Self and World Revisited Quassim Cassam When Tony Cheng first suggested a conference to mark the 20th anniversary of the publication of Self and World I was both pleased and sceptical. Naturally I was pleased by attention but I doubted that enough people would still be interested in the book for such a gathering to be worthwhile. The book has always been seen, not entirely accurately, as following in the footsteps of P. F. Strawson and as very much part of an Oxford neo-Kantian tradition that appears to have gone of fashion. The issues discussed in Self and World were very much alive in Oxford until the 1990s but it is hard to think of many people today who still work in that tradition. It’s a good question, which I will come back to, why these issues are so much less prominent today than they were twenty years ago. Even twenty years ago the appetite for the kind of philosophy I was doing in Self and World was waning, and the initial reception of the book was not what I had hoped for. There was an enthusiastic review in French by Pascal Engel but lukewarm reviews by Peter Sullivan, Daniel Hutto, Michael Morris and Sydney Shoemaker. Worst of all was a spectacularly hostile review in Mind by Andrew Brook. Mike Martin, who was Mind’s Reviews Editor at the time, had sent the book to Brook and he responded with a review that was so vitriolically negative that it made me wonder what I had done to offend him. The published version of the review was bad enough but still milder than the first draft which Mike had asked him to tone down. On reflection, I can see that what was driving Brook was partly a deep hostility to the philosophical tradition in which I was working. He objected to the fact that there is no empirical psychology in Self and World and seemed to feel that the entire methodology of the book was suspect. I imagine that Brook wasn’t the only person who held this view though few have expressed their reservations in the trenchant way that Brook expressed his. Rereading the book recently I came away with a mixture of pride and embarrassment. I felt that the book has some good things in it but I was embarrassed by the turgid writing style, which makes it very hard to read, as well as some of its less convincing arguments. It’s a strange criticism of a philosophical work but the fact is that there are far too many arguments in Self and World. The book is packed with theses, arguments, objections, responses to objections and responses to responses to objections. I should have just said what I wanted to say and not spent so much time worrying about possible rebuttals. The fact that I proceeded as I did was partly the result of the natural defensiveness of someone writing their first book. But there was more to it than that. Although I was a huge admirer of Strawson, and in particular of The Bounds of Sense, I shared the frustration of many readers of Strawson at the elusiveness of his arguments and the sense that, beneath the elegance of his prose, one never really knew exactly how they were supposed to work. Nobody could accuse Self and World of elegance but I was determined to make all my key premises and moves absolutely explicit. The Bounds of Sense is a work of genius whose moving parts are, like the moving parts of a Rolls Royce, hidden from view. Self and World strikes me as almost painfully scrupulous, and, for better or worse, its moving parts are about are visible as they could be. If The Bounds of Sense is a philosophical Rolls Royce one can only speculate what that makes Self and World. Mention of Strawson brings me back to the question of the relationship between my work and his. In a recent paper in the European Journal of Philosophy Lucy Allais, a former pupil of mine whose doctoral thesis was examined by Strawson, describes a conversation with an unnamed British philosopher who asked her how she saw the difference between my work and Strawson’s and her work and Strawson’s. The implication, as Allais interpreted the question, was that there was no difference between the three of us. Yet when I was writing Self and World I thought I was arguing against Strawson and breaking with the Oxford neo-Kantian tradition. I was more conscious of being influenced by Merleau-Ponty and the work of my Oxford colleague Michael Ayers. From Merleau-Ponty I took the idea that the bodily self is a ‘subject-object’, not a ‘mere’ object but an object that is a point of occupancy for psychological properties. From Ayers I got the related idea that the bodily self is the ‘presented subject of experience and action’ and that ‘our experience of ourselves as being a material object among others essentially permeates our sensory experience of things in general’. Strawson had little philosophical interest in bodily self-awareness, and my insistence that what self-consciousness requires is intuitive awareness of oneself ‘qua subject’ as a physical object rather than the conception of oneself as a physical object was intended to mark a clear dividing line between my position and Strawson’s. My rejection of the Strawsonian emphasis on conception went back to my D. Phil. thesis on transcendental arguments. Arguments of this form try to show that the truth of some disputed proposition P is necessary for experience. Since we know that we have experience this puts us in a position to know that P is true, and hence to refute varieties of scepticism that question our knowledge of P. Following the publication of Stroud’s paper on transcendental arguments in 1968 Strawson conceded that transcendental arguments are too strong and that the most that they can show is that in order for us to have experience we must believe that P is true. So, for example, what self-conscious experience requires is only the belief that some of our experiences are experiences of mind-independent objects or the belief that the subject of experience is a corporeal object among corporeal objects. But claims about what we must believe to be the case are, in their own way, even harder to establish than claims about what must actually be the case. The sceptic about the external world claims not to believe that his experiences are experiences of mind-independent objects and the Cartesian dualist claims not to believe that the subject of his thoughts and experiences is corporeal. Yet both the sceptic and the dualist have self-conscious experience. So what do ‘belief-directed’ transcendental arguments have to say about them? This led to the formulation in Self and World of the ‘problem of misconception’. The issue comes to a head in Strawson’s account of the Paralogisms, the chapter of the first Critique in which Kant takes on the so-called ‘rational psychologist’, who thinks that he has a priori knowledge of the nature of the self. According to Strawson, the rational psychologist thinks that the “I” of Kant’s “I think” refers to a peculiar kind of entity – an indivisible and immaterial soul substance- because no criteria of personal identity are invoked in the self-ascription of experiences. Such self-ascriptions are ‘criterionless’ but - and this is Strawson’s response to the rational psychologist - “I” can be used without criteria of subject-identity and yet refer to a subject because, ‘even in such a use, the links with those criteria are not in practice severed’. What are the criteria of subject-identity and in what sense are links with these criteria not in practice severed in criterionless self-ascription? On the first question Strawson, unlike Kant, takes it that the criteria of subject-identity ‘involve an essential reference to the human body’. The sense in which links with bodily criteria aren’t severed in criterionless self-ascription is that even in soliloquy “I” is used by a person who would, as Strawson puts it, ‘acknowledge the applicability of those criteria’. The claim is not just that the referring use of “I” requires the acknowledgement of the applicability to oneself of some criteria of subject-identity but the acknowledgement of the applicability to oneself of criteria of identity that make essential reference to the human body. In that case, what do we make Cartesians or Lockeans who think of personal identity in psychological rather than bodily terms? This issue was made especially vivid to me by the fact that one of my earliest Oxford colleagues was Richard Swinburne, who I took to be a card-carrying dualist. Conversations over lunch at Oriel College in the mid-1980s left me in no doubt that Swinburne did not lack self-consciousness in the philosophical sense. They also left me in no doubt that he did not acknowledge the applicability to his thinking self of the bodily criteria of identity favoured by Strawson. Strawson seemed to want to say that Swinburne didn’t really believe that his thinking self is immaterial and was only pretending to believe it. This was what I described in Self and World as the first response to the ‘problem of misconception’. I think I was right to reject this response, although I’m less sure that I was right to reject some other alternatives to the one I eventually sided with. What is ‘the problem of misconception’? This is how I characterised it in chapter 4 of Self and World: Even if persons are in fact physical objects, it does not follow that self-consciousness requires the conception of oneself as a physical object unless it is also accepted that self-consciousness requires a coherent, accurate, and substantive conception of what one is…. Consider, once again, the Cartesian dualist who regards the persisting subject of her thoughts as an immaterial substance. This belief may well be philosophically indefensible….. but this surely has no bearing on her ability to think first personally (1997: 127). Leaving aside the possibility of insisting that dualists doesn’t believe what they say they believe it seems to me now that the most promising response to the problem of misconception is to take transcendental arguments as being concerned to show what, simply by virtue of being self-conscious, we are rationally committed to believing, or ought to believe, regardless of what in fact we do believe. In Self and World I rejected this response because, like Strawson, I accepted a ‘knowing which’ requirement on self-reference and couldn’t see how one could count as satisfying this requirement in virtue of what one ought to believe rather than what one does believe. The ‘knowing which’ requirement is associated under that description with Evans who in turn attributes it to Russell. As formulated by Evans the principle, which he calls ‘Russell’s Principle’ states that ‘a subject cannot make a judgement about something unless he knows which object his judgement is about’ (Evans 1982: 89). Strawson’s anti-Cartesian argument in the Bounds of Sense and quite a few of my arguments in Self and World rely on some such principle. However, the principle in its general form is controversial and its application to self-reference is even more controversial, as I well knew when I was writing the book. One of the very few people with whom I had philosophical conversations at that time was John Campbell and he used to argue that not being subject to Russell’s Principle requirement is a distinguishing feature of first-person thought. On Campbell’s account the reference of “I” is fixed by the token-reflexive rule – the rule that any token of “I” refers to whoever produced it – and knowing which thing one is, or what kind of thing one is, doesn’t come into it. The fact that I helped myself to Russell’s Principle without seeing any need to defend it shows how deeply influenced I was by Evans. Apart from questions about the plausibility of the ‘knowing which’ requirement itself perhaps the most serious question about my relying on it in the book is that it is hard to reconcile with the supposed metaphysical neutrality of my account of self-awareness. In Self and World I repeatedly assert that my account is an account of how we must experience ourselves as being and that the claim that the presented subject of experience is a material object among others ‘would be compatible with maintaining that the subject is in fact non-physical’ (1997: 58). However, I also argue that I count as knowing which thing my first-person judgements are about in virtue of being intuitively aware of one’s spatial properties and spatial boundaries. Bodily awareness, I claim, delineates the object that is in question in first-person thought. But can bodily awareness provide one with knowledge of which thing one is if one is in fact immaterial? It’s hard to see how this can be so. If this is right then reliance on Russell’s Principle is incompatible with metaphysical neutrality. Whether this is an argument against replying on Russell’s Principle or for abandoning claims to metaphysical neutrality is another question. I suspect that if I had confronted this issue head on in 1997 I would have been more inclined to abandon neutrality than Russell’s Principle. Despite these reservations I do think there is something right about the idea that what matters for self-conscious self-reference is not one’s self-conception but rather the nature and content of one’s ‘intuitive’ self-awareness’, that is, the nature and content of one’s experience of oneself. Even I am deluded as to the type of thing I am I still have a concrete sense of where I end and the rest of the world begins, and I have this sense in virtue of being intuitively aware of myself as a bounded physical object among others. I argued that this type of self-awareness is necessary if one is to be in a position to conceptualise one’s perceptions as perceptions of mind-independent objects. This is the ‘objectivity argument’ that is developed at great length in chapter 2 of Self and World, which was the first chapter I wrote. In chapter 4 I develop the ‘identity argument’ according to which intuitive awareness of oneself as a physical object is required for consciousness of one’s own identity as the subject of different representations. In developing these arguments I insisted that the requisite awareness is not just awareness of what is in fact one’s thinking, experiencing self as a physical object but consciousness of oneself ‘qua subject’ as a physical object (1997: 25). However, this move introduces a range of further difficulties which deserve more detailed discussion than is possible here. The first concerns the difficulty of reconciling awareness of something ‘qua subject’ with awareness of it as an object. A possible objection, which I called the ‘incompatibility objection’, is that these are mutually exclusive modes of awareness. I tried to deal with this issue in the book but in a paper published in 2006 Béatrice Longuenesse argues against my view and in favour of what she interprets as Kant’s view that ‘consciousness of oneself as the subject of thought is not and cannot be consciousness of oneself as an object’ (2006: 284). What Longuenesse attributes to Kant is a kind of dualism, not mind-body dualism but subject-object dualism. One of the aims of my book was to undermine subject-object dualism but Longuenesse argues in her insightful paper that I was not successful in this. Longuenesse is absolutely right to represent Self and World as propounding a view of self-consciousness that is, in certain key respects, fundamentally anti-Kantian. I wrote in the final paragraph of Self and World that Kant was right to insist that self-consciousness doesn’t involve ‘an intuition of the subject as object’ (B421) if this means an intuition of the subject as an immaterial substance. Kant’s mistake, I said, was to conclude that self-consciousness ‘does not involve being presented to oneself as an object at all’. The correct view, I said, is that self-consciousness is ‘intimately bound up with up’, and indeed requires, ‘awareness of the subject “as an object” – not as an “immaterial” substance but as a physical object in a world of physical objects’. My view, as Longuenesse describes it, is that ‘we could not ascribe our perceptions and representations to ourselves, take them to be our own – unless the very same state of awareness that is awareness of ourselves qua subjects of thought and knowledge, were also an awareness of ourselves as a physical object among physical objects’(2006: 284-5). What this formulation implies is that on my account there is a single state of awareness of ourselves as subjects and as objects. On this reading, I’m not just saying that each of us is aware of himself or herself as a subject and as an object, as if these are two separate states of awareness. The view that there are two separate states of awareness is what might be called the ‘conjunctive view’. Longuenesse thinks that what I am defending in Self and World is the stronger ‘single state’ view. This is where, according to Longuenesse, I company with Kant. As she reads him, Kant would have agreed that self-consciousness is intimately bound up with consciousness of one’s own body but he would not have agreed that ‘this consciousness is a consciousness of ourselves qua subjects as a physical object’ (2006: 285). On this point Longuenesse thinks that Kant is right and I’m wrong, although she is kind enough to say that in the preface of I, Me, Mine that the project of that book grew out of an attempt to understand the relationship between my view and Kant’s. So why does Longuenesse prefer Kant’s view? The crux of the matter is my conception of what it would be to be aware of anything ‘qua subject’. Longuenesse’s point is that awareness of one’s bodily self ‘qua subject’ is incompatible with that same awareness counting as awareness of that same self as an object. The best I can hope for, she argues, is the conjunctive view. What is my conception of awareness of something ‘qua subject’? Drawing on Sydney Shoemaker, I suggest (1997: 60) that for a form of awareness to count as awareness of the self qua subject it must at least be a form of awareness on the basis of which it is possible to make first-person statements that are immune to error through misidentification. In addition, it must be a form of awareness that does not require one to keep track of the identity of the referent of “I” in one’s self-ascriptions. I argue that bodily self-awareness satisfies both the immunity to error condition and the no tracking condition. For example, ‘when a subject is disposed to judge “I am in front of Marble Arch” and a few moments is disposed to judge “I was in front of Marble Arch”, there is no question of the transition from the former to the latter disposition needing to draw upon an observationally based re-identification of the subject or an ability to keep track of the subject' (1997: 70). But awareness of something as an object does require the exercise of an ability to keep track of an object. So the considerations that support the idea that bodily self-awareness is awareness of oneself qua subject also imply that it isn’t and can’t be awareness of oneself as an object. It isn’t a solution this problem to argue that bodily self-awareness, including awareness of one’s physical location from one moment to the next, does require keeping track. That would allow it to count as awareness of one’s body as an object but now it isn’t awareness of it qua subject. In the book I respond to this worry by distinguishing two senses of object awareness. In a broad sense, to be aware of something as an object is to be aware of it as a physical object, that is, as shaped, located and solid. There is no requirement that one needs to keep track of the object. The tracking requirement is, however, built into a different and stricter sense of object awareness. Then my claim is that it is only in the former sense that awareness of oneself qua subject is awareness of oneself ‘as an object’: even if one doesn’t have to keep track of one’s bodily self from one moment to the next or, as Longuenesse puts it, exercise ‘any particular skill for identification or reidentification’ (2006: 296) one is still aware of it as shaped, located and solid. Longuenesse is suspicious of this attempt to overcome the subject-object dichotomy. She argues that the ‘qua’ in ‘qua subject’ introduces a mode of access to what we are aware of. In contrast, the ‘as’ in ‘as an object’ introduces the ontological category to which the thing we are aware of belongs. My thesis is that ‘even in the peculiar mode of awareness of oneself that is awareness of oneself qua subject, one is aware of oneself as a physical thing among physical things’ (2006: 297). What I can’t claim – because such a claim would make no sense – is that ‘one is aware of oneself qua subject qua object’ (2006: 297). So even on my view, there is an incompatibility between these two forms of awareness. Longuenesse does not wish to deny that the thing one is aware of qua subject is the same entity that one is aware of as an object but, as she puts it, the relation between these forms of awareness is ‘one of conjunction, not one of identity’ (2006: 298). In other words, my argument fails as a defence of the ‘single state’ view but not as a defence of the ‘conjunctive view’. Presented in this way, Longuenesse observes, my position is very close to that of Merleau-Ponty. Rereading Self and World I can find no decisive evidence of an understanding of the distinction between the single state and the conjunctive view. I think Longuenesse is correct, though, that what I was trying to defend was probably the single state view even if I didn’t explicitly think of the issues in these terms. I think she is also correct that the conjunctive view is the best I can hope for. In retrospect I don’t have a problem with this. There is evidently some sense in which if one is aware of something ‘qua subject’ one can’t, in being aware of it that way, also be aware of it ‘qua object’. If, on the other hand, it is possible for one to be simultaneously aware of one and the same thing ‘qua subject’ and ‘as an object’, and the thing that one is simultaneously aware of in these ways is a physical thing, then I am content with that. If, as Longuenesse claims, the conjunctive reading of where I end up leaves me in close proximity to Merleau-Ponty’s view of the body as a “subject-object” then that’s fine too. While I can’t claim to have entirely overcome Kant’s dualism of subject and object I think I can claim to have weakened its grip at least to some extent by bringing into focus the possibility that one and the same thing can be presented as the subject of experience and as physical object in a world of physical objects. From the fact that such dual awareness is possible it doesn’t follow, of course, that it is necessary. Why wouldn’t it be possible, for example, for me to be in a position to think of some of my experiences as experiences of mind-independent objects even if I’m not aware of myself qua subject as a physical object? Perhaps, in order to have experiences as of mind-independent objects I need to be conscious of myself as having a body but why do I need to be conscious of myself as, in any stronger sense, a bodily being? And why couldn’t I still think of different experiences as mine, as belonging to one and the same self, even if I lack bodily awareness? If I were placed in a sensory deprivation tank would that wipe out my ability to think of myself as a single subject of diverse experiences? What has bodily self-awareness got to do with consciousness of one’s identity? In Self and World I tried to answer these questions but my answers seem to have persuaded few readers. Contrasting my position with Kant’s, Longuenesse writes that when I attach the Kantian “I think” to my representations, the subject to whom experience is ascribed as mine is the referent of “I”, whatever it is, but ‘there is no necessity to think of this subject as a physical object’ (2006: 293) and no need to experience ourselves as shaped, extended and solid. When I was writing Self and World John Campbell even questioned the need for me to be aware of myself as located among the objects to which my experiences relate. To make his point he introduced the notion of a ‘film subject’. A person watching a film is conscious of the actions and events being portrayed as mind-independent but has no sense of himself as located in the space in which those actions and events occur. Even if the film subject is an embodied being and is aware of his body as shaped, extended and solid such awareness plays no role in his awareness of the world he experiences as an objective world. These and other similar examples bring out how difficult it is to establish substantive necessary conditions of self-conscious experience. Why should arguments that try to establish experiential or ‘intuitive’ necessary conditions be any more immune to counterexample than ones that try to establish conceptual or doxastic conditions? Both types of argument face the same problem: for any putative necessary condition someone is going to have the intuition that experience is possible even if that condition isn’t fulfilled. One always has the option of attacking the intuition but who is to say in such cases that it is the intuition rather than the argument that is untrustworthy? This is not a question to which, in Self and World, I offer a satisfactory answer or, indeed, any answer. If there is a reason for thinking that ‘experience-directed’ transcendental arguments are any better than belief-directed arguments I don’t say what that reason is. On reflection, however, it’s not clear why necessary conditions are important anyway. Necessary conditions are certainly relevant where anti-sceptical transcendental arguments are concerned, that is, where the aim is to show that the truth of some proposition P which the sceptic doubts is necessary for the truth of some other proposition which the sceptic doesn’t doubt. But for the most part in Self and World I’m not arguing with the sceptic. Who, then, am I arguing with? It’s clear from the beginning of the book that my primary target is the view that the self is, in some important sense, systematically elusive from the perspective of self-consciousness. Hume and Kant both defended versions of this ‘elusiveness thesis’ and even 20th century materialists have endorsed it. According to Shoemaker, for example, ‘when one is introspectively aware of one’s thoughts, feelings, beliefs and desires, one is not presented to oneself as a flesh and blood person, and does not seem to be presented to oneself as an object at all’ (1984: 102). If this is one’s target then there is no need to show that introspective awareness of oneself as a flesh and blood object is strictly necessary for consciousness of one’s own identity or for experience of mind-independent objects. It is enough that the type introspective awareness that Shoemaker describes is possible and, at least in some cases, actual. All I really needed to show, and I think did show, is that with the exception of those in sensory deprivation tanks or victims of other sorts of sensory deprivation, we are introspectively aware of ourselves as abiding material substances, as physical objects among physical objects. If such awareness is also necessary for experience of objects or whatever then so much the better. As far as refuting the elusiveness thesis is concerned what matters is what is actually the case, not what has to be the case. I think I focused on necessary conditions because, in the neo-Kantian tradition in which I was brought up, necessary conditions were always the objective. I said at the outset that I saw myself and breaking from this tradition but it clearly wasn’t a clean break, and my methodology was much more ‘transcendental’ than it needed to be. In chapter 1 of Self and World, which was the last chapter I wrote last, I presented the idea that the self cannot be an object for itself as part of an argument for the thesis, which I labelled the ‘Exclusion Thesis’, that the thinking, experiencing self is not an object among others in the world and therefore not a physical object among physical objects. My argument against excluding subjects from the world is not a transcendental argument. Instead, the focus is on the role of abstraction in sustaining the Exclusion Thesis. For the self that supposedly is not part of the world is not the person or human being but what Wittgenstein describes as the ‘philosophical self’ or the ‘metaphysical subject’. For Husserl it is the ‘transcendental Ego’ that has no place in the world and my objection to all these claims is that they are much less surprising and less interesting than they appear once it is understood that the philosophical or metaphysical or transcendental self is nothing more than an abstraction from the personal or human self. There are about as many different ways of abstracting the philosophical from the human self as there are philosophers interested in performing this feat of abstraction but the basic point is that the exclusion of the philosophical self from the empirical world is ‘no more surprising than the exclusion from the empirical world of anything that is essentially abstract or formal’ (1997: 20). I stand by this argument, about which there is nothing especially ‘transcendental’. It’s just an argument, and I was struck on rereading Self and World how often I make the point that Kant-inspired views of the self to which I object always seem to be in the business of abstracting: sometimes it is the transcendental self that is abstracted from the empirical self and sometimes it is transcendental self-consciousness that is abstracted from empirical consciousness. My view was, and is, that these abstractions are deeply suspect. Despite my insistence here and in the book that my approach to the self and self-awareness is different from Strawson’s there is no question that the book has a strong neo-Kantian flavour. The Bounds of Sense sets the agenda for the book even though I end up not agreeing with it on many issues. Campbell and I met for lunch every two weeks during term over a long period and in the early days these lunches were followed by a couple of circuits of Christ Church meadow. I’ve lost count of the number of time the conversation reverted to The Bounds of Sense, pages 97-112. Those page numbers are still etched on my brain because we both thought that there is where the deepest action is in the book. As is well known, Kant in the Transcendental Deduction related the unity of consciousness to experience of objects, sometimes seeming to argue from experience of objects to unity of consciousness and sometimes from unity of consciousness to experience of objects. In The Bounds of Sense Strawson tried to reconstruct and defend both arguments and these arguments were also the inspiration for the central chapters of Self and World. Chapter 2 is my take on the argument from objectivity to unity and chapters 3 and 4 are my take of the argument from unity to objectivity. Perhaps it isn’t surprising, therefore, that Self and World is seen as a neo-Kantian book. The fact remains, however, is that while my discussion is structured along Kantian lines, I don’t follow Kant in representing the unity of consciousness as a purely ‘formal’ unity and, as Longuenesse’s discussion brings out, many of my theses and arguments are ones that Kant and Strawson would have rejected. As I said at the beginning, it’s a very interesting question why the kind of Kant- and Strawson-inspired philosophy that was so central to Oxford philosophy faded so rapidly in the early 2000s. Partly I think it is because this kind of philosophy never really crossed the Atlantic, except in the person of John McDowell, though it’s telling that even McDowell talks more about Sellars than about Strawson. American Kantians have always been much more interested than I was in what Kant actually meant, and people specialising metaphysics and epistemology were more interested in David Lewis. Doing Kant inspired epistemology but not engaging in Kant scholarship made one a member of a very small club. It resulted in the feeling, which I have never entirely shaken off, that one’s work always ends up falling between two stools: not Kantian enough for the serious Kantians and far too Kantian for everyone else. In addition, one should never underestimate personal factors in a person’s philosophical influence. When I was an undergraduate the professors in Oxford were Strawson, Dummett, Ackrill and Hare. I read a huge amount of Hare, who totally dominated the moral philosophy syllabus at all levels. When he retired, his work largely disappeared from view. With the passing of the generations, Oxford philosophy is under new management, and it is no longer safe to assume that graduate students there or elsewhere in the UK will have immersed themselves in Strawson, Dummett and Evans, as I did. The retirements of Strawson and Dummett and departures in a relatively short space of time of Peacocke, McDowell, Campbell, Brewer and me represented a dramatic changing of the guard. My own interests have evolved too, and my current work in epistemic vices couldn’t be more different from Self and World. Looking back on work one did a long time ago it’s natural to ask whether one would do anything differently now if one were writing on the same subject. The answer in my case is a resounding ‘yes’. One thing that’s worth noting about Self and World is that it is clearly the work of someone young. What gives the game away isn’t just its defensiveness but also its basic thesis. I wrote Self and World in a single four month burst in the summer of 1995. I was 34 then and had the idea that all you needed to do to write a book is write a thousand words a day, five days a week for sixteen weeks. That is more or less what I did, the writing done by lunchtime, the afternoons spent watching cricket in the Parks and the evenings ruined by editing my morning’s work. There didn’t seem anything strange about the idea of identifying oneself so closely with one’s body. I didn’t want to say, with Bernard Williams, that persons are bodies because that sounds needlessly paradoxical and reductive. On the other hand, I did want to say that one is aware of oneself as a bodily being, a corporeal object among corporeal objects, and in this sense identifies oneself with one’s living body. What this underestimates, I think, is the extent to which it is possible to become alienated from one’s body or bodily self as a result of ageing or physical disability. It is easy to identify with a healthy body, but identifying with a malfunctioning or decaying body is a different matter. Imagine the 90 year old in constant pain as a result of arthritis, limited mobility and the sense, which many older people report, of no longer recognising the face the see in the mirror every morning as their own. They haven’t changed, they say, but their body has. From being an expression of their agency, the concrete form that their presence in the world takes, their body has become a hindrance, something that no longer responds to their instructions and stops them from doing what they want to do. In these circumstances I can see how one’s body might be viewed and experienced as an alien object. Perhaps not surprisingly, these possibilities and were not at the forefront of my mind as a 34 year old. They aren’t at the forefront of my mind now for that matter but they are not as remote as they once seemed. It’s important, of course, not to exaggerate the significance of the phenomenon of bodily alienation. One might hold that the sense in which one can become alienated from one’s own body is compatible with its continuing to serve as the presented subject of experience. I argue in Self and World that the bodily self presents itself to itself as ‘the subject of thought, sensation, and perception’ (1997: 1997) and it is arguable that this would still be true if one’s body has decayed or is in the process of decaying. The bodily self-awareness I was discussing in the book is far too basic and primitive to be called into question by the phenomenon of bodily alienation except in peripheral and highly unusual cases of complete or near complete sensory deprivation or what I refer to as ‘body blindness’. Part of the problem here is to understand in what sense it is and is not possible to become alienated from one’s own body, not just in cases of brain damage or physical injury but in the normal process of ageing. This issue merits more detailed consideration than I gave it in Self and World. I want to end by saying again what a pleasant surprise it is to find that there is still some interest in the musings of my 34 year old self. I don’t find it especially easy to identify with the author of Self and World. My overall assessment of the book is that it is not a something but not a nothing either. It subjects a range of well-known Strawsonian or neo-Kantian arguments to a searching examination and finds them wanting. My criticisms of those arguments still seem to me to be quite forceful even if the ‘experience-directed’ transcendental arguments that I put forward are far from unproblematic. As Longuenesse notes, the official aim of the book is to show that self-consciousness requires bodily awareness but I end the book with an apparently weaker claim to the effect that self-consciousness is ‘intimately bound up with’ awareness of the subject as a physical object in a world of physical objects. I should have stuck to explaining and defending this weaker claim, which I still believe is correct. 18