The Sources of Self-Consciousness

III

The key theoretical difference between representation-dependent judgments and representation-independent judgments is that the former but not the latter involve taking at face value a representational content in which the first person features. It is natural to ask how we should understand these first-person contents. We have already seen one example of a representation-dependent judgment—namely, the judgment that I am in front of a cathedral made on the basis of a perception of the cathedral. It is very clear that my perception of the cathedral does have an important first-person dimension. An integral part of what I see when I see the cathedral is that it stands in certain spatial relations to me. But what is the relation between the content of the perceptual state and the content of the self-ascription made on the basis of that perceptual state?

Peacocke himself maintains a studied neutrality here:

We can more generally consider examples in which the thinker’s reason for making his judgment “I am F” is his being in some state, other than a belief state, which represents a certain content C as correct. In the class of examples I want to consider, the content C may, but need not, be the same as the content “I am F.” Some theorists believe that perceptual states have nonconceptual contents which are distinct in kind from the contents which feature in beliefs. What I have to say is orthogonal to that issue. In the class of examples on which I want to focus, it will be the case that the content C, even if it is distinct from the content “I am F,” is still one which stands in an implicational relation to the content “I am F.” (1999, 264)

There are, as far as I can see, three possible positions that might be adopted. The first position is that the content C just is the conceptual belief content “I am F”—that is to say, there is a unity of content between the perceptual state and the self-ascription. The second position is that the content C is a first-person nonconceptual content. According to this second position the self is explicitly represented in the content of the justifying perception in a way that underwrites, but somehow falls short of, how the self is represented in the content “I am F.” The third position is that content C is a nonconceptual content that has no first-person dimension. The self is not represented explicitly at all.

Maintaining the third option appears to place pressure on the distinction between representation dependence and representation independence. If the self is not represented at all in the justifying perceptual state then it is hard to see how the content “I am F” can be reached by taking the content of that state at face value. It would be much more like a case of representation independence, where the thinker capitalizes on the implicational relation between the justifying perception (e.g., the perception that there is a telephone on the table) and the first-person belief formed on its basis (e.g., “I am looking at a telephone”). It looks, therefore, as if we do not have a choice between the second and third options. If we do not go with the first option then the second option is mandatory.

But can we at least be neutral between the first and second understandings of the relation between justifying perception and first-person belief? Here matters are a little more complicated. According to the first understanding, it will be remembered, there is an identity of content between the target perceptual state and the corresponding self-ascription. That is to say, the self is represented the same way in the original perception and in the self-ascription made on the basis of that perception. If the identity of content thesis is true then the self-ascriptive judgment simply duplicates the content of the original perception. The perceptual content is, roughly, I am in front of the cathedral. The self-ascriptive judgment endorses that and is best represented as the propositional attitude of belief being taken to the content I am in front of the cathedral. The difference between the self-conscious judgment and the perceptual state on which it is based is simply at the level of force.

The identity of content thesis has had, and continues to have, considerable currency in contemporary philosophy and mind and epistemology. It features most conspicuously in the epistemic theory of perception, according to which perception should be analyzed in terms of the acquisition of dispositions to believe (Armstrong 1968), but it is also maintained by philosophers who reject the epistemic theory. John McDowell is a case in point:

Suppose someone is presented with an appearance that it is raining. It seems unproblematic that if his experience is in a suitable way the upshot of the fact that it is raining then the fact itself can make it the case that he knows that it is raining. But that seems unproblematic precisely because the content of the appearance is the content of the knowledge. (McDowell 1982, 213–214)

The identity of content thesis is a consequence of McDowell’s brand of direct realism—it is because what we believe is what we see that what we know is available to us through perception.

The identity of content thesis raises an obvious question. What has happened to the familiar distinction between the content of belief and the content of perception? If, as many authors (including of course Peacocke himself) have argued, there is a fundamental difference between the content of perception and the content of belief (Peacocke 1992), then how can the content of a belief merely duplicate the content of a perception? We are familiar, for example, with the idea that propositional attitude content is digital, while perceptual content is analog. Why is this not enough to block the identity of content thesis?

The simple answer is that the analog/digital distinction does not map precisely onto the perceptual/doxastic distinction. Many theorists, once again including Peacocke, have proposed that propositional attitude contents can contain elements that are fundamentally perceptual in form—what are often termed perceptual-demonstrative modes of presentation (Peacocke 1983, 1992). The example we are considering seems clearly to fall into this category. The doxastic I am in front of the cathedral is best viewed as containing the relevant cathedral under a demonstrative mode of presentation.

But adverting to perceptual-demonstrative contents in this way does not, I think, address the real worry with the identity of content thesis. In essence, the problem is not to do with how the cathedral is represented, but rather with how the self is represented. Whereas there is a certain plausibility in suggesting that there is a single perceptual-demonstrative mode of presentation of the cathedral that can stand in the content both of perception and of belief, it is hard to see how anything comparable holds for the representation of the self in perception and perceptually based beliefs. The self does not feature in perception in anything like the way in which it does in beliefs—even when the beliefs in question are perceptually based.

This conclusion follows, I think, from some basic reflections on the nature of belief. I am taking the content of belief to be an abstract object, something that can be believed by many different people and that can be expressed in a declarative sentence. It must be possible for different people, and indeed the same person at different times, to take different attitudes to the same belief content. Partly this is a matter of it being possible for you to disbelieve something that I believe. Partly it is a matter of it being possible for me to believe something that for you is merely an object of hope—or for me to come to believe that something is the case when previously I simply feared that it might be the case. It’s uncontroversial that any plausible general account of the objects of belief must allow this to be the case. And that has certain consequences. Some of them are fairly obvious. It cannot be the case, for example, that I could believe something that it is logically impossible for you to believe. But it also follows, and this is slightly less obvious, that the content of belief must be divorced from the grounds on which one might come to believe that content on any particular occasion. Let’s suppose, for example, that I come to the true belief that I am in front of the cathedral because I see it in front of me, while at the same time you come to the erroneous belief that at that very moment I am not in front of the cathedral because you have been told by someone whom you believe to be a good authority on my movements that I am in fact in the museum. There is a single belief-content here that we are evaluating in different ways—namely, the belief-content that I am in front of the cathedral. And if we are to disagree about its truth-value then it cannot be the case that what I believe is tied to, or in any way reflects, the fact that I arrived at the belief on the basis of perception—because if it was then it wouldn’t be in conflict with your belief, which you arrived at on the basis of testimony.

But if this is right then it cannot be the case that what I believe is what I see. What I believe, the content of my belief, must be something that is independent of the particular form in which my being in front of the cathedral became manifest to me. As I have stressed, the fact that I am looking at the cathedral is part of what I see—and the self is represented in perception as standing in certain specifiable relations to the cathedral associated with my physical location at the origin of a particular visual array. But then this cannot be the way in which the self is represented in my belief that I am in front of the cathedral. Any such representation of the self would be too specific to feature in a suitably objective belief-content.

IV

So how is the self represented in belief? We can remain neutral on the question of whether the first-person concept should be analyzed as the sense of the first-person pronoun (as suggested, e.g., in Evans 1982) while still noting that there are significant commonalities between how the self is picked out linguistically and how it is represented in belief (and other propositional attitudes). The key to both the first-person pronoun and the first-person concept is the type/token distinction. The first-person pronoun is a linguistic type whose meaning is exhausted by the rule specifying that any appropriately uttered token of that type refers to its utterer. Similarly, the first-person concept is a type that, when tokened in thought, plays a familiarly distinctive functional role for the subject who tokens it. Adequate accounts of both the first-person pronoun and the first-person concept, therefore, will have to operate at two levels. One familiar way of doing this for the first-person pronoun is to define the linguistic meaning (or character, as it is sometimes called) of the first-person pronoun as a function from contexts to contents, where contents are functions from possible worlds to the individual of the context (Kaplan 1989, 505–507). One would expect a structurally similar account to hold for the first-person concept, so that the first-person concept is that concept whose tokening generates thoughts with a highly specific and irreducible functional role for the subject who has them. In the case of both pronoun and concept, the account is pitched at the level of the type rather than the token. The meaning of “I” is located at the level of the function from contexts to individuals, rather than at the level of the individuals picked out by particular tokenings of “I.” So too is the nature of the first-person concept fixed in a way that is independent of its tokening in particular individuals.

In light of this, the difference between the representation of the self in perception and in thought becomes clear. Frege remarked in “The Thought” that “everyone is presented to himself in a special and primitive way in which he is presented to no one else” (Frege 1918, 12). Appearances suggest that Frege was intending this peculiar perspective to be reflected at the level of thought, and his comments have usually been interpreted as applying to the sense of the first-person pronoun. What I would like to suggest, however, is that Frege’s pithy phrase applies quite accurately to the way in which the self is presented in perception, but has no application at all to the way in which the self is represented in the content of belief and other propositional attitudes. There is nothing perspectival about the first-person concept.³

This basic thought can be motivated further by considering the range of representation-dependent thoughts available in other sensory modalities. I can come to the belief that I am near the church because I hear the ringing of the nearby church-bells. My auditory perception has a first-person content, namely, that I am not far from the source of the sound. This is part of the content of perception. It is not something that I infer from what I hear. So, the self must be represented in auditory perception—and yet in a fundamentally different manner from how it is represented in beliefs based on auditory perception. The same holds with even more force for self-conscious beliefs derived from somatic proprioception. Somatic proprioception is a source of first-person contents. Nonetheless, the self does not feature in the content of somatic proprioception in anything like the way it does in the contents of beliefs based on somatic proprioception.

V

Representation-dependent judgments are judgments of the form “I am F” that are made by taking a certain justifying perceptual content C at face value. There is nothing wrong with the idea of a content being taken at face value. It is quite clear that when I form the belief that I am in front of the cathedral on the basis of a perception of the cathedral in front of me I am taking my experience at face value. I am not, for example, making any sort of inference from my perception to the presence of the cathedral. The perception provides an immediate reason for the belief. But the question remains of how best to understand the content that is taken at face value. We have rejected the identity of content thesis, according to which there is no difference in content between the justifying perceptual state and the self-ascriptive belief that is formed on its basis. If the self is explicitly represented in perception then it cannot be represented there in the same way as it is in belief. Since we have also rejected the suggestion that there is no representation of the self in the content of the justifying perceptual state, we are forced to accept that representation-dependent judgments are grounded in a form of representation of the self that is essentially nonconceptual. What this means, of course, is that an account is urgently needed of the grounds of representation-dependent first-person judgments. We need an account of the first-person dimension of the target perceptual states.⁴

Let me approach this at a tangent. There is a general issue here about the extent to which perception is fully conceptual. We are familiar with the idea that perception is concept laden—that the content of perception falls, as McDowell often puts it, within the space of reasons (McDowell 1994). We do not see objects in a neutral way, attending simply to their geometric, sensory, and kinematic properties and then, on the basis of that neutral perception, work out how they fit into our conceptual categorization of the world. The concepts are already built into the way in which we perceive the world, although nobody has ever managed to make clear how this works. Nonetheless, to say that perception is concept laden is not to deny that it has a nonconceptual dimension. In a sense it is completely obvious that perception must have a nonconceptual component because otherwise there would be no qualitative difference between the content of perception and the content of propositional attitudes. The critical question is not whether perception has a nonconceptual content, but whether it has a nonconceptual representational content—and several authors have argued persuasively that the nonconceptual content of experience is partially (and indeed perhaps completely) representational. These authors, however, have focused primarily on certain limited aspects of the content of perception (see, e.g., Peacocke 1992, chap. 3). They have stressed, for example, how the content of perception is analog rather than digital; how it presents distances in a unit-free manner; and how it supports a fineness of grain that is arguably unavailable to conceptual thought.⁵

However, these three features of the content of perception do not by any means exhaust its nonconceptual representational content. The content of perception includes, I shall suggest, an ontogenetically primitive form of representation of the self. Full-fledged self-consciousness emerges, as I have stressed elsewhere (Bermúdez 1998 and chapters 1 and 2 in this volume), from a foundation of more primitive types of self-awareness—in the propriospecific dimension of exteroceptive perception, in somatic proprioception, in the ability to locate oneself spatially within the world, and in the self-representation implicated in primitive forms of social interaction. These primitive forms of self-consciousness are prior to full-fledged self-consciousness in the order of development—and indeed are necessary conditions for the emergence of full-fledged self-consciousness. This genetic story by no means entails, of course, that once full-fledged self-consciousness has been attained, these primitive forms of self-consciousness have any further role to play in the cognitive economy. It could simply be that they just drop out of the picture, rather like some of the basic perceptual expectations that very young infants have about the behavior of distal objects.

Yet, there is a fundamental disanalogy between the perception of distal objects and the perception of the self in this respect. It is plausible to think that young infants parse the perceived array into object-like segments that obey certain basic geometrical and kinematic principles. These have been studied at some length by developmental psychologists. With the acquisition of concepts and language, however, this way of perceiving the world is overridden by the conceptual seeing-as that I have already briefly mentioned. Little if anything remains of the infant perceptual universe in adult perception. But the perception of the self is not at all like that. The fundamental differences in content between first-person representational states and first-person beliefs based on those states point us toward the persistence of a fundamentally nonconceptual representation of the self in ordinary perception. This provides the key to the way in which perceptual states provide grounds for the representation-dependent first-person judgments that are made by taking them at face value (although not, as we have seen, by duplicating them).

VI

The best account of the nonconceptual representation of the self in visual perception comes, I think, from J. J. Gibson’s ecological approach to perception, and I will simply briefly sketch some of the principal respects that he has drawn to our attention in which the self is represented in visual perception (Gibson 1966, 1979).

Gibson stresses certain peculiarities of the phenomenology of the field of vision. Notable among these is the fact that the field of vision is bounded. Vision reveals only a portion of the world to the perceiver at any given time (roughly half in the human case, owing to the frontal position of the eyes). The boundedness of the field of vision is part of what is seen, and the field of vision is bounded in a way quite unlike the way in which spaces are bounded within the field of vision. The self appears in perception as the boundary of the visual field—a moveable boundary that is responsive to the will.

In addition to the boundedness of the visual, the self becomes manifest in visual perception (according to Gibson) because the field of vision contains various body parts that hide, or occlude, the environment. The nose, for example, is distinctively present in just about every visual experience. The cheekbones, and perhaps the eyebrows, also occupy a (slightly less) dominant position in the field of vision. And so too, to a still lesser extent, do the bodily extremities, hands, arms, feet, and legs. These protrude into the field of vision from below in a way that occludes the environment, and yet that differs from how one nonbodily physical object in the field of vision might occlude another. All objects, bodily and nonbodily, can present a range of solid angles in the field of vision (where a solid angle is an angle with its apex at the eye and its base at some perceived object), and the size of those angles varies according to how far the object is from the point of observation. The further away it is, the smaller the angle will be. But the solid angles subtended by occluding body parts cannot be reduced below a certain minimum. Perceived body parts are, according to Gibson, “subjective objects” in the content of visual perception.

These self-specifying structural invariants provide only a fraction of the self-specifying information available in visual perception. There are two more important types of self-specifying information.

The mass of constantly changing visual information generated by the subject’s motion poses an immense challenge to the perceptual systems. How can the visual experiences generated by motion be decoded so that subjects perceive that they are moving through the world? Gibson’s notion of visual kinesthesis is his answer to this traditional problem. Whereas many theorists have assumed that motion perception can only be explained by the hypothesis of mechanisms that parse cues in the neutral sensations into information about movement and information about static objects, the crucial idea behind visual kinesthesis is that the patterns of flow in the optic array and the relations between variant and invariant features make available information about the movement of the perceiver, as well as about the environment.

As an example of such a visually kinesthetic invariant, consider that the optical flow in any field of vision starts from a center, which is itself stationary. This stationary center specifies the point that is being approached, when the perceiver is moving. The aiming point of locomotion is at the vanishing point of optical flow. Striking experiments have brought out the significance of visual kinesthesis. In the so-called moving-room experiments, subjects are placed on the solid floors of rooms whose walls and ceilings can be made to glide over a solid and immoveable floor. When experimental subjects are prevented from seeing their feet and the floor is hidden, moving the walls backward and forward on the sagittal plane creates in the subjects the illusory impression that they are moving back and forth. This provides strong support for the thesis that the movement of the perceiver can be detected purely visually, since visual specification of movement seems to be all that is available. When young children are placed in the moving room, they actually sway and lose their balance.

A further important form of self-specifying information is available to be picked up in the field of vision, according to the theory of ecological optics. This is the result of the direct perception of a class of higher-order invariants that Gibson terms affordances. It is in the theory of affordances that we find the most sustained development of the ecological view that the fundamentals of perceptual experience are dictated by the organism’s need to navigate and act in its environment, because the animal and the organism are complementary. According to Gibson, the possibilities the environment affords are not learned through experience. Nor are they inferred. They are directly perceived as higher-order invariants. And of course, the perception of affordances is a form of self-perception—or, at least, a way in which self-specifying information is perceived. The whole notion of an affordance is that of environmental information about one’s own possibilities for action and reaction.

VII

Gibson’s characterization of the phenomenology of visual perception offers important insights into how the self is represented nonconceptually in the content of perception. To return to our earlier discussion of representation-dependent judgments, it is the “ecological self” that features in the content of basic perceptual states and that provides the epistemic grounds for self-conscious judgments made on the basis of those perceptual states. More generally, representation-dependent judgments are grounded in the primitive, nonconceptual representations of the self discussed in more detail in the other chapters in this volume.

This brings us to the second of the two questions identified at the beginning of section II. To understand the epistemology of full-fledged self-conscious judgments it is not enough to understand the grounds on which those judgments are based. We need also to understand why, when those judgments count as knowledge, they should do so.

It will be helpful to begin by turning briefly to how representation-independent judgments are grounded. These judgments are, in essence, self-ascriptions of mental states, and it follows from the definition of representation-independence that the self-ascriptive component of the representation-independent judgment is not grounded in the content of the state that the subject self-ascribes. A representation-independent judgment is made when a subject makes the transition from φ-ing that p to the self-ascriptive judgment “I am φ-ing that p,” but there is nothing in the content p that corresponds to the first-person concept that appears in the self-ascriptive judgment (in obvious contrast to representation-dependent judgments). So how does the subject make the transition? The simple answer is that the transition is available to subjects who have a basic mastery of what might be termed a simple theory of introspection (by analogy with the simple theory of perception and action discussed in Evans 1982 and Campbell 1994). In this case the simple theory of introspection amounts to nothing more than some level of mastery of the a priori link between being aware of a thought and it being the case that one is thinking it—which is, of course, the same a priori link that governs the status as knowledge of the resulting representation-independent judgment (according to Peacocke’s delta account).⁶ What we see in this case is that the subject’s reason for forming the judgment is an appreciation of precisely the factor that qualifies the judgment as knowledge. Let me describe epistemic situations of this type as being canonically internalist.

To see how representation-dependent judgments can be just as canonically internalist as representation-independent judgments, we need to explore the relation between representation dependence, representation independence, and the phenomenon that Sydney Shoemaker has described as immunity to error relative to the first-person pronoun. As Shoemaker puts it (with reference to statements rather than judgments):

To say that a statement “a is φ” is subject to error through misidentification relative to the term “a” means that the following is possible: the speaker knows some particular thing to be φ, but makes the mistake of asserting “a is φ” because, and only because, he mistakenly thinks that the thing he knows to be φ is what “a” refers to. (Shoemaker 1984, 7–8)

Immunity to error through misidentification relative to the first-person pronoun (henceforth abbreviated to: immunity to error through misidentification) is a property of judgments that are not based on an act of identification of a particular object as oneself. They are judgments that one has a certain property based on grounds such that knowing that something has that property is ipso facto to know that one has that property oneself. It is clear that representation-independent self-ascriptions are immune to error in this sense.⁷ If I form the judgment that I am thinking about Pythagoras’ theorem because I am aware of thinking about Pythagoras’ theorem then that judgment is not based on an identification of myself as the author of the relevant thought. Rather, knowing that there is a thought about Pythagoras’ theorem just is knowing that that thought is a thought that I am having. It is equally clear that representation-dependent thoughts very often share this feature of being immune to error through misidentification. If I form the judgment that I am in front of the cathedral because I see the cathedral in front of me, this is not because I form the belief that someone is in front of the cathedral and then identify that person as myself. Knowing on the basis of perception in this way that someone is in front of the cathedral just is knowing that I myself am in front of the cathedral.

The immunity to error through misidentification of first-person judgments is intimately linked with their status as knowledge, in virtue of the connection between immunity to error through misidentification and the defining property of the first-person concept, namely, that of picking out the thinker of any thought in which it features (in the appropriate way). In a judgment of the form “I am F” that is immune to error through misidentification the only possibility of error is with respect to the predicative component. And hence, provided that the information grounding the predicative component has been acquired in an appropriate manner, one would expect the judgment to count as knowledge. It is natural to conclude, therefore, that an explanation of the immunity to error of first-person judgments will provide the key to explaining their status as knowledge.

Nonetheless, representation-independent and representation-dependent judgments secure their immunity to error through misidentification in fundamentally different ways. In the case of representation-independent judgments the very feature that secures their status as knowledge is what underwrites their immunity to error through misidentification—namely, the a priori link between being aware of a thought and its being the case that one is thinking it. In the case of representation-dependent judgments, on the other hand, immunity to error through misidentification has to be a function of the states on which they are based. Let me offer the following conjecture. Representation-dependent judgments that are immune to error through misidentification are so because they involve taking at face value first-person contents that are themselves immune to error through misidentification. It is easy to see how this works in the case of judgments based on perceptual states, such as the example we have been considering of my forming the belief that I am in front of the cathedral in virtue of seeing the cathedral in front of me. When I see the cathedral in front of me I do not see that the cathedral is in front of someone and then work out that I am that person. Seeing that someone is in front of the cathedral just is seeing that I am in front of the cathedral. And, given this, it is hardly surprising that the resulting judgment is itself immune to error through misidentification. It is plausible to conjecture that a broadly similar account is true of other types of representation-dependent judgments (such as those based on the deliverances of somatic proprioception, for example).⁸

In a sense, however, this merely pushes the problem one step further back. Representation-dependent judgments are immune to error because they are grounded in first-person states that are themselves immune to error through misidentification. But in virtue of what are those first-person states immune to error through misidentification? Immunity to error through misidentification is a formal property of representational states. We need to know what is responsible for it. The answer, in brief (and at least as far as visual perception is concerned), lies in the simple fact that the information about the self provided within the content of visual perception is such that it is manifest to the perceiver that it could not be about any creature except the self. To return to the earlier example, the information that I have about my own spatial position relative to the cathedral could not be information about anyone else’s spatial position relative to the cathedral. This is because the body is manifest as the originating point of perception in such a way as to ensure that the self-conscious judgment that I am perceiving the cathedral is immune to error through misidentification.

To get a better understanding of why this is the case we need to revert to the insights into the phenomenology of visual perception gleaned from Gibson. The self is presented in perception as a distinctive and peculiar form of object in virtue of certain features of the content of visual perception. Some of these derive from the fact that the world is presented to the perceiver largely in terms of the affordances it offers. Others arise from the features that led Gibson to describe perceived body parts as subjective objects in the field of view—the fact, for example, that body parts can vary in perceived size only within a limited range. The way in which the body appears as the frame and boundary of the visual array is also extremely important. The embodied self is the originating point of perception in a much richer sense than the camera lens is the originating point of a photograph. The way in which the embodied self appears in perception, therefore, reflects the immunity to error of the relevant perceptual content and, by extension, that of the resulting belief content.

None of these features of the content of visual perception are transmitted to the content of self-conscious judgments made on the basis of visual perception, but nonetheless they underwrite the status as knowledge of those self-conscious judgments. They do this because they make it the case that the perceiver’s reason for forming the judgment includes an appreciation of precisely those factors that qualifies the judgment as knowledge. The immunity to error of representation-dependent judgments is a function of the fact that the bodily self is the origin of perception. The distinctive way in which the bodily self is presented in visual perception ensures that this is manifest to the perceiver—it is part of what is perceived. This ensures that, to return to the terminology introduced earlier, epistemic situations leading to the formation of representation-dependent judgments are canonically internalist.

Notes

4. This problem arises not only for self-conscious judgments that are representation dependent. It arises also for an important class of representation-independent self-ascriptions. Peacocke’s list of representation-independent first-person judgments include autobiographical memories (my self-ascribing the memory of attending the birthday party) and self-ascriptions of perceptions (my self-ascribing the perception of the phone on the table). The second of these is quite clearly derived from a first-person content. What makes it true that I see that the phone is on the table is my occurrent perception of the phone on the table—and, as we’ve seen at some length with the example of my seeing the cathedral, my occurrent seeing that the phone is on the table is a first-person content. The same holds for the third example—my autobiographical memory of attending the birthday party (assuming, of course, that this memory is episodic). My episodic memory of attending the birthday is a memory of my having attended the birthday party. In both these cases of representation-independent judgment we need to distinguish, just as we did in the representation-dependent case, between the way in which the self is represented in the original perception or memory and the way in which it is represented in the doxastic states formed on the basis of that target state.

3 The Sources of Self-Consciousness

I

II

III

IV

V

VI

VII

VIII

Notes

References