psychology can reveal that we have greater cognitive capacities than are demonstrated by the “analysis of empirical capacities” which initially demonstrates that the antecedent of the conditional analysis is fulfilled. In other words, transcendental psychology can yield anti-sceptical transcendental arguments.

Space does not permit a detailed review of the contents of Kitcher’s work. It will have to suffice to say that in spite of her initial disdain of interpretations which hold that the transcendental deduction is a mere preparation for the argument of the Principles of Empirical Judgement, especially the Analogies of Experience (p. 62), her “long argument” demonstrates that in the end Kant attempts to connect his concept of apperception with the objective validity of the categories only through an epistemic analysis of the conditions of perceiving spatial and temporal arrays, an analysis which is suggested, primarily by example, in section 26 of the B-deduction but only carried out in the Principles. This account thus ends up being similar to my own, which Kitcher generously acknowledges. In what may be the most original part of her work, however, she suggests that it may be possible to ground at least the most central of Kant’s categories more directly in the argument of the deduction than this sort of interpretation does. This would be by means of her analysis of the concept of synthesis itself, where a synthetic representation of an object is shown to be a representation (recognized as) preserving aspects of the content of earlier representations which in turn represent states of affairs other than representations themselves (see Ch. 4, e.g. pp. 110-11, as well as Ch. 5, e.g. pp. 177-18). In fact, although at this stage of her exposition Kitcher talks only of abstract “functions of synthesis” with unspecified rules, this is a thoroughly causal analysis of synthesis, demonstrating that we can comprehend a representation as a synthesis of prior representations only by tracing out causal connections between later and earlier representations and then between the latter and objects. By this means at least the central category of causation could be argued to be implicated in the very notion of synthesis itself, not just in the further details of perceiving spatial and temporal arrays. (By this means Kant could also be shown to be attempting to answer Hume at the most basic level, showing that Hume’s own conception of the relation between impressions and ideas is incomprehensible unless he is willing to concede the indispensability of causal connections; Kitcher suggests as much at pp. 95-6.)

Among further topics, Kitcher’s treatment of the Paralogisms (chapter 7) is extremely useful, stressing that Kant needed them not merely to correct the errors of previous metaphysicians but also to warn against misinterpretation of his own conception of apperception; and her discussion of Kant’s conception of systematicity as a scientific heuristic (chapter 8) is a suggestive treatment of a topic that was to become of increasing importance to Kant himself in the final decade of his career. All in all, there is much to reflect on in this work, and no serious student of Kant’s and/or Kantian epistemology could well afford to ignore it.

COPYRIGHT 1993 Oxford University Press

COPYRIGHT 2004 Gale Group