Subjectivity in Descartes and Kant Author(s): Hubert Schwyzer Source: The Philosophical Quarterly, Vol. 47, No.

188, (Jul., 1997), pp. 342-357 Published by: Blackwell Publishing for The Philosophical Quarterly Stable URL: http://www.jstor.org/stable/2956419 Accessed: 20/05/2008 19:08
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=black. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We enable the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

http://www.jstor.org

ThePhilosophical Vol. Quartery, 47, No. 188 ISSNoo3I-8094

July 997

SUBJECTIVITY IN DESCARTES AND KANT

Kant's critique of Cartesian scepticism is often characterized in the following sort of way. Descartes represents our inner life, our subjectivity, as if it were something independent and unsupported, as if our (my) conscious states, our thoughts and experiences, could somehow be the whole of what is real - without requiring the reality of anything else, and in particular without requiring the reality of the objects our thoughts and experiences. And of Kant argues, we are told, that the one reality presupposes the other, that subjectivity is impossible without objectivity, that a necessary condition of our having thoughts and experiences at all is that the objects of those thoughts and experiences actually have a certain character. And it is pointed out that Kant argues, in the Principles, that experience would be impossible if its objects were not in their own right quantifiable, substantial, causally inter-related, and so on. I think this picture of Kant's response to Descartes is widely held. I shall call it 'the standard picture'. I know of no one who has spelt it out more carefully and worked out its consequences and difficulties more powerfully than Barry Stroud.' In this paper I shall for the most part focus on Stroud's version of the picture. I do not, of course, wish to reject the picture entirely. It is surely correct to say that Kant means to argue that experience would be impossible if its objects were not really thus and so. My objection to the picture is that it presents Kant as subscribing, or coming very close to subscribing, to
I 65 Originally in 'Transcendental Arguments', Journal of Philosophy, (1968); later in The of Scepticism(Oxford UP, 1984); and most recently in 'Kantian Significance Philosophical Argument, Conceptual Capacities, and Invulnerability' ('KACCI'), in Paolo Parrini (ed.), Kant and Contemporary (Dordrecht: Kluwer, I994), pp. 231-51. Epistemology
? The Editorsof ThePhilosophical Quarterl', 1997.Publishedby BlackwellPublishers,Io8 Cowley Road, Oxford ox4 Main Street, Maiden, MAo2148,USA.
IJF,

UK, and 350

SUBJECTIVITYIN DESCARTES AND KANT

343

Descartes' dichotomy of the inner and the outer, the subjective and the objective, and, above all, to Descartes' view of the nature of subjectivity. We have, on that view, two discrete realities. The first is the purely subjective inner world, containing only our conscious states, and, for some who subscribe to the view, ourselves qua conscious beings. This reality is known only from within, and there can be no doubt about what it is really like: it is as it seems to its possessor (thatis its subjectivity).The other reality contains precisely what the first does not; most notably, it contains things in space, with their characteristics.It is not known, directly, from within; only the first reality can be known that way. But it is known, if it is known at all, only via what is known from within. For the Descartes of the first two Meditations, this second reality is, as Kant puts it, 'problematic': there seems to be no valid path to it from the first, from which the only possible path to it must lead. For Kant, according to the picture at issue, the second reality can be known if the first is known, for the first depends on the second; if there were no spatial world there could be no inner world, no thoughts or experiences. The difference seems to be that whereas Descartes, initially, disallows a valid inference from the inner to the outer, Kant insists on it. What the Kart of this picture shares with Descartes is the view that consciousness per se delineates a world of objects complete unto itself, inner objects to be sure, but objects (of awareness) none the less. The only question is whether that world needs the support of another world. I shall argue in this paper that there is for Kant no such inner world of objects initially (or 'directly') apprehended by us, as there is for Descartes. There is no such inner world from which an objective outer world is to be inferred or out of which an allegedly objective world is to be constructed or 'constituted'. Moreover, as I hope to show, Kant has a powerful argument against the possibility of taking one's inner states as any sort of starting-point for metaphysical theorizing. So Kant's critique of Descartes is considerably more radical than the standard picture allows. This does not mean that for Kant there is no such thing as subjectivity, if that means no such thing as a first-personpoint of view on one's experience. What it does mean is that for Kant the first-person point of view does not itself determine a domain of objects of awareness. We shall examine later what the implications of this might be for Cartesian scepticism. But first it will be helpful to look further at what Stroud has to say about Kant. He claims (KACCI p. 233), perhaps rightly, that the demonstration that certain principles of nature (like that of causation) are necessary conditions of the possibility of experience is 'what Kant himself saw as the distinctive and most important payoff of his transcendental philosophy'. But, Stroud asks (p. 234), how is this demonstration supposed to work?
? The Editorsof ThePhilosophical , 1997 Quarterl,

344

HUBERT SCHWYZER

... how can truthsabout the world which appear to say or imply nothing about human thought or experiencebe shown to be genuinely necessaryconditionsof such psychological facts as that we think and experience things in certain ways, from which the proofs begin? It would seem that we must find, and cross, a bridge of necessity from the one to the other. That would be a truly remarkablefeat, and some convincing explanationwould surelybe needed of how the whole thing is possible.

Kant's answer, Stroud tells us (KACCI p. 235), is transcendental idealism. And that means at least this: that the real, non-psychological, world
is not really a world which is in every sense fully independent of all thought and experience. It is a world which, transcendentally speaking,depends on or is 'constituted' by the possibilityof our thinkingand experiencingthingsas we do.

So Kant, on Stroud's view, has paid a heavy price for the licence to build his bridge of necessity, and that is that the bridge does not really reach the other side; all of it, when the last span is lowered into place and the last bolt tightened, remains within the realm of subjectivity. Not that nothing philosophically important has been accomplished. The bridge, in Kant's able hands, may well take us to places we have never before seen, or even dreamt of- to ever deeper levels of the nature of our thinking. We might even come to see what anypossible conception of reality would have to be like. But that will still be, in the last analysis, about conceptions, and so on the 'psychological' side of the divide. And Kant's bridge might also have important consequences for how philosophers can intelligibly theorize. It may be that we can no longer follow Descartes in playing the devil's advocate and supposing 'that all the things which I see are false ... that body, figure, extension, motion, and place are merely figments of my mind' (Second Meditation). For if it is true, as Stroud allows that Kant may have shown, that we haveto think in terms of extended bodies if we are to be able to think at all, then we are incapable of any such supposing. In that case, belief the in an outer world - which is not the same as there actually being outer world is a necessary condition of the very possibility of thought and experience. And this leads, in Stroud, to a subtle and interesting enquiry into whether and how our sceptical urges can be appeased by these sorts of findings.

II Let us turn now more directly to Kant. We might start by noting that in and experience his formulations whereas Stroud constantly conjoins thought in 'necessary conditions of thought and experience') and lumps them (as together as 'our psychology', Kant typically separates them. The conditions
? The Editorsof ThePhilosophical Quarterl, 1997

SUBJECTIVITYIN DESCARTES AND KANT

345

of thought, that is, of all possible judgements, are worked out early in the of Analytic of the Critique PureReason,in 'The Clue to the Discovery of all the Pure Concepts of the Understanding',2 before question of whether our thinking is true of the real world so much as arises. The argument of the Clue is, in outline, as follows. Since, as logic has shown, all thoughts or judgements must have a certain form (subject/predicate, conditional, quantified, etc.), it follows that all thought aboutanything ('about objects') would have to conceive those things in certain ways (as substances and attributes, causes and effects, singularities and pluralities, etc.). This means that Kant takes himself already to have established that we have to think about the world, about any world that can be thought about at all, in certain specific ways. We cannot conceive of things without conceiving them under the categories of the understanding, as substances, as causally inter-related, etc. That, if you like, is our conceptual scheme. That we have to think in these categorial ways is, as we shall see Kant fully recognizing, simply a fact about us. It tells us nothing about what the things that we have to think about in those ways are really like. But there is, for Kant, that further question, which he takes up in the Transcendental Deduction of the Categories, the question, as he puts it, of the 'objective of validity' how we think about the world. That question is not settled by what he takes to be an established fact, that we have to think about it in those ways. This distinction, between the argument of the Clue and that of the Transcendental Deduction, is crucial. It shows that Kant is not content with a 'descriptive metaphysics', a metaphysics that does no more than lay out how we think, or even have to think, about the world. Nor is he reducing metaphysics to psychology, as if the world's being thus and so were itself to be a function of how we think. The Transcendental Deduction is meant to do something quite different from the Clue. We need to see what this new argument, for the objective validity of how we think about the world, actually amounts to. But it might well be thought that that argument in Kant, however it may in fact proceed, is doomed in advance. For either, surely, it will try to show in itself, quite independently of 'our psychology', must that the world as it is our be thus and so, somehow because psychology is thus and so - in which case we have, in Stroud's politely restrained phrase, the 'truly remarkable' and bridge of necessity. Or it is only appearances, not things as they are in themselves, that must be thus and so, as necessary conditions of our psychology, in which case the importance of the distinction we have been
of Critique PureReason,tr. N. Kemp Smith (New York: St Martin's Press, I965), A70/B95ff. (All references to Kant in this paper are to this work.) Commentators often call the Clue the 'Metaphysical Deduction', following a remark by Kant (BI59).
? The Editorsof ThePhilosophical Quarterly, 1997

346

HUBERT SCHWYZER

talking about is radically undermined. The allegedly objective validity of how we think will not really be objective after all; we shall only have learned about the world-as-it-appears-to-us,not as it really is. This is a world relativized to our psychology. So the Transcendental Deduction will either make impossible claims (the remarkable bridge), or it will add nothing substantive to the conclusion of the Clue (the world as we think it conforms to the conditions of thinking). But this is all prejudgement. We have not even begun to see what Kant's argument about objective validity, the argument of the Transcendental Deduction, actually amounts to. But when we turn to the opening pages of the Deduction, our first impression might well confirm our suspicions. For the way Kant sets up the problem makes it look as if he means to equatethe objective validity of the ways in which we have to think about the world (the categories) with their being necessary ways of conceiving what we experience. He begins by noting that we have, with the categories, a problem which we did not have with the forms of sensibility (space and time), viz., the problem of 'how conditions thought can have objective subjective of validity' (A89/BI22, italics origThe categories, as introduced in the Clue, are after all nothing but inal). subjective necessities (this is how we have to think);it still needs to be shown that they are valid of objects - which presumably means true of real things. So far, so good. But the continuation of the very same sentence seems that is, can already to relativize objectivity to us: '... can have objective validity, furnish conditions of the possibility of all knowledge of objects'. And a little later he explains that what needs to be shown is that 'only as thus presupposing them [the categories] is anything possible as object of experience.... The objective validity of the categories ... rests, therefore, on the fact that, so far as the form of thought is concerned, through them alone does experience become possible' (A93/BI26). This certainly has the appearance of reducing objectivity to what is essentially psychological. But we should not jump to that conclusion before seeing the actual argument. How can Kant have supposed that if something is necessary for thought, it is, for all that, only 'subjectively' necessary, whereas if it is necessary for experience it is objectively valid? Well, there is this difference between 'experience' and 'thought' in Kant: 'experience' is defined in terms of objects; 'thought' is not. To think is merely to connect thought-elements ('representations') with one another according to rules of judgement; you are thinking if your representations are combined thus and so (see Table ofJudgements, A70/B95). Experience, on the other hand, is always ofsomething. (This is of course an artificial distinction. Thought too is ofsomething, or has objects: it is representational. Kant's point is that one can abstract that feature from it,
? The Editorsof ThePhilosophical Quarterl!, 1997

SUBJECTIVITYIN DESCARTES AND KANT

347

and consider it purely formally, as one does in logic.) Kant defines experience as 'empirical knowledge of objects'. Now it does not follow from this definition (nor does Kant think it follows) that the mere having of an experience is proof of any kind of mind-independent reality of its 'object'. The object of experience might, for all that, be something that is itself subjective, like an inner state. But though the definition does not of itself solve a problem, it does bring into focus a question central to the argument of the Transcendental Deduction, that of what can be meant by an 'object of experience'. And one of the chief lessons of the Deduction, as I understand it, is that it is wrong to take it as a giventhat a representation (that which is presented to consciousness in experience) has an (intentional) object at all, or that it represents something to its possessor. Kant disputes Descartes' belief that the fact that my representations represent something to me, that my ideas mean something to me, is epistemologically fundamental and unconditioned. This representational capacity of 'our representations' itself needs to be explained; the Transcendental Deduction sets out to explain it. Whether such an explanation succeeds in bestowing objective validity on anything remains to be seen. And that question amounts to this: if Kant has succeeded in showing that without the categories there could be no such thing as mental representation at all - that in their absence nothing whatever could be represented to me, even something inner or subjective - will that have shown that the categories are objectively valid, in a sufficiently robust sense of that phrase? Or will things still seem to be objectionably relativized to us? III The opening premise of the main argument of the Transcendental Deduction (2nd edition) lays down (BI3I-2) a condition for the possibility of a representation'srepresenting anything to me:
for It must be possiblefor the 'I think'to accompanyall my representations; otherwise something would be representedin me which could not be thought at all, and that is equivalentto saying that the representationwould be impossible,or at least would be nothing to me.

Kant is saying here that there is a condition, the accompanyability of the 'I think' (which we have yet to examine), on the possibility of any representato tion, or datum of consciousness, being anything me.And this means, working backwards through the quoted passage, on the possibility of its beinga representation at all, i.e., on its representing anything to me (Kant clearly takes it, rightly, it seems to me, that for an item to represent something is for
? The Editorsof ThePhilosophical Quarterly, 1997

348

HUBERT SCHWYZER

it to represent something to someone); to put it another way, there is this or, condition on anything's entering my thought. Let us pause on this - not yet on the condition, but on what the condition is a condition on. Kant's claim is completely general. It means that I cannot represent something as, say, cheese, unless the condition is met; but it also means that I cannot represent unless the condition is met. I cannot rethis as a (or my) sensation cheese of present myself as walking oras thinking that I am walking, or even simply as thinking, or for that matter as being in any state whatever. But this seems odd. Does not Kant's formulation of the condition clearly exempt the representation of myself as thinking from being itself subject to the condition? If it must be possible for the 'I think' to accompany all my representations, then that representation of myself as thinking mentioned in the condition surely will not need further such accompaniment. If the accompanyability of the 'I think' is to be a condition on anything's being represented to me, my own thinking must already, independently of the condition, be capable of being represented to me. But there is a trap here. The 'I think' that is here at issue is not,as we shall soon see, the representation of myself as thinking; it does not, as it does in Descartes, express the proposition that I am thinking. So the 'I think' is not really a representation at all; it does not represent anything as anything. On the other hand, the representation of myself as thinking, proposition 'I am thinking' is a bonafide as a bonafide representation, is subject to the condition at issue, the and, accompanyability of the 'I think'. At first blush it looks as if Kant's claim about the 'I think' is nothing but a weakened form of Descartes' doctrine that all thinking makes implicit reference to oneself, the thinker. Descartes seems to have held that if I have the thought that there is a book on the table, what is really in my consciousness there is a book on the on a is not there being book thetable,but rather thatI think table. When I see you across the room, what I am really aware of is notyou So across room, thatI see(i.e., think I see)you there. I am the true subject the but of all my thoughts. And by that is meant not merely that I am their agent (as I am also the agent of all my eating and walking), but that my thoughts are really aboutme. The first-person singular pronoun, 'I', is the proper grammatical subject of every proposition that expresses a thought; every asserted proposition 'p' is, when fully articulated, of the form 'I think p'. The sceptical doubts of the First Meditation must surely have confirmed Descartes in this doctrine, but the doctrine does not really depend on those doubts. It seems to be philosophically puzzling in any case how I could have a anything like this book, real solid object, in my consciousness or 'present to the mind'. Must it not instead be the thought, or the idea, of the book that is really there? And indeed my thought, my idea? It is worth remembering in
? The Editorsof ThePhilosophical Quarterlv, 1997

SUBJECTIVITYIN DESCARTES AND KANT

349

this connection that Locke, who had no time at all for Descartes' doubts, also held that all thought is, initially in any case, about one's own ideas. However, the relationship between this doctrine of Descartes' and the doubts of the First Meditation is an interesting one; I shall talk about it later. This doctrine is at the heart of Descartes' subjectivism: the objects of consciousness, or in any case its direct objects, are one's own inner states. It might look as if Kant's remark about the 'I think' is simply a weaker, or perhaps more careful, version of this doctrine. Instead of 'wherever there is a thought, there must be an "I think" too', he is saying 'wherever there is a thought, it must be possible for there to be an "I think" too'. And this seems to be a negligible difference - especially when one notes that Descartes could well have used Kant's formulation. It seems that there is no reason why Descartes should not have allowed that I do not haveto make explicit to myself, on each occasion of thinking, the fact that my thought is mine; it is of enough that I be capable doing so - that suffices to show that I am aware of it as my thought. And so it will appear that Kant, whatever his final aims and arguments may be, must initially be endorsing Descartes' subjectivism: consciousness fully articulated has the form 'I think p'; the immediate object of consciousness is one's own thought. And from this point the standard picture of Kant's response to Descartes will seem inevitable. Since we know, or have good reason to believe, that Kant means to argue for further conditions of 'thought-and-experience', conditions that go beyond the 'I think' condition and point to an allegedly objective realm, we shall anticipate that he will be arguing that things must really be thus and so if our thought and experience about them are to be what they are; that is, he will be building a metaphysical bridge from the subjective to the objective. But, as I shall try to make clear, the 'I think' is, for Kant, no starting-pointfor such a bridge; and in particular it is not, as it was for Descartes, my recognition of myself as being in a conscious state. So what is it? And what role is it playing in Kant's argument for the objective validity of the categories? There is no difficulty in supposing that Descartes and Kant will agree on this: that if I am to think of or be aware of something, if something is to be represented to me, if there is to be an (intentional) object of my consciousness, then I must be capable of noting that fact. Any thought of mine to the effect p that I can phrase as 'p' to myself I must be able to rephrase as 'I think p'. Consciousness has that sort of reflexivity about it. If it did not, the data striking my retinae, or otherwise at work on me or in me, would, as Kant says, be 'nothing to me'; I would be no better than a camera or a computer, affected by stimuli or going through otherwise relevant motions but to which these stimulations or motions meant nothing whatever; I would have no Intentional consciousness, understanding, is a situation for understanding.
? The Editorsof ThePhilosophical Quarterly, 1997

350

HUBERT SCHWYZER

it or someone, to someone, is not merely a situation in someone. And that 'forsomeone factor' marks something essentially first-personal. It is what Kant calls 'pure apperception'. Up to this point Descartes and Kant are together. But from here on they diverge. For whereas Descartes sees the essential first-person involvement in consciousness as bedrock, as defining, or exhibiting to one's inner self, the very nature of consciousness, and so, as it were, as solvingthe problem of what it is to represent something to oneself, Kant sees that involvement as calling for an explanation, and so as posingthat problem. Descartes regards the 'I think', the involvement of the self in consciousness, as itself an object, indeed theultimate object, of consciousness (viz., that I am thinking, cogito). is the only item whose presence to consciousness is It unmediated. Every other item, e.g., cheese, or you across the room, needs mediating: it needs to be explained how such an item can be present to consciousness. And the explanation is that such items are present to my consciousness precisely by being represented, in the form of ideas, and thus by cast the content of my thinking - where my thinking's presence to my becoming consciousne.s is presumed to need no explanation. Kant has arguments against taking the 'I think' to be any kind of object of consciousness, something of which one is aware. First, what is given to consciousness lacks the requisite unity: 'I should have as many-coloured and diverse a self as I have representations of which I am conscious to myself' (BI34).This 'to myself', the 'for-someone factor', as I called it above, must be one and the same thing in all my consciousness. Moreover, and much more importantly, my thinking, taken as object of my consciousness, even if it were universally and univocally present, could not in any case capture the 'to myself', the 'I think' of pure apperception. For if I am conscious that I am thinking, then thatI am thinking is a situationfor me.If it were not, then I would not be conscious of it and that would be the end of it. But if it is, then that 'for me' factor remains, unreduced and unaccounted for. It has not been, and cannot be, cashedin for 'I am thinking', or for any other thought. One cannot eliminate the 'to myself' that must qualify every object of consciousness in favour of some further and ultimate object of consciousness that as it were speaks for itself, spelling out its own that-it-is-for-me. Neither my thinking nor any other item of which I am conscious can account for the 'to myself'; any attempt to locate this condition of consciousness in the contents or among the objects of consciousness 'has always already', as Kant says, 'made use of its representation' (A346/B404). So with the 'I think' of pure apperception we have not, pace Descartes, reached the bedrock of consciousness. Far from exhibiting the very nature of consciousness to the one who is conscious, the 'I think' does no more than mark the
? The Editorsof ThePhilosophical Quarterl,, 1997

SUBJECTIVITYIN DESCARTES AND KANT

351

fact that consciousness is a situation for one. We still have the question of what it is that accounts for that fact. In virtue of what is anything a situation for one? And this is the question of what makes it possible that a datum, say something striking the sense-receptors, should represent something to me. That it does so does not, after all, follow from the fact that the sensereceptors are stimulated (see esp.
A90/BI22-3).

To say that something is

inflicted on someone's sensibility is to say one thing, to say that that means something to that person is to say another thing. We still need an explanation of the for-me factor; it cannot, as Descartes thought it was, be a brute datum of consciousness. But this raises another question, on a higher level. What sort of explanation of the for-me factor is Kant looking for? Is it to be an external explanation, in terms of one's physical make-up or environment, etc.? Or is it to be an internal one, in terms of first-person access to the contents of one's consciousness? Well, we know it cannot be the latter; that has been precisely Kant's argument against Descartes. Any internal explanation will have to the employ for-me factor, and will therefore be unable to account for it. But on the other hand Kant is hardly likely to be interested in external, causal, factors, like neural or environmental conditions. These are empirical and contingent factors, and not relevant to these kinds of philosophical concerns. There is a short-cut, if not fool-proof, procedure for determining what kind of explanation of the for-me factor Kant is after, and that is to jump ahead and see what explanation he in fact gives. Let us then look, briefly, at the outcome the Transcendental Deduction. What makes the for-me factor of Kant claims, is a kind of actionthat we perform, viz., the act of possible, And judgement. judgement, or judging, as has been shown earlier, in the Clue, is a matter of synthesizing, or bringing together, the data of intuition, by means of certain conceptual operations, the categories. By judging (as it were forming mental sentences) we create a unity(a sense, a meaning). This is the (synthetic)unity of consciousness itself. Without it, there can be no forme factor, nothing can mean anything to me, there is no representation of anything, no intentional object for my consciousness. A brief example might be of help here. Someone, in daylight, with eyes open, and with cheese in front of him, is being sensorily affected, that is, there is a cheese-image on his retina, his brain is functioning appropriately, etc. In virtue of what is this a situation for him such that he can say to himself 'I'm seeing cheese'? What is the process or state of affairs that makes it clear and intelligible to us as philosophers, a priorienquirers, that this is a situation for him? This much internality is needed. It will not be enough to be told that nature and the organism in question are simply such that when circumstances (light, oxygen, retina, brain, etc.) are thus and so,
? The Editorsof ThePhilosophical Quarterly, 1997

352

HUBERT SCHWYZER

representation occurs. Kant would not doubt that;but it does not address the philosophical question. What we need is to make the nature of the for-me This is something which, for factor, the first-person perspective, perspicuous. quite different reasons, neither an external causal story, nor a story like Descartes',fromthe first-personperspective, can accomplish. Kant's answer to our present question (what makes it clear and intelligible to us that this is a situation for him?) is not in the end a recondite one. It is because, and only because, the subject sees thatthere cheese font of him is in that he can say, reflectively, 'I'm seeing cheese'. And that means that he forms the thought, makes the judgement, thatthere is cheese here. The forme factor is possible not because one is conscious of oneself, but because one can make judgements about objects. And to do that is to conceive them under the categories, to see them as substances, causes, totalities, etc. If one could not do that, one would lack the for-me factor. Now whether or not this explanation of Kant's is plausible in detail, it is clear that some explanation of the for-me factor along the general lines indicated is called for. For the for-me factor is not, as Descartes thoughc, somehow self-explanatory. We need to come to understand, and to articulate, from a non-first-person point of view, what it is in virtue of which something is a situation for one.

IV
But now, to return to Descartes, someone might be wondering whether the needs to be understood in the way we have proposed, as nothing more cogito than pure apperception, the non-propositional for-me factor, parading illegitimately as a thought. Perhaps Kant is wrong, not in arguing that the for-me factor is non-propositional, but in believing that the cogito exemplifies the error he is pointing out. We need to come to terms with this question, especially so if Kant's argument is to have any repercussions for what I have called Descartes' subjectivism, the view that the (direct)objects of consciousness are always and only one's own inner states. Cannot the cogito,as it functions in Descartes, and despite Kant's insistence to the contrary, be understood straightforwardlyas the thought that I am thinking, and not as the for-me factor? And to this we might add a further question: even if the has to be understood as nothing more than the for-me factor, cannot cogito Descartes' subjectivism be seen to stand independently of it? Could 'Cogito',as it functions in Descartes, be understood as simply registering what happens to be the present content of consciousness, like 'I'm looking out of the window', or 'I'm hungry'? Well, one wants to say,
? The Editorsof ThePhilosophical Quarterly, 1997

SUBJECTIVITYIN DESCARTES AND KANT

353

unlike these, whose truth depends on circumstances that might or might not 'I obtain, 'Cogito', am thinking', is alwaystrue. - But of course it is not. I am frequently not thinking. What is always true, as Descartes indeed recognizes, is only 'I am thinking' when I thinkit. But then it, the actual content of thought, becomes irrelevant; 'I am thinking' will equally be true when my thought is that the refrigerator is empty, or anything else - for it is not whatI think but thatI am thinking that is at issue. 'Cogito' it does not record the content of a particular thought, but only of any of my thoughts you please, its beingmy thought. this means that it records precisely the for-me factor But that all my thoughts must possess if they are to be my thoughts. For nothing else but the for-me factor is present in all my thoughts simplyin virtue their of if So being thoughts. 'Cogito', it is to play the role that Descartes intends it to my play, can be nothing more than the 'I think' of pure apperception. It is, then, precisely Descartes' insistence that 'Cogito' always true, whatis ever the content of one's thinking, that unmasks it as the for-me factor. And Descartes' great mistake - surely one of the most interesting and influential mistakes ever committed in philosophy - was to suppose that the for-me factor could itself be captured in the content of one's thinking - as if there could be that primal thought, 'I am thinking', that, unlike all other thoughts, does not have to represent how things are for one, precisely because it succeeds in absorbing into its own articulation what it is for anything to be represented to one. I have tried to spell out Kant's reasons for supposing that this cannot be done. I believe they are good reasons. The 'I think' that it must be possible to append to all my representations if they are to be anything to me cannot itself be caught in a thought. V Two major points have emerged from our discussion. First, the 'I think', the for-me factor, is essentially something non-propositional. Hence the preferthe ability of calling itfor-me,rather than I think; first person's necessary presence in consciousness is grammmatically dative, not nominative. That presence is not itself a thought at all, and so it is not something true, and not something that can be known to be true, and therefore not something from which other truths can be derived. (This is not to deny that when the for-me factor obtains something is thereby true. It will be true, always true, that the person in question, who might happen to be me, is thinking. But the proposition that is always true here is the trivial 'If something is a situation p for someone, then that person is thinking p'. This has none of the other features that Descartes needs in the cogito; is not essentially first-personal, it is not it
? The Editorsof ThePhilosophical Quarterly, 1997

354

HUBERT SCHWYZER

even singular, it is not categorical. And its being true does not in any way imply that the for-me factor is itself a truth.) Second, and connectedly, the 'I think' is in another way non-basic: rather than being the foundation of explanations of consciousness, it itself stands in need of an explanation in terms of a genuine act of consciousness (the 'I think' is itself no such act). The first of these two points has consequences for Cartesian subjectivism. While the fact that the 'I think' of pure apperception is not itself a thought does not directly refute the view that one is immediately aware only of one's subjective states, it does succeed in undermining Descartes' justification for subjectivism. For Descartes' grounds for subjectivism consist precisely in taking the for-me factor to be itself the thought that one is thinking. And so the very nature of consciousness - that it is a situation for one - makes it inevitable that one should be directly aware only of one's own states. The formula that expresses consciousness is not 'p' (e.g., 'This is cheese'), but 'I am thinking p' (e.g., 'I am having sensations of cheese'). The 'I think' is always present where consciousness is present, and since it is understood propositionally, as asserting that I am thinking, it is therefore invariably that of which I am conscious in so far as I am conscious at all. Other items (e.g., cheese) are possible as objects of consciousness only in so far as they are collected under the mantle of the 'I think', that is, as ideas. To take away the propositionality of the 'I think', as Kant has done, is to take away Descartes' reason for subjectivism. There is nothing in the nature of consciousness, as Descartes thought there was, to prevent cheese, or the presence of cheese, from being what I am directly aware of. There is no longer that principled objection to unmediated awareness of outer things. And so we need not be troubled by the thought that the very nature of consciousness makes it the case that we have to provide a special argument (perhaps a metaphysical bridge) to lead us from the inner to the outer. There is no such original confinement to the inner. I can imagine someone not being persuaded by what I have been arguing, and objecting as follows. Surely Descartes' subjectivism has a source that is independent of the cogito, since that doctrine is already established in the First Meditation (as a result of the sceptical doubts) before there is any mention of the cogito.If this is so, it will not be right to attribute the subjectivism to a misreading of the for-me factor, as I have argued. So now we have three items in the air to juggle with: (a) the scepticism of the First Meditation; (b) the misrendering of the for-me factor in the cogito; and (c) the subjectivism. I have claimed that (b) is the source of (c); I have not yet talked much about (a). The objection claims that one can move straight from (a) to (c) without going via (b).
? The Editorsof ThePhilosophical Quarterly, 1997

SUBJECTIVITYIN DESCARTES AND KANT

355

I do not think this objection succeeds. It seems to me that to the extent that the doubts of the First Meditation entail subjectivism, they will also entail a misrendering of the for-me factor. So, in so far as subjectivism is is present in the First Meditation, the cogito also prefigured there. This is because merely raising philosophical doubts about reality, or seeing such doubts as intelligible, probably does not of itself commit one to subjectivism at all. But Descartes' doubts take the form of a 'thoughtone'sexperience (how things are for one) in experiment' in which one compares the case where one supposes there to be external objects producing it with one's experience in the case where one supposes there not to be such objects. And the result of that thought-experiment is that there is no crucial difference between the two cases that the experiencer can detect. All that I am really aware of in experience, the thought-experiment reminds us, is the experience itself that I am having. This is certainly (an instance of) subjectivism. So the objector is right in claiming that subjectivism is present in the First Meditation. And it is also true that the cogitohas not yet been mentioned. But this is not the end of the matter. While the thought-experiment of the First Meditation, in which one compares experiences, might explain why one would want to say that all I am aware of, if I am aware of anything, in experiencing something is the experience itself, it does not explain but takes for granted that, in experiencing something, I am, and indeed inevitably am, aware of the experience itself. Where does this thought come from? Why suppose that, whatever the content of my experience might be, I am always conscious thatI am having the experience? Well, one might say, that is just what experience is like; it is reflexive, self-aware, that is its whole essence. But what shows that this is so? And to this question the only possible answer, it seems to me, is in terms of the for-me factor. Experience is indeed something for me. What has happened here is that, once again (or rather, already), the for-me factor has been recast as the object of one's awareness. this It is as if because or that state of affairs, e.g., my sitting at the computer in my dressing gown, is a situation for me, one of the things I am therefore aware of, i.e., that is therefore thus and so for me, is thatit is thus and so for me. In the absence of this transformation of the for-me factor into a proposition I make about myself there is, as far as I can see, no plausibility at all in the claim that in experiencing anything I am always aware of having dark work is already under way the experience. If this is so, then the cogito's in the First Meditation; and that means that Descartes' misrendering of the for-me factor infects his scepticism as well as being responsible for his subjectivism.
? The Editorsof ThePhilosophical Quarterlr, 1997

356

HUBERT SCHWYZER

VI Let us now draw some threads together. We have seen that Kant renounces the Cartesian view that the 'I think' is a fundamental truth, and with it the doctrine that only what is subjective can form the starting-point of knowledge (the near side of any cognitive bridge). So it should be clear that whatever Kant's philosophical enterprise amounts to, it will not be a matter of arguing from how things are with us to how they are in reality, from the subjective to the objective, from inside consciousness to what is outside it. I have rejected the standard picture of Kant's response to Descartes. And of course I have argued for more than this: I have argued that Kant's repudiation of Descartes' picture of consciousness is actuallyjustified. But how are we to assess this victory over Cartesian subjectivism- assuming it to be a victory? It is perhaps unclear just what has been achieved. And Stroud's question might look as if it can be raised again: have I not simply argued, someone might ask, that if we are to represent anything to ourselves, we must do so in objective terms - that is, we must make judgements of the form 'this is cheese', where that does not reduce to 'I think this is cheese'? Is must be like? And is it not really much that not still about what our thinking the same as saying that if we are to represent anything, be conscious of in anything, we have to believe objective states of affairs? I do not know whether this suspicion can be finally laid to rest by what I have been arguing. But, first, if what is being said is on the following lines, then I think I have answered it: 'So you are saying that we must represent things in objective terms if we are to be conscious of anything, that we must judge "p",not "I think p". But this is still a statement about us as judging, it is not about how things are apart from us. And that matters because, after all, all that we are really aware of in judging p is thatwe are judgingthatp, not thatp itself.' If this is the suspicion, then I have answered it. It is the familiar twist on the 'I think'. 'But', you will say, 'all that has been established by in Kant is that we haveto represent things certain ways; and is that not simply a but that does not mean that in representing those fact about us?'. Certainly, things in those ways we are not directly aware of them. The fact that we represent things does not itself sentence us to being directly aware only of our representations. But this is not enough, it will be thought, to put aside the Stroudian suspicion. The fact that we have to represent things as thus and so, it will be said, does not, after all, mean that the things represented must themselves be thus and so. For all we know they might in actual fact be quite otherwise. So
? The Editorsof ThePhilosophical Quarterly, 1997

SUBJECTIVITYIN DESCARTES AND KANT

357

the fact that we have to represent them as thus and so can still only mean that they have to be thus and so in our minds,which is surely the same as them to be, that saying that they have to seemto us, or that we have to believe way. Again this could, in another context, be harmless. There is nothing intrinsically threatening to our knowledge of real things in the thought that we must represent those things as this or that; or even, if one insists on the tendentious terminology, that we must believe them, or that they must seem to us, to be this or that. That thought will not be threatening unless we take the fact of representation, or of believing or seeming, to block our access to how things actually are. But it can surely only do that if we take the fact of representation, by contrast with what is represented to us, to be what we are really (or directly) aware of in representing anything. A footnote in Werner Pluhar's new translation of the Critique illustrates the continuing influence of the standard picture.3 Pluhar explains that representations (he calls them 'presentations') 'are such objects of our direct awareness as sensations, intuitions, perceptions, concepts, cognitions, ideas, and schemata'. The implication seems clear: things in space are not objects of our direct awareness. But it is only if we suppose that its seeming to us that p is what we are really aware of when it seems to us that p, that its seeming to us that p will stand in the way of our being aware that p when it seems to us that p. If we do not suppose that it is our representation that p that we are really aware of when p is represented to us, then there will be no temptation to say that in so far as p is represented to us we are not directly aware of p. But this is where we were before, with Descartes' trick of transforming the fact of representation into what is represented thereby, the manoeuvre of converting intentional ccnsciousness into its own intentional object. If we resist this manoeuvre then there will be nothing restrictive, or ominously 'merely psychological' or 'purely subjective', about the fact that we represent things to ourselves, or about the fact that we do it, or have to do it, in this or that way. The fact, then, that Kant is telling us about the conditions of representing objects to ourselves does not have the consequence that he cannot consistently maintain that we are directly aware of those objects.4 at University Califomia SantaBarbara of

am grateful to Tony Brueckner and Rudy Winnaker for many helpful comments on an earlier draft of this paper.
? The Editorsof ThePhilosophical Quarterly, 1997

4I

3Indianapolis: Hackett, 1996, p. 22.