Is there a Paradox of Self-Consciousness?

by Andy Hamilton

Dept. of Philosophy
Durham University
Durham DH1 3H

1. Undermining the Paradox

José Bermúdez' book attempts to resolve what he terms "the paradox of self-consciousness" - the apparent strict interdependence of self-conscious thought and linguistic self-reference. He attempts this resolution from the standpoint of "philosophical naturalism". This latter position is left largely undefined, but apparently involves the requirement that it must be explained how self-consciousness is acquired by the subject in the normal course of psychological development. The boundary between conceptual and psychological enquiry is fluid, and empirical work in scientific psychology can help to sustain or disprove philosophical claims, Bermúdez believes (pp. xi-ii).< 1 >

The first part of the book deals with the Paradox in general terms through a discussion of the nature of concepts and a defence of the idea of nonconceptual content. The second part, from the discussion of Somatic Proprioception onwards, discusses bodily awareness and the embodied subject directly; a third part treats psychological self-ascription, followed by a resolution of the paradox. I found much of the discussion very illuminating and there is much I agree with - notably the centrality of proprioception to self-consciousness, and the analysis of circularity concerning "I". But I disagree with the resolution of the paradox, with the naturalist assumptions underlying that resolution, and most fundamentally with the claim that there is a paradox in the first place. I will first outline my disagreement with the overall project, and then discuss issues arising from Bermúdez' account of proprioception. I do not at this stage have a firm rival analysis to offer, but I will offer some considerations which favour a more humanistic approach.

The first strand of the paradox of self-consciousness, Bermúdez claims, is that there seems to be no account of how someone could acquire a grasp of the first-person pronoun, without having some prior first-person knowledge: "Any theory that tries to elucidate the capacity to think first-person thoughts through linguistic mastery of the first-person pronoun will be circular, because the explanandum is part of the explanans..." (p. 16). Anscombe is wrong to claim that the self-reference rule is circular as it stands, he believes; to say this is to confuse the semantics and pragmatics of "I". < 2 >But there is an explanatory circularity, and also a capacity circularity, Bermúdez argues: the capacity for reflexive self-reference by means of "I" presupposes the capacity to think thoughts with first-person contents, and so cannot be deployed to explain that capacity. That is, a degree of self-consciousness is required to master the first-person pronoun, and vice versa.

This capacity circularity helps to generate the paradox of self-consciousness, Bermúdez believes; to dissolve the paradox, he wishes to reject the claim of capacity circularity. He is not opposed to "local holisms" as such. However, if the various abilities at the root of self-conscious thought form a local holism and consequently can only be explained in terms of each other, then it is impossible to explain how these abilities can be acquired in the normal course of cognitive development, he argues. Thus what Bermúdez terms the Acquisition Constraint is violated (p. 19). On his view, it becomes impossible to explain how young children can "bootstrap themselves into self-conscious thought". If there is complete interdependence between reflexive self-reference and first-person thought, it follows that there are no cognitive abilities more basic than self-consciousness, in terms of which it can be explained: "The acquisition of self-consciousness then becomes a complete mystery. Self-consciousness depends upon the capacity to take a first-person perspective (to think first-person thoughts), but we have no understanding of how entry into that first-person perspective is possible" (p. 123).

These claims are highly questionable. It is not clear why the circularity between the capacity to self-refer using "I", and the capacity for first-person thoughts, should cause a problem, yet other capacity circularities do not. There surely is a local holism here, and it is not clear how it violates the Acquisition Constraint. It is simply that any explanation will have to refer to the simultaneous acquisition of the capacities; or alternatively it could be argued that one must refer to the acquisition of a single complex capacity. An analogy would be with personal memory and factual memory. These are conceptually interdependent. It is hard to conceive of someone whose knowledge of their own past is derived entirely from merely factual memory - if they could not retain personal memories, how could they retain information that they were told? Indeed, it is difficult to conceive how they could so much as understand testimony concerning their own past - that is, the personal memory-reports of others. But the fact that the two capacities have to emerge together does not mean that they are inexplicable.

Surely Bermúdez should be arguing that primitive self-consciousness does not require the use of either the first-person pronoun or the thinking of genuinely first-person thoughts. As it stands his claim that the capacity to think genuinely first-person thoughts does not require any linguistic or conceptual capacities seems close to a self-contradiction (p. 27). Bermúdez would say that this is just a sign of a deep unreflective commitment to the Thought-Language Principle, the principle that the capacity to think a certain range of thoughts must be analysed by analysing the capacity for their canonical linguistic expression (p. 13). But it may be that when Bermúdez talks of "having first-person thoughts" he means "being sensitive to self-specifying information". For instance, he seems to equate "primitive first-person contents" implicated in perceptual experience at the ecological level with the presence of "self-specifying information" (p. 163). This is not a notion I find unproblematic, however.< 3 >

2. Naturalism and Psychology

The suggestion of a paradox requiring resolution arises from Bermúdez' confidence that self-consciousness is a well-defined and unified concept, and that there is a determinate concept of a "first-person perspective". In fact, self-consciousness is a theoretical philosophical construct which is open to a variety of interpretations. A creature is described as self-conscious when its behaviour has attained a certain level of complexity, and self-consciousness thus involves a range of capacities, as I will argue below. (Discussion of non-indexical self-reference will also cast doubt on the idea of determinate first-person perspective.) However, it is natural to think that the use of a device like the first-person pronoun is central to self-consciousness. The problem lies in defining what "a device like 'I' " really is like. The essence of "I", I would argue, is more to do with immunity to error through misidentification (IEM) - and it is a merit of Bermúdez' account that he gives due consideration to this phenomenon - than indexicality.< 4 > In fact I am torn between defining higher-level self-consciousness in terms of IEM, and denying that there is a determinate phenomenon of self-consciousness to be tracked.

The idea of a determinate phenomenon of self-consciousness is sustained by Bermúdez' naturalism and his attitude to psychology. Bermúdez sees Philosophy and Psychology very much as on a continuum. I think this is a mistake. Philosophy is essentially a conceptual investigation concerning presuppositions of enquiry. Scientific psychology has to respect those presuppositions in its endeavours. This is not a claim about relative importance of the disciplines, simply a matter of a distinction between conceptual and empirical enquiry. For instance, psychologists sometimes say that the distinction between personal and factual memory - or the various psychological analogues of that distinction - is one that is supported by psychological theory. But it is, or should be, a datum for such theorising, not subject to experimental confirmation or disconfirmation. Likewise, the analysis and explanation of self-consciousness are distinct if related questions. Psychology can, of course, help to illustrate philosophical claims, and much of Bermúdez' psychological discussion does this very interestingly.

Although he is not uncritical of the philosophical assumptions underlying the experimental work of scientific psychology, Bermúdez is still far too charitable. An example is his discussion of the false-belief task - where a child is apparently unable to recognise the false belief of another child concerning the location of an object such as a bar of chocolate. Possessing the concept of belief requires that one comprehend the possibility that another subject has a false belief, and on the experimental evidence of the false-belief task, this is not achieved until the child is approaching four years old. Bermúdez claims that "the central that no subject who fails the false-belief task can be ascribed the concept of belief" (p. 280). It does not need experimental evidence to show that this is questionable. Hiding games enjoyed by young children surely involve something like the thought "Mummy does not know where I am", or "Mummy thinks I'm in the living room but I'm not". 10-month Tracey is regarded by Bermúdez as having protobeliefs such as "Mother recognises what I have set out to do". Rather than regarding protobeliefs as a distinct, independent category, it would be better to say that there is evidence that Tracey believes that her mother... etc. Tracey is showing clear signs of recognising the existence of other minds, which is what is involved in possessing the concept of belief.

3. Nonconceptual content and levels of self-consciousness

Although I have questioned whether there is a paradox, I will say something about the attempted resolution. An account of self-consciousness in relation to animals and neonates is, after all, required in any serious treatment of self-consciousness. In his critique of the Thought-Language Principle in Chs 2-4, Bermúdez criticises the "classical view of content", and in particular the Conceptual Requirement Principle: the range of contents that one may attribute to a creature is directly determined by the concepts that the creature possesses. He believes that it is plausible to maintain that: "there are ways of representing the world that are nonconceptual, in the sense that they are available to creatures who do not possess the concepts required to specify how the world is being represented" (p. 43). Thus is opened up the possibility that there are "nonconceptual first-person contents" that do not require mastery of the concept "I" and the first-person pronoun, and so the paradox of self-consciousness may be resolved.

Although I recognise that this way of talking, in terms of "representing the world" and "nonconceptual content", is widespread, what Bermúdez refers to as the familiar conception of content, that the only kind of content is propositional content, seems to me correct. One needs to examine what proponents of nonconceptual content think they are defending. Crucial to the claim of nonconceptual content is the idea of "representing the world". This is a rather underdeveloped notion in Bermúdez' treatment. He comments on the need to give an account of what it is for a creature to be representing the environment. A creature must be credited with this capacity when its behaviour cannot be explained simply by how things are in the immediate environment, but how they are taken to be (pp. 85, 88). Bermúdez talks of "primitive forms of inference" concerning elements in the representation where conceptual content is not involved (pp. 92-3).

When an explanation is required in terms of how things in the environment are taken to be, I would argue, we are in the region of beliefs. There is no clearly definable intermediate category of nonconceptual contents. Take the well-known example of Chrysippus' hunting dog. In pursuit of its quarry, it reached a point where the path branched three ways. It sniffed the first path, then the second, each time drawing a blank; then took the third path without sniffing. Chrysippus argues that the dog must have engaged in some simple reasoning - "Either A or B or C; but not A, and not B; therefore C". < 5 >That seems a reasonable thing to say. It is more plausible than the more theoretical discussion Bermúdez offers in terms of a transition from nonconceptual to conceptual content which in part involves moving from partially recombinable to globally recombinable content constituents. This seems to me cognitive science speculation. I am not convinced that Bermúdez shows that there is another fundamental kind of explanation of behaviour apart from those in terms of either the attribution of propositional attitudes - however analogical - or of stimulus and response.

Clearly much more needs to be said on the question of animal beliefs and inferences. It does seem right to concede a lower-level self-consciousness possessed by non-language-users. But I would argue for the "humanistic" claim that higher-level self-consciousness is primary; other varieties constitute proto- or primitive self-consciousness, and there is a continuum between them. The primitive level will involve some kind of self-recognition or self-awareness. This sense of "self-conscious" conforms with everyday uses of the term. For instance, someone who is highly self-conscious is highly aware of the impression they are making on others, and of others' reactions to them.

It seems that chimpanzees have primitive self-consciousness to a more marked degree than other monkeys. A striking illustration of the way in which the chimp can view itself as a chimp among other chimps appears to be the way it can be trained to recognise its own mirror-image. Chimps will notice in the mirror a spot of paint placed on their forehead, and rub it. Monkeys, in contrast, treat their images as they do other monkeys, usually by threatening them. < 6 >Presenting the chimp with taped sounds of its own calls would be another test of self-recognition. Indeed, Bermúdez cites the results of just this test in the case of newborn infants. Day-old infants cried significantly less when played a recording of their own crying than that of another infant; he claims that this is "a clear case of self-perception. Infants can discriminate their own crying from the crying of other infants of the same age" (p. 125). Whether or not this is a clear case of self-perception, it interestingly parallels the example of the chimps.

Self-recognition in a mirror is, however, only a particularly apt expression of self-consciousness in chimps. Their degree of self-consciousness is illustrated in a whole range of intelligent behaviour: an ability to deceive (hiding bananas), to plan strategies (of becoming dominant), of gaining help in these strategies. Monkeys perhaps exhibit such behaviour to a lesser extent. The concept of self-consciousness is part of a framework of concepts involving intelligent or intentional agency and rationality.

The connection between self-consciousness and the linguistic or grammatical form which gives it expression is a deep question central to the relation of thought and language. It is right to suppose that higher-level self-consciousness could not exist prior to or independently of some such expression. But there are grey areas. Children can use their own names as devices of self-reference before they grasp "I". When Jacob says "Jacob wants a biscuit" his utterance may plausibly be interpreted as "I want a biscuit" - for instance if he is upset when not given a biscuit, and so on. We interpret Jacob as self-consciously self-referring, because although he lacks a grasp of the normal grammatical expression, he seems to be on the verge of acquiring it. Comprehending use of the term "I" goes together with a pattern of intentional behaviour which exhibits self-consciousness. Given the behaviour without the use, one may sometimes ascribe the ability to self-consciously self-refer; conversely, given the use without the behaviour, one may decline to ascribe self-conscious self-reference.

4. Eliminating indexicality

The possibility of eliminating indexical self-reference throws the whole debate concerning self-consciousness into unclarity, however. Yet it does seem plausible to say that self-conscious self-reference may be achieved without the use of an indexical term. In proposing this view, I am not directly criticising Bermúdez's account - although I think it causes problems for it - but suggesting complications for the traditional view that self-consciousness is expressed by means of a device like "I".

Definite descriptions, like proper names, may be used self-consciously to self-refer. "The present writer" is used by the present writer in order to avoid the intrusive "I". Certain historical personages have self-consciously self-referred using their own name - Julius Caesar, Henry James, De Gaulle, indeed Andy Hamilton on his ansaphone, and in the present sentence.

Thus there seems no reason why this use of one's own name - or of a special, speaker-restricted name used only by oneself - could not become a general or universal practice. There could then be an equivalent to the self-reference rule, a rule that each one applies, generating a term that may be used knowingly and intentionally to self-refer, viz: "Their own name is the word which each one uses in speaking of him- or herself". The result may be termed a community of name-users. There is a uniform rule across persons, but not a uniform term; the speaker's name would not count as an indexical, since its reference does not vary with the context of utterance, viz. who uses it. "De Gaulle" always refers to De Gaulle, but only De Gaulle can use it as a device of self-conscious self-reference. Speakers would use their own name to express first-person thoughts.

The scenario has more or less radical variants. "I" may be abolished; or most radically, all personal pronouns and 1st-, 2nd- and 3rd-person singular and plural verbal inflections are dispensed with in favour of a uniform "3rd-person". In each case, except for a certain convenience arising from the presence of a uniform term, nothing essential seems to have been lost. It is true that "I" has an unambiguous reference, such that its use enables the audience to identify the referent in a way that the speaker's use of their own name may not. Thus there would be confusion in situations where I did not know the name of the other person, and so did not know whether they were self-referring by means of their own name or referring to someone else of that name. But in a small community where everyone knew everyone else this would not be a problem. Perhaps each person could have two names, one for self-reference and one for other-reference. If a stranger appeared in the community, they could receive the name "Stranger", or "Stranger 1" and "Stranger 2", and so on.

A different kind of drawback arises if I forget my own name. It will then be essential for me to make one up, or to say "This person here", pointing to myself. "This person here" will substitute for "I" only when supplemented by a pointing gesture at oneself; it is not clear, therefore, that it is a true indexical. Even "The present speaker" may be unclear, owing to the indeterminacy of "present".

It may seem that in forgetting my own name, I have lost a piece of empirical knowledge, while in forgetting "I", I have lost a species of linguistic competence, the capacity self-consciously to self-refer. But a speaker might simply forget the word "I". Brain-damage has bizarre effects, and one could imagine a victim asking "What's that word, you know, the one for 'This person here [pointing to himself]'?". So in both our present notation and the name-user notations, it is possible that the speaker forgets the word, but has not lost the linguistic competence involved in self-referring. In the community of name-users, the need for a new name in cases of amnesia would be more urgent - that is all.

How are names learned in these novel scenarios? Where only "I" is abolished, "I am AH" could become "Name is AH", and "What is my name?" could become "What is name?" But assume the more radical scenario where all personal pronouns are abolished, and 1st-, 2nd- and 3rd-person singular and plural verbal inflections dispensed with. Then it is hard to see how the knowledge "I am AH" could be expressed, except perhaps in the form "The person uttering this very sentence is AH". But it is not obvious that knowledge of one's own name has to exist in propositional as well as practical form. The name-users could simply cotton on to people's names, including their own - as young children do before they acquire the use of personal pronouns, and as indeed they do in grasping these also.

The possibility of non-indexical self-conscious self-reference reduces a general reflexive analysis, as proposed by O'Brien and others, to emptiness. < 7 >For it is the intention to self-refer that the appeal to a general device of self-reference was meant to explain. In the case of "I", the intention seems built in because one does not have to distinguish self-consciously self-referring uses of the term from other uses; there are no other comprehending uses. This seems to be the unique feature of indexical self-reference. Oedipus understands the definite description without realising that it refers to him. But a speaker does not have any understanding at all of "I" if they do not realise that it is a device of self-reference; it is unambiguously self-referential. "I" is exclusively a device of self-reference, and it is this difference which makes generality appear more important than it is.

However, there is a further, rather elusive consideration which may suggest that the requirement of indexicality has not been dispensed with after all. One could claim that the scenario where there is no general device will still have to be interpreted in terms involving "I". For the name-users, the person's name apparently does all the work formerly performed by personal pronouns and verb inflections. But as interpreters we are still imputing understanding of an analogue of the 1st-/3rd-person distinction. Note for instance that it was said of the name-user scenario that speakers use their own name to express first-person thoughts.

These considerations illustrate a fundamental problem for thought-experiments which attempt to eliminate some apparently essential linguistic device. The name-user scenario tried to achieve this by separating grasp of the first-person device, and the intention to self-refer. If this attempt is successful, one may conclude that all that is necessary to interpret primitively self-conscious animals is the employment of a uniform rule of self-reference; a uniform device is not required. We would have to interpret the creatures as if they had self-referring thoughts, without these being first-person ones. But in the case of "I", the device and the idea of self-conscious self-reference seem to be intertwined, and it is not clear what is left behind when the device is imagined away.

But whatever one may think about whether a self-conscious creature has to use a device of self-reference, it is plausible to say that it is impossible to interpret a creature as exhibiting self-consciousness without using such a device in ones interpretation. Thus, to return to the earlier example of the chimp recognising itself in the mirror, the best explanation one can give of the chimp's behaviour will involve the use of "I"-sentences, or indirect reflexive pronouns cognate with "I". What the chimp realises is: "I have a spot on my forehead". But the question of interpretation versus use remains deeply problematic.

5. Proprioception and self-consciousness

In Bermúdez' book it is naturalism rather than materialism which takes centre-stage. Concerning proprioception, the work of psychologist James Gibson is key inspiration. But the influence of what Cassam in his recent book calls "materialism concerning self-consciousness" - a position owing a lot to Gareth Evans and indirectly Husserl and Merleau-Ponty - is also felt. However, there is no actual mention of Husserl and Merleau-Ponty. It seems to me that Gibson and the phenomenological tradition each offers great insights into the nature of proprioception, and it should be a task of a philosophical account of the body and self-consciousness to attempt to reconcile them, insofar as this is possible.

Bermúdez is inclined to reject what he terms the Schopenhauer/ Wittgenstein view that the self cannot be perceived, and has no place in the content of perceptual experience. He wishes to subscribe instead to a Gibsonian perspective whereby "the self is the limit of the visual field...[but] it is precisely in virtue of this that the self features in the content of perception. The self appears in perception as the boundary of the visual field, a moveable boundary that is responsive to the will" (p. 106). "The crucial thought is that even if it is conceded, pace Gibson, that the self is not directly perceived, it is still the case that the self has a place in the content of perceptual experience in virtue of the self-specifying information that is an integral part of that perceptual experience" (p. 108).

Bermúdez interprets Gibson's "ecological optics", his theory of perceptual invariants - higher-order patterns that remain constant during change in the perceiver's field of vision caused by their own movement or that in the environment - in terms of what he calls "self-specifying structural invariants in the field of vision". Gibson regards vision as kinaesthetic in that it registers body movement as much as the muscle-joint-skin system and the inner ear; like these, "vision obtains information about both the environment and the self" (Gibson (1979), p. 183). Bermúdez concludes that "information specifying the movement of the perceiver is present in visual perception". Since this is self-specifying information, the self has a place in the content of visual experience. Perceptual experience does not only provide information about the external world: "information about the ambient environment is inextricably combined with self-specifying information, without which the former would be of little use" (p. 114). Bermúdez believes that Gibson's ecological approach to perception shows that perceptual experience is a source of first-person nonconceptual contents.

Setting aside the doubts about a determinate concept of self-consciousness discussed earlier, I am sympathetic to Bermúdez' view that proprioception counts as a "genuine form of self-consciousness" (p. 144), and indeed with the idea that there is a close connection between IEM and the essence of first-person judgments (also p. 144; and elsewhere). His Gibsonian picture leads to some persuasive arguments in favour of the view that proprioception is a modality of perception. Bermúdez considers the constraints that perception must involve identification or tracking of objects over time; and that ordinary modes of perception involve the perceiving of a multiplicity of objects. His discussion of proprioceptive content is fertile and suggests many questions for further discussion, especially concerning the dual criteria for location and the concept of a hinge (pp. 154-61).

But there is surely at least a tension between his espousal of Gibson's picture of proprioception as continuous with other perceptual modalities, and the claim that proprioception is a "form of self-consciousness". "Self-consciousness" would have to include consciousness of ones immediate environment, and "the self" would as a result seem to be a very fluid concept.

The fact that proprioception constitutes knowledge of ones body makes it seem innocuous to say that it is a mode of perception. However, the perceptual model carries implications in tension with IEM and the phenomenological concept of a body-subject or lived Body (of which more below). Despite appearances it is, I think, possible to locate a principled distinction between proprioception and sensory perception. It is not enough to point out that there is no organ of proprioception except the nerve-receptors, for the same is true of exploratory touch, one of the five senses. A more promising suggestion is that no kind of action is required in order to gain knowledge in the case of proprioception.

The absence of the need for action is expressed in the fact that I just know when my legs are crossed, and by a particular use of the term "feel". This is really the crux of the matter. As noted earlier, one can always become aware of specific feelings in proprioception, but not in the sense required by a "body-image" or a perceptual account. Feelings of pressure, touch, muscle tension or skin stretching, or tension in the shoulder are not felt in a perceptual sense, nor should they be described as, for instance, "my legs feeling crossed". There are different senses in which a region is felt. "How did you know your legs were crossed?" To answer "I felt they were crossed" implies an action of touching, which is wrong. Compare "How did you know there's a bump on your knee?" "Because I felt it". "They felt (as if they were) crossed" is no better, since it implies a feeling distinct from the knowledge that my legs are crossed. And do my legs really "feel as if they are crossed"?

The correct answer to the - purely theoretical - question "How do you know that your legs are crossed?" is "I just know". Wittgenstein (1980) gives a subtle and interesting justification of this view:

      My lower arm is now lying horizontally and I should like to say I feel that; but not as if I had a feeling that always goes with this position (as one would feel ischaemia or congestion) - rather as if the 'bodily feeling' of the arm were arranged or distributed horizontally... So it isn't really as if I felt the position of my arm, but rather as if I felt my arm, and the feeling had such and such a position. But that only means: I simply know how it is lying - without knowing it because....As I also know where I feel pain - but don't know it because.....< 8 >

The crucial claim is that it is "as if the 'bodily feeling' of the arm were...distributed horizontally". "I felt my arm" means something like "My arm wasn't asleep/ frozen/ anaesthetised".

These remarks suggest that the idea of an object of perception should be replaced by that of a subject of proprioception. Certainly it is at least the case that proprioception has an ambivalent status; it involves knowledge of an object distinct from the awareness of it, yet that object is comprised within the body-subject in Merleau-Ponty's sense. Proprioception offers an intermediate case between perceptual knowledge from the five senses, and the case of pain and other sensations.

So the perceptual model is wrong to claim that I am simply moving my attention from objects that lie outside one of my bodily boundaries to what is going on at or beneath that boundary. Its proponents usually acknowledge that proprioception is unique among modes of perception in providing a way in which one can come to be aware of only one object and its parts - unless they are proponents of q-proprioception - but they fail to explain why this should be. I am not convinced that Bermúdez avoids this fundamental problem.

The crucial discussion of these questions is found on pp. 137-51 of The Paradox of Self-Consciousness. In defending the perceptual model, Bermúdez argues that it meets identification and multiple-object constraints. But satisfaction of the latter constraint, while not allowing the possibility of q-proprioception - which Bermúdez firmly rejects (p. 144) - does give an implausibly broad extension to the term self-consciousness. More importantly, we need to know how a perceptual model can explain the IEM status of ordinary judgments of posture and position. This is not a problem with the account I would favour, an account which asserts the conceptual interdependence of proprioception and bodily identity.

6. The lived Body and the mere body

In this final section I will be concerned to elucidate this claim of conceptual interdependence. Partly it is simply a matter of a defence of the IEM status of bodily self-ascriptions. < 9 >But the idea of interdependence further amounts to the anti-materialist, and probably anti-naturalist, claim that "my body" is that of which I alone have knowledge and am able to move in an IEM-exhibiting way. The phrase "my body" is intended to emphasise the subject's knowledge of and concern with their own (living) body.< 10 > It will be some improvement to say, still using the terminology of the traditional distinction between mental and physical, that the body is a psychological as well as a material unity. But as Merleau-Ponty argued, "The experience of our own body...reveals to us an ambiguous mode of existing" - neither thing nor consciousness. < 11 >I will outline an account of the body and proprioception which is influenced by the phenomenological tradition, though it disagrees with it in important respects. I will argue that "my body" is the primary sense of "body", and that this implies that the body is neither simply a mental nor a physical entity.

Husserl was probably the first philosopher to recognise the importance of proprioception and the physical-intentional dichotomy concerning the body. It is worth considering his account for the crucial things it gets right and also for its questionable aspects. In Ideas Book II he claims that the human body is defined by intentional attributes of action and proprioception as well as spatio-temporal material attributes. He distinguishes two aspects, der Leib, the "animated flesh of an animal or human being" - the "lived Body" or what I have referred to as "my body" - and der Körper, "inanimate physical matter", the "mere" body or body viewed purely as a physical object. < 12 >He writes:

      ...the Body is originally constituted in a double way: first, it is a physical thing, matter; it has extension, in which are included its real properties, its color, smoothness, hardness, warmth...Secondly, I find on it, and I sense "on" it and "in" it: warmth on the back of the hand, coldness in the feet, sensations of touch in the fingertips...

      [The Body is] a material thing which, as localization field for sensations and for stirrings of feelings, as complex of sense organs, and as phenomenal partner and counter-part of all perceptions of things...makes up a fundamental component of the real givenness of the soul and the Ego.< 13 >

The lived Body is sometimes referred to by commentators as the "mindful body"; a better alternative might be "the body-in-action". < 14 >But talk of the lived Body or body-in-action is useful chiefly as indicating essential features of, simply, "my body".

Husserl's account is a great advance on those of earlier philosophers, who insofar as they considered it at all, almost universally regarded the body as a purely material entity.

But despite the efforts of recent commentators to portray him in an anti-Cartesian light, there is a Cartesian bias in Husserl's treatment which was rejected completely by Merleau-Ponty. < 15 >The metaphor of animation is a key instance. Husserl describes "my body" as "animated flesh" (and bones), and refers to the "psychological I" as that aspect of the lived Body which animates it. Husserl's claim that the lived Body is unique in "being moved 'spontaneously' or 'freely' by the will of the Ego" implies Cartesian mechanistic talk of "control". < 16 >

The metaphor is misleading because I do not, ordinarily, control or exert control over my body. In unusual circumstances I may, as when one hand is trembling but I am able to control it with the other. It would be some improvement to say that I act through or by means of my body, though even this sounds wrong. Action is, essentially, a bodily phenomenon - though this does not mean it is purely physical. This is brought out by considering the two senses of "move" as in "I am moving my arm" - one which generates IEM and one which does not. The former is a basic action in Danto's sense; the latter is not, that is, it is achieved by doing something else.

This is what was referred to earlier by the definition of "my body" as the one which I can move in an IEM-exhibiting way. In the case of the non-basic action, errors of identification are possible. It could be that coincidentally with my moving someone else's arm, someone else is moving mine, so it seems as if I am moving my arm in the non-basic sense. Or I could be paralysed, initially without realising it, and think that I am moving my arm in the non-basic way when in fact someone else is. This is one sense in which Husserl's account is Cartesian. But there is a subtler difference from the account I wish to defend, in that Husserl treats intentional attributions as made to the lived Body, whereas the present account regards them as ascribed to persons.

The claim of a conceptual circle between proprioception and the lived Body is challenged by what may be termed the Body-body problem. The problem concerns the relation between "my body" and the mere body - between bodily ascriptions to persons, and predications of human bodies which are not also ascriptions to persons. < 17 >The idea of a Body-body problem should displace that of the mind-body problem. It might be said that this is because there are now two relations to consider, between mind and lived Body and mere body respectively; but the emphasis on the lived Body brings into question the very definition of "mind" in the first place.

To say that the body is a physical object is just to say that predicates may be applied to it which apply with just the same sense to non-persons. Thus there is a sense in which it has mass, chemical constitution, size, and traces a spatio-temporal path through the world. When I say "I weigh 12 stones", "I am 6 feet tall", "My body contains X litres of H20", these predicates apply with just the same sense to non-persons - as, for instance, in "My bath contains X litres of H20". As Husserl puts it, the Body is "integrated into the causal nexus of material nature". < 18 >"Holding the block of ice made my hands numb" or "The weight of the stone pressed down on my leg" are examples of this integration. "Physical body" seems to be the primary sense of "the body" as far as physiologists are concerned; their interest is with the "mere" body.

It is not quite right to assume that the Body-body problem concerns a divide between the first and third persons. This tends to be Merleau-Ponty's assumption, though he does write: "It is indeed not enough to say that the objective body belongs to the realm of 'for others', and my phenomenal body to that of 'for me', and we cannot refuse to pose the problem of their relations, since the 'for me' and the 'for others' co-exist in one and the same world...". < 19 >But he is wrong to talk of a "for me" and "for others" in the first instance. I make physical attributions of my own body; it is also possible to treat one's body merely as a physical object. This applies not just to a part of my body, but also to my whole body. To take an heroic example, "I interposed my body between my partner and the gunman"; or "I used my bodyheat to try to prevent the child suffering from hypothermia, by holding him close to me". This is to recognise that my body has physical properties like any other animate or inanimate object - the ability to act as a barrier, possibly with fatal consequences to myself, or as a heat-producer. It is to exploit the body's merely physical properties. But it is not the case that one always does this when acting.

How is the Body-body problem to be resolved? Although clearly it does not concern the interaction of distinct substances, one may still wonder whether, like the mind-body problem, it is a pseudo-problem. Certainly, I will argue, it would be a mistake to regard "Body" and "body" as two distinct entities.

The position defended here may be better understood by considering the alternatives to it:

(1) A Cartesian mechanistic account, whether it was fully endorsed by Descartes himself or not. This simply resurrects the mind-body problem; so it is more a rejection of the terms of the Body-body problem than a resolution of it. A clear advocate of this approach is Thomas Reid, who subscribes more closely to what is traditionally regarded as a Cartesian position than does Descartes himself in some respects. On Reid's view, I am lodged in my body like a pilot in a vessel, and the body is the instrument of the mind. Descartes himself is more equivocal. Arguably he gave a scientific, mechanistic account of the interaction of psychic and physical events, and a more subtle explanation of how mind and body can co-exist in one whole subject. He maintained that although experience shows that mind and body are not really distinct, they are nonetheless conceptually distinct.

(2) A Fregean "different sense, same reference" account - the lived Body and the mere body are the same object viewed in two different ways. This response seems to be favoured by Husserl and Merleau-Ponty. Husserl writes that from a phenomenological point of view, "the physical and the lived Body are essentially different", though they are the same object. Merleau-Ponty says that the objective body and the phenomenal body are two aspects of one and the same thing. But he also says that the "body-subject" involves a part-whole relationship between physiological elements and the embodied subject; the whole is more than the sum of the parts, rather than one object with two aspects.

(3) The (possibly materialist) claim that there is one sense of "body", qualified by "living" and "dead". On this view, the distinction between lived Body and mere body simply amounts to the fact that the body has certain capacities when the person is alive, which it loses when they die. It is "body" in the sense of "corpse" which is the primary sense in which the body is a merely physical entity.

None of these options is satisfactory. The Cartesian mechanism of (1) is quite unviable. But the idea that the lived Body and mere body are the same object viewed in two different ways, though more plausible, is not correct either. Although it is right to say that I have proprioceptive knowledge, as lived Body my body is subject rather than object. A general principle concerning sense and reference must be invoked here. On the Fregean view, it is right to speak of "the Morning Star" and "the Evening Star" as having different senses but the same referent, because the methods of viewing are essentially similar - observation at different times, with Venus in different positions in the sky. The sense-reference distinction applies only to objects with the same kind of identity and individuation criteria.

Stars (Morning and Evening), persons (Cicero and Tully), and mountains (Everest and Gaurisankar) satisfy this condition; "pain" and "C-fibre firings", or "Caesar" and "0", do not. With the latter cases, there is no appropriate sortal concept that covers them both. "Object", "event" and "process" - at least in the very abstract philosophical sense in which these terms might be deployed here - are not appropriate sortal concepts. One cannot talk of criteria of identity for anything as general as "object". It is possible to say that the Morning Star and the Evening Star are the same object, but only because one can make the prior claim that they are the same heavenly body.

This leaves option (3). It is true that the body has certain capacities when the person is alive, which it loses when they die. Also in its favour is the fact that, although criteria for individuating senses are notoriously vague, "body" does not seem to be ambiguous except perhaps in the ordinary sense in which the (living) body is contrasted with the corpse. So the difference between the lived Body and the mere body is not that between two senses of "body". However, option (3) fails to acknowledge that talk of the "lived Body", "body-subject" or "bodily self" does have a use. It is a means of showing the diversity and significance for self-consciousness of what are traditionally regarded as uniformly physical ascriptions to persons. Expressions like "lived Body" are reminders that many such ascriptions - or what Strawson called "M-predicates" - are not really physical, but are in the ambiguous category outlined here. They are, however, ascriptions to persons and not bodies.

The lived Body is the primary concept; the corpse is very much the ex-lived Body. As Merleau-Ponty argues, the mere or physiological body is a theoretical construct or abstraction, particularly apparent to the subject in cases of illness where their capacities are restricted. Such is the divide between living Body and corpse that the idea that criteria of bodily identity connect the two, and give equal treatment to each, sounds odd. The criteria of bodily identity are therefore essentially dependent on those of personal identity. It is not clear that we know what we are talking about when we talk of "bodily identity".

Talk of the "lived Body" is metaphorical, a way of emphasising certain kinds of ascription to persons, and perhaps in this sense the Body-body problem is a pseudo-problem. The consequences of taking "lived Body" literally are illustrated by David Bell's presentation of key features of Husserl's concept. < 20 >It is, Bell claims, immediately expressive; is sensitive, to pain, heat and cold, etc.; has motility or power to act; and functions as the absolute point about which all spatial relations are experienced as orientated. These are all said to be properties of the lived Body - it is the lived Body which yawns, smiles or cries out, or has the power to act.

But these are all things which I do - not my body, lived or not. Perhaps it is right to say that the mere (physical) body is sensitive to heat and cold - and so I can get sunburned or frostbitten - and in a different sense that I am. But it is not my body that is sensitive to pain; rather, I am. Attributions to the lived Body are, strictly speaking, part of the range of attributions to persons. (Of course proponents of the lived Body do not deny that there are attributions to subjects which are not also attributions to the lived Body, for instance thoughts.) There is no need to postulate another bodily entity or aspect of bodies, separate from that of persons.

A similar response applies to Merleau-Ponty's concept of the body-subject. His discussion suggests that the proper distinction is not between different kinds of object - phenomenal and physical bodies - but between the body as object and as subject. Merleau-Ponty wrote: "I observe external objects with my body, I handle them, examine them, walk round them, but my body itself is a thing which I do not observe". "My body is constantly perceived [cf. observed]. It is therefore an object which does not leave me. But in that case is it still an object?" < 21 >As a corrective to Cartesian mechanism the claim has some justification; but it remains the case that I am the subject, not my body. Part of what is meant by the term "body-subject" is that central varieties of bodily self-ascription are immediate and IEM.

But am I not an embodied subject, the "me" a "bodily me"? In a sense this is undeniable. But the following claims are more puzzling than is usually allowed: "I am embodied", "I have a body", "the subject must be identified with a material object". What use does "I am embodied" have? Perhaps it just means "I am alive". The claim constitutes an attempt to deny a strictly inconceivable position - that I might be a disembodied or non-embodied subject. Of course it is true that I am AH, and that AH is a living human being. To be embodied is not merely to be related to a body; though it is not correct either to say I am identical with my body.

Clearly this account needs much further development. As I stated earlier, I think that the most persuasive account of proprioception would involve a synthesis of the Gibsonian account which Bermúdez favours, with those elements of a phenomenological account developed here. I am inclined to argue that naturalism can acknowledge the body-subject only in the most attenuated form; but that is material for another occasion.


<1> Otherwise unqualified references are to Bermúdez (1998).

<2> His analysis here was in fact anticipated by Lucy O'Brien (1994). I will question it below.

<3> For one reason, it implies an impersonal concept of information, which is criticised in Hamilton (1995) and (forthcoming).

<4> I developed a detailed account of IEM in Hamilton (1995).

<5> The example is interestingly discussed in Hyman (1999).

<6> These questions are interestingly discussed in Cheney and Seyfarth (1990), especially p. 242.

<7> This analysis claims that anyone who uses "I" with the understanding that it is a device of self- or reflexive reference inevitably succeeds in self-referring self-consciously.  It is defended by O'Brien (1994), (1995), and Rumfitt (1994).

<8> Wittgenstein (1980) Vol I, §786.

<9> This was discussed in Hamilton (1995) and is developed at greater length in Hamilton (forthcoming).

<10> It was used in this sense by Carnap in his (1967), sections 129-131 - a discussion which shows some influence from Husserl.

<11> Merleau-Ponty (1962), p. 198.

<12> Husserl (1989), for instance section 62, p. 297.

<13> Husserl (1989), sections 36 and 40, pp. 153, 165.

<14> It has also been termed the "living Body" by Cassam (1997), p. 52, and by Bell (1990), p. 208; but this has unintended biological connotations.  I am grateful for discussion on these points with Paul MacDonald, and his (1996).

<15> See B. Smith (1995), in B. Smith and D. Smith eds. (1995), pp. 406-7.  D. Smith (1995, pp. 324-6) argues that the break with Cartesian dualism in Heidegger and in Merleau-Ponty is a development of insights found in Husserl.

<16> Husserl (1989) p. 167.

<17> D. Smith uses the term "body-body problem" in his (1995).

<18> Husserl (1989), p. 167.

<19> Merleau-Ponty (1962), p. 106.

<20> He refers to it as the "living body", as does Cassam; Bell (1990), pp. 208-10.

<21> Merleau-Ponty (1962), p. 91; see also Bell (1990), p. 208.


Bell, D. (1990), Husserl, London: Routledge.

Bermúdez, J. (1998), The Paradox of Self-Consciousness, Bradford/MIT Press.

Carnap, R. (1967), The Logical Structure Of The World, trans. R. George, Berkeley and Los Angeles: University of California Press.

Cassam, Q. (1997), Self and World, Oxford: Clarendon.

Cheney, D., and Seyfarth, R. (1990), How Monkeys See The World, Chicago: Chicago University Press.

Hamilton, A. (1995), "A New Look at Personal Identity", Philosophical Quarterly, Vol 45 No 180, pp. 332-49.

------------- (forthcoming) Memory and the Body: A Study of Self-consciousness.

Husserl, E. (1989), Ideas Pertaining to a Pure Phenomenology and to a phenomenological Philosophy, Second Book, trans. R. Rojcewicz and A. Schuwer, Dordrecht: Kluwer.

Hyman, J. (1999), "How Knowledge Works", Philosophical Quarterly 49, pp. 433-51.

McDonald, P. (1996), Descartes, Husserl and Radical Conversion, Durham University: PhD, 1996; forthcoming as Descartes and Husserl: The Philosophical Project of Radical Beginnings, New York: SUNY Press.

Merleau-Ponty, M. (1962), Phenomenology of Perception, trans. C. Smith, London: Routledge and Kegan Paul.

O'Brien, L. (1994), "Anscombe and the Self-Reference Rule", Analysis, 54, pp. 277-81.

-------- (1995), "The Problem of Self-Identification", Proceedings of the Aristotelian Society, XCV, pp. 235-51.

Rumfitt, I. (1994), "Frege's Theory of Predication: An Elaboration and Defense, with Some New Applications", The Philosophical Review, Vol. 103, No. 4.

Smith, B. and Smith, D. (1995), The Cambridge Companion to Husserl, Cambridge: Cambridge University Press.

Smith, B. (1995), "Common sense", in Smith, B. and Smith, D. (1995).

Smith, D. (1995), "Mind and Body", in Smith, B. and Smith, D. (1995).

Wittgenstein, L. (1980), Remarks on the Philosophy of Psychology, two volumes, Oxford: Blackwell.

top | back to symposium index | Bermúdez's reply