¹ So Kant believed: see the originality claim at PFM 375n. and the related claim at A120n. (about being the first psychologist ever to think that imagination might be a necessary ingredient of perception itself). In this essay, I shall employ the following abbreviations for the standard editions: A–/B– (Critique of Pure Reason), CPrR [Critique of Practical Reason), CJ (Critique of Judgement), PFM (Prolegomena to Any Future Metaphysics), Logic (the Jäsche text), MFPNS (Metaphysical First Principles of Natural Science), Ak. (Akademie Ausgabe), THN (A Treatise of Human Nature), and EHU (An Enquiry Concerning Human Understanding) and ECHU {Essay Concerning Human Understanding). All pagination in Kant's writings is that of Ak., though occasionally I shall refer to texts by section numbers.

² This phrase was suggested to me by a reader of an earlier draft of this paper who advocates the position opposed to mine. Paul Guyer conveys the idea well when he compares the role of the psyche in Kant's theorizing to hardware on which a computer program is run: what does it matter whether the Kantian ‘software’ runs on electronic ‘hardware’, ‘brainware’, ‘mindware’, or whatever? See ‘Psychology and the transcendental deduction’, in Förster, E. (ed.), Kant's Transcendental Deductions (Stanford, CA: Stanford University Press, 1989).Google Scholar

³ More specifically, a psychologists fallacy is committed when one confuses the actual processes of thought, and the natural laws which govern them, with the logical, grammatical and other normative rules which constitute (wholly? or only in part?) the difference between thought and other species of mentation. For a history of the use of the term ‘psychologism’ and the negative connotations that came to be associated with it, see Kusch, Martin, Psychologism: A Case Study in the Sociology of Philosophical Knowledge (London: Routledge, 1995).Google Scholar

⁴ Cited from Posthumous Writings, ed. Hermes, Kambartel and Kaulbach, by Lanier Anderson in ‘Neo-Kantianism, normativity and psychology’ (part of a larger paper in preliminary stages of publication).

⁵ I expect that Kant would have regarded mathematical logic, down to and including the calculus of truth-functions, as a species of mathematics, falling outside the province of logic as he demarcated it; if so, then it would have to be considered part of the explanandum, not the explanans, of transcendental philosophy (the problem of the possibility of synthetic a priori judgements). I examine this topic in detail in a work still in manuscript stage, entitled ‘Time out of mind: self and understanding in Kant and British empiricism’.

⁶ See B131n. and Ak. 20:271.

⁷ Even the logical functions are held to be dependent on the subjective constitution of the human mind: ‘Of this peculiarity of our understanding, that it can bring about unity of apperception only in this precise manner and through this exact number of categories, the reason can no more be stated than why we have precisely these and no other functions of judging or why time and space are the unique forms of our possible intuition’ (B145–6); ‘Other forms of intuition (besides space and time), as well as other forms of the understanding (than the discursive forms of thought, or of cognition through concepts), even though they should be possible …’ (A230/B283). Texts such as these make clear that Kant did not feel he was in a position to deny the possibility of sensibly conditioned intelligences with logical functions different from those listed in the table of judgements — anymore than he could deny the possibility of different forms of intuition than space and time. The clear implication is that the fundamental logical functions of judgements Kant deemed definitive of our understanding (see A69/B94 and A81/B106) — categorical subject-predicate form, hypothetical form, universal, affirmative, et al. — are dependent on the special constitution of our minds in the same sense of ‘subjective’ he applies to space and time and to sight, hearing, etc. (see A28–9/B44, C; 189, and Ak. 20: 268). Many resist this implication (see e.g. P. F. Strawson ‘Sensibility, understanding, and the doctrine of synthesis’, in Forster, E. (ed.), Kant's Transcendental Deductions, pp.70–2Google Scholar yet, their resistance seems to be due not to the absence of convincing textual evidence (B145—6 and A230 could hardly be clearer or more specific) but to prior anti-psychologistic commitments related to protecting logic from psychological intrusions.

⁸ See reference in n. 4.

⁹ What all these views have in common is a tendency to discount or explain away Kant's insistence that metaphysics, beginning with transcendental philosophy, is something altogether beyond the purview of conceptual analysis, and is rather fully as ampliative a science as Euclidean geometry or Newtonian physics (see A5–6/B9–10, A10/B13–14, and B23). More generally, there has long been an extreme reluctance to accept the viability, must less the necessity, of the ‘synthetic a priori’ as a class of proposition, even within the domain of critical philosophy (to which Kant in effect restricted it: see PFM 270). ‘Transcendental argument’ is a particularly egregious example. This species of analytical argument came to be so named as a well-intentioned homage to Kant, in the belief that reasoning from the necessary conditions of our thought of objects to the conditions of the objects themselves is the engine driving the ‘austere strictly transcendental philosophy’ that remains after everything psychological has been stripped away. The fallacy implicit in this species of reasoning has been well described by Barry Stroud: ‘What calls into question the validity of the last step of would-be transcendental arguments from the way we think to the way things are is the apparently simple logical observation that something's being so does not follow from its being thought or believed to be so. Something's being so does not follow from everyone in the world's believing it to be so, from everyone's fully reasonably believing it, even from every reasonable person's being completely unable to avoid believing it … That we all think things are a certain way is one thing; that things are a certain way is another.’ ‘Kantian argument’, Parrini, P. (ed.), in Kant and Contemporary Epistemology (Dordrecht: Kluwer Academic Publishers, 1994).CrossRefGoogle Scholar This is quite correct; and Kant, who always insisted on the need to distinguish the indispensability of concepts and principles to our thought of objects from the question of their actual objective validity (see PFM 258–9) would have criticized the reasoning that so many attribute to him in much the way Stroud does, as the following attests: ‘never can one succeed in proving a synthetic proposition from mere pure concepts of the understanding, e.g. the proposition that everything that exists contingently has a cause. The furthest one could go is to prove that, without this relation, we could not even so much as conceive the existence of the contingent, i.e. cognize the existence of such a thing a priori through understanding. But from this, it does not follow that this same condition is also a condition of the possibility of the things (Sachen) themselves’ (B289; see also B168); indeed, Kant went so far as to assert that ‘One can ascribe all illusion to taking the subjective condition of thought for a cognition of the object’ (A396). From this, and similar texts (e.g. A129, B138, B168, A221, A243–6, and A400–1), it seems clear that Kant would never have dreamed to argue against the Humean sceptic in so patently fallacious a manner as ‘transcendental argument’. (Admittedly, other texts, such as A213/B259–60 and A346–7, appear to exhibit a contrary tendency; yet, when subjected to scrutiny, it becomes evident that in none of these does Kant draw conclusions of the kind he deemed ‘synthetic a priori’ solely on the basis of conceptual analysis, without also invoking our immediate, pre-conceptual consciousness of the manifold of sense in pure intuition — the first, and indispensable, condition under which synthetic a priori judgements are possible: see B73.)

¹⁰ See n.l above. Kant also claimed originality for recognizing the need to demonstrate the applicability of pure concepts to objects that can only be given a posteriori, and devising the method of transcendental deduction to meet it (see PFM 260–1). And it is true: the method of transcendental deduction is undoubtedly the most important and innovative aspect of his psychologism. Nevertheless, the problem it was devised to solve is, in my view, self-generated: a consequence of the psychologistic accounts he gave of the origins of pure space and time in the aesthetic and of the categories in the metaphysical deduction. The same is true, I believe, for Kant's many other innovations (including his account of freedom: see n. 11): all presuppose the specific context established in the transcendental aesthetic together with the problems and challenges peculiar to it.

¹¹ ‘The system of the critique of pure reason turns on two cardinal points: as system of nature and of freedom, one leading with necessity to the other. The ideality of space and time and the reality of the concept of freedom, the first leading inexorably and analytically to the second. According to the one, synthetic-theoretical cognition a priori; according to the other, synthetic-practical cognition, likewise completely a priori’ (Ak. 18: §6351, pp. 1796–8). It is unfortunate how little stress is placed on the presupposition of the transcendental idealist context in Kant's moral philosophy in recent secondary literature on the topic.

¹² That pure intuition (the space and time of the transcendental aesthetic) is the product of the spontaneity of imagination rather than the receptivity of sense is the view I have advanced in earlier writings, beginning with Kant's Model of the Mind: A New Interpretation of Transcendental Idealism (Oxford: Oxford University Press, 1991). Readers with an interest in a detailed examination of Kant's doctrine of pure intuition are advised to consult that work and two follow-up papers published in the Review of Metaphysics (1993, 1995). For purposes of this paper (the thesis that Kant's transcendental philosophy is simply a system of a priori psychologism), it suffices to recognize that the doctrine of pure sensible intuition is the foundation for everything that follows in the transcendental philosophy (e.g. no transcendental logic is possible without transcendental aesthetic, as stated explicitly at A76–2/B102; also A15–16/B29–30 and A55/B79–80). For if this is admitted, then we cannot afterwards treat transcendental aesthetic as an episode within the normative epistemology with which we may wish to assimilate transcendental logic; we are, on the contrary, obliged to accept it at face value as a theory of the possibility of sense perception prior to and independently of the conditions for uniting the resulting perceptions in a single, unified experience.

¹³ I have adapted this phrase from EHU IV/ii. 33, where it is used in reference to the principle of the uniformity of nature. Since this principle presupposes the customary associations that, according to Hume, yield the idea of necessary connection, and so too that of cause and effect, it seems to me fair to say that, on Hume's view, all the most fundamental and important ideas for human and animal cognition, including those that are the focus of philosophers’ concerns (cause, necessity, existence, self, etc.), are acquired very early in life. I see Kant's transcendental philosophy in much the same terms: it describes the ‘original acquisitions’ (Ak. 8:221–3) of the human mind in its state of nature, prior to, and as condition for, all learning and acculturation. A detailed working out of this view will be given in the work mentioned in n. 5 above.

¹⁴ At A725/B753, Kant stresses the need ‘to investigate the origin of the pure concepts of understanding and in so doing determine the extent of their validity’. See also A763/B791.

¹⁵ It should be stressed that psychologism, so construed, only claims that an understanding of a representation may (not that it inevitably must) help us to determine (at least) some of the indispensable elements of its sense. For in many contemporary accounts, the notion that the nature and workings of our minds may be the source of some of the essential constituents of thought often gets distorted into the thesis that thought can be explained completely in terms of ideas and their relations. Since thought includes language, the inability to explain the syntactical, semantical and other (similarly normatively constituted) features of thoughts in terms of the individual, isolated psyche has given wide currency to the view that early modern philosophers are guilty of committing the ‘psychologistic fallacy’. As Hacking, Ian pointed out in Why Does Language Matter to Philosophy} (Cambridge: Cambridge University Press, 1975)CrossRefGoogle Scholar, this is a caricature even of Locke's views on language: the most commonly adduced example of the misguided psychologistic approaches to language that supposedly flourished during the early modern period. What one should realize is that Locke, far from being concerned to offer a theory of language capable of reckoning with its syntactic and semantical features, merely pointed to language as the only means at our disposal to communicate our ideas to one another. For, so considered, Locke's position is perfectly compatible with the existence of other, non-psychological (normative) conditions of language (and perhaps artificial sign systems generally). This is still more evident in Hume (whom, along with Kant, Hacking does not consider), whose careful treatment of language in THN III. gives full due to the dependence of language on conventionally grounded norms, which cannot be reduced to, or eliminated in favour of, the natural operations of the individual, isolated human understanding (see ch. 3C of my book, Hume's Theory of Consciousness, Cambridge: Cambridge University Press, 1994). In general, early modern thinkers up to Kant (by contrast with many of their nineteenth- and twentieth-century successors) did not suppose that a satisfactory account of mental representation can double as an account of linguistic representation.

An example of another kind of widely current error is the notion that the British empiricists relied primarily on introspection, and endeavoured to base their philosophizing on the supposed infallible certainty of private exhibitions. For, setting the more problematic case of Berkeley aside, neither Locke nor Hume can be suspected of having placed much trust in introspection, and indeed virtually all of their reasoning is extrapolative in character. Thus, the questionable character (or impossibility, as implied by the private language argument of Wittgenstein) now widely attributed to private exhibitions of concepts is far less obviously a problem for Hume than it is for later philosophers such as Husserl, James and Russell.

¹⁶ I do not include the transcendental ideas of reason here because my focus in this paper is restricted to transcendental analytic: Kant's logic of truth, not of the semblance (Schein) of truth (= dialectic). My inclusion of the ‘I think’ is based on the following authority: ‘[it] was not classified in the general list of transcendental concepts but nevertheless must be counted among them … One sees easily … that it is the vehicle of all concepts in general, thence also of transcendental concepts, and thus that it is always conceived along with these too and is just as transcendental as they. But it can have no particular title because it serves to present all thinking as belonging to consciousness’ (A341/B399). As we shall see in more detail in part II (to be published in the next issue of Kantian Review), the ‘I’, construed as analytic unity of apperception, is not just the psychological source but the psychologistically determined meaning of the concept ‘logical universality’.

¹⁷ Hume's investigation of the origins of ideas to determine their contents should not be confounded with the patently causal account of causal inferences (experience causes customs which cause us to draw inferences). The implicit, and often explicit, use of causal notions in Hume's science of human nature can be resolved in the cash value of the introspectible contents of customary transitions of thought, and most particularly of the imagination-immanent affects of facility and vivacity. Kant, by contrast, made little use of affects immanent to mental activity in his a priori psychology of cognition. The one exception I have been able to identify is the feeling he deemed indispensable for the communicability of cognition in C), §21.

¹⁸ See also A44/B61–2 (‘subjective constitution’, plus the parallel remark at PFM 290 stressing that the counterpart to a logical difference of the sort in Leibniz and Wolff is ‘the genetic one of the origin of cognition itself’), A78/B103 (the goal is to ‘determine the first origins of our cognitions’), A128 (‘origin and truth’), A195–6/B240–1 (how we are ‘first led to make for ourselves the concept of cause’), A204/B249 (critique is concerned ‘with the sources of synthetic a priori cognition’), A270–1/B326–7 (‘understanding and sensibility [are] two sources of representations’), A725/B753 (the demand is ‘to investigate the origin of the pure concepts of understanding and in so doing determine the extent of their validity’) and A758/B786 (Kant's critical investigation of ‘the primary sources of our cognition’).

¹⁹ ‘Whether concepts are mere educta or producta. Preformation or epigénesis, producía either through physical (empirical) influence or through consciousness of the formal constitution of our sensibility and understanding on the occasion of experience, hence still producía a priori, not a posteriori. The doctrine of innate ideas (ideis connatus) leads to fantasizing (Schwärmerei). Aqvisitae are a priori or a posteriori acqvisilae. The former are not always intellectual. Thus, the division of cognition into sensitive and intellectual is not the first, but rather that into a priori and a posteriori’ (Ak. 18: §4851, p: 1771; see also §4859). The preformationist ‘middle course’ described at B167–8 thus seems to be a version (attributed to Crusius at Ak. 17: §§4275) of the innatism of educta.

²⁰ It has been objected to me that Kant insists that ‘we are not talking about the origin (Entstehen) of experience’ (PFM §21a). But this criticism would be valid only if I had referred to texts in which ‘Entstehen’ is mentioned to support my claim that origins are Kant's prime concern in transcendental philosophy. However, unintentional errors aside, I would only rely on texts with ‘Ursprung’ or ‘Quelle’ to support my claim. For no translation must be allowed to obscure the fact that these are two totally different notions:. ‘Entstehen’ is c-oordinate with ‘Vergehen’, but ‘Ursprung’ and ‘Quelle’ ate not. The PFM text my critic references seems merely to be an echo of Bl, where Kant contrasts what begins (‘anhebt’) with experience to what originates (‘entspringt’) from it, where the latter is identified as Kant's primary concern in the Critique (‘entspringen’ clearly betokens origins, that is, ‘Ursprung’). In both cases, Kant's phraseology implies a rejection not of psychology per se but simply of empirical psychology.

²¹ In responding to his critic, Kant embraces faculties innatism, while explicitly rejecting representational innatism: ‘This first formal ground alone, e.g., of the possibility of a space intuition is innate, not the space representation itself’ (On a Discovery, Ak. 8: 222). In Kant's Model of the Mind (ch. ID. and 3), I related this text to a passage in the Critique in which one finds the form of intuition characterized as a ‘subjektive Beschaffenheit der Sinnlichkeit’ and formal intuition equated with ‘time and space’ themselves (A267–8). This seems to me to coincide exactly with what Kant is saying in the On a Discovery text with respect to space and time, and the same difference applies to the case of the categories and their relation to the (innate) logical functions of judgement which are constitutive of the understanding as a capacity to judge or think (see n. 22).

²² Kant uses this expression at A81/B106, where he identifies the capacity to think with the capacity to judge, which he had earlier used to define the understanding exclusively in terms of the logical functions of judgement (seeA69/B94).

²³ ‘Discursivity’, in Kant's usage, always betokens representation by means of universals, that is, the use of a representation as a conceptus communis. Many interpreters suppose that he used this expression with reference to the imagination or its products, but, to my knowledge, this was never the case (nor would it be consistent with his view that imagination is not a faculty of concepts: see A78/B103).

²⁴ This is where the expression Vermogen zu denken recurs, thus linking this text quite particularly to Axvii via A81/B106.

²⁵ The recurrence in the B edition of the A edition focus on the possibility of the understanding represents a major stumbling block for those scholars who deny, or doubt, that the distinction between an objective or subjective deduction carries over from the A edition to the B edition. The primary basis on which to form a judgement is the lengthy footnote in the MFPNS preface, in which Kant publicly announced his intention to rewrite the part of his project concerned to explain ‘the manner how experience first becomes possible’ by discovering the ‘underlying a priori principles of the possibility of thought itself’ (Ak. 4: 476). This ‘how’ question seems to coincide quite precisely with the A edition characterization of the subjective deduction as ‘so to speak, a search for the cause of a given effect’. Still more tellingly, in the MFPNS preface footnote, the how question is treated as inessential by contrast with another question that seems to correspond exactly with the task assigned to the objective deduction in the A preface: proving that ‘all employment of pure reason can never concern anything but objects of experience, and, because nothing empirical can be the condition in a priori principles (Grundsätzen), they can be nothing more than principles (Prinzipien) of the possibility of experience in general. This alone is the true and sufficient foundation (Fundament) of the determination of the boundary of pure reason, but it is not the solution of the problem: ‘how’ experience is possible by means of these categories and only by their means. This latter task, although the building stands firm without it, nevertheless has enormous importance and, as I now realize, equally great ease, since it can be carried out almost by a single inference from the precisely determined definition of a judgement in general (an act through which given representations first become cognitions of an object). The obscurity which attaches to my previous treatment in this part of the deduction, and which I do not disclaim, is attributable to the usual fate of the understanding in inquiry, that the shortest route is not the first that comes to notice.’ Since this is just to say the material added in B (B129–69) to replace the exposition in A (A95–130) is concerned with a how question that is not strictly essential to the Critique's chief task of establishing the boundary of pure reason, it seems to follow that the subjective deduction, far from being downplayed or omitted in B, was the specific concern of Kant's revision. Accordingly, I shall treat the B edition transcendental deduction as a subjective deduction concerned with ‘the possibility of the understanding, even as regards its logical employment’ (B131).

Where then is the objective transcendental deduction to be found? Kant tells us in the A preface that, so far as the objective deduction is concerned, ‘what is said on pages 92 to 93 [A pagination] alone may be sufficient’: it could well be the case that this passage is the objective deduction, so that what follows ‘is not strictly essential’. Though the textually most straightforward answer, many interpreters baulk at the suggestion that everything after the section which this passage brings to an end (A95–130 and B129–69) is inessential to the transcendental deduction of the categories. Yet, such resistance may diminish if we recall that, in the MFPNS footnote, Kant makes clear that the essential purpose of the Critique can be achieved if three things are conceded to him, coinciding more or less exactly with the doctrines of the transcendental aesthetic (concession 3), metaphysical deduction of the categories (concession 1), and the objective transcendental deduction (concession 2, as furnishing the principle of possible experience on which the transcendental schemata and principles of the analytic of principles depend). Viewed in this light, the claim of the objective transcendental deduction all but falls out as a corollary of the metaphysical deduction, which in turn is built on the foundation laid in the transcendental aesthetic (see A76–7/B102); for, in essence, the objective transcendental deduction merely extends to empirical synthesis (perceptions as combinable in one experience) what the metaphysical deduction asserts with respect to pure synthesis. Why the subjective deduction, although not strictly essential, is of the utmost relevance to Kant's purpose will emerge shortly.

²⁶ The term ‘metaphysical’ attached to space and time should be understood in the light of Kant's Comments on a Dissertation by Kästner, Ak. 20: 419–21 a passage, space permitting, I plan to examine in part II of this essay.

²⁷ For a more detailed consideration, see my article, ‘Kant on the possibility of thought: universals without language’, Review of Metaphysics, 48 (June 1995), 809–58.