Consciousness, Self-Consciousness and Essential Indexicality


Dan Zahavi argues for a Sartrean conception of consciousness which he regards as endorsed by numerous other thinkers in the phenomenological tradition and as superior to both contemporary “higher-order” and “one-level” (neo-)Brentanian conceptions. As Zahavi points out, all three conceptions—the ostensibly phenomenological, the “higher-order” and the Brentanian—assume that consciousness is to be explicated as a form of self-awareness. Only under this assumption, thinks Zahavi, can one endorse the thesis deriving from Nagel that a psychological state or experience is conscious only if there is something it is like to be in this state or experience. For according to Zahavi there being something it is like to be in a certain psychological state or experience implicates the givenness of something to the subject, hence what Zahavi calls the dative status of the subject. I argue that one can accommodate, at least for conscious perceptual experience, the intuition which leads Zahavi to speak of givenness and the dative status of the subject without endorsing the thesis that consciousness is a form of self-consciousness (which I regard as implausible and for which Zahavi has no good argument). I do this by extracting what I take to be the real lessons implicit in Perry’s deliberations about essential indexicality: essential anaphora and essential spatial demonstrativeness. These two notions, taken as capturing the structure of distinctively perceptual intentional content, permit an account of conscious perceptual experience which allows one to construe what it is like to undergo such experience in a way which does not entail any kind of self-awareness.

This is a slightly altered and extended version of a paper held on February 17th., 2015, in the Research Seminar of the School of Philosophy at the ANU.

Section 1: Zahavi and Sartre on Consciousness

Over the last two decades Dan Zahavi has defended a conception of consciousness which he claims to find, at least in general terms, in all thinkers in the phenomenological tradition, from Husserl, Scheler and Heidegger through Sartre and Merleau-Ponty to Henry and Ricoeur.1 He distinguishes this conception from two competitors, namely, contemporary higher-order conceptions on the one hand, and Brentano’s and neo-Brentanian conceptions on the other. 2

The most well-known advocate of this allegedly phenomenological conception of consciousness is Sartre.3 According to Sartre,

consciousness is essentially characterized by intentionality. It is as such a consciousness of something. He also claims, however, that each and every intentional experience is characterized by self-awareness. Thus, Sartre takes self-awareness to constitute a necessary condition for being conscious of something. (Sartre 1943, pp.18, 20 and 28; Sartre 1948, p.62; quoted in Zahavi 2004, p.82).

Importantly, the two other conceptions of consciousness also maintain this. Both higher-order and Brentanian and neo-Brentanian positions also assume that, to put the matter crudely, consciousness entails self-consciousness.4 In other words, for all three conceptions the consciousness of a conscious intentional state or experience consists in its subject’s being in some way aware of this intentional state or experience. The difference lies in how this self-awareness is conceived. According to higher-order theories, it is a distinct higher-order intentional state or experience which is not intrinsic to the conscious state or experience of which it is an awareness. By contrast, Brentano and the neo-Brentanians deny that it is a distinct intentional state or experience. Rather, this self-awareness is an intrinsic structural feature of the conscious state or experience itself. It is, however, genuinely intentional; it is awareness of this conscious state or experience. Against this, Sartre and Zahavi maintain that precisely because it is indeed an intrinsic structural feature, the self-awarenesss in which consciousness consists cannot be genuinely or strictly intentional. Pace Brentano and such neo-Brentanians as Kriegel, it is not an awareness of anything.

But why do Sartre and with him Zahavi deny that this intrinsic self-awareness is a form of intentionality in the strict sense, i.e., a form of what Zahavi calls object-consciousness?5 Because they regard the self-awareness in which consciousness consists as a form of intentionality in the strict sense, Brentano and the neo-Brentanians de facto appeal to notions of representation and reference which entail the distinctness and extrinsicality of higher-order theories. In short, they formulate their version of the one-level account in a fashion which persistently insinuates the higher-order account they seek to reject. Sartre and Zahavi agree with Brentano and the neo-Brentanians that

self-awareness (or inner consciousness) differs from ordinary object-consciousness. The issue of controversy is over whether self-awareness is merely an extraordinary object-consciousness or not an object-consciousness at all.(Zahavi 2004, p.85)

Sartre and Zahavi think one must endorse the latter option because otherwise one’s account of consciousness becomes incoherent and unstable.

But why do all three camps maintain that the consciousness “should be accounted for in terms of self-awareness, i.e., a conscious mental state differs from a non-conscious mental state by entailing self-awareness”? (Zahavi 2004, p.86) In his essay on Brentano Zahavi just asserts that this is so. Any conception of what makes a conscious state conscious must, he says, explain how such a state is

given as my state, as a state that I am in. Why? Because this first-personal givenness is an ineliminable part of what it means for a state to be conscious—it concerns the fact that a conscious mental state feels like something for somebody —and for a theory of consciousness to leave this aspect out is to leave something absolutely crucial out. (Zahavi 2004, p.68)

This is just mere assertion. In another essay (Zahavi 2009), however, Zahavi attempts to argue. He first points out that the consciousness of conscious experience has often been explicated in terms of there being something it is like to undergo the experience. Zahavi thus agrees with Nagel’s claim that there being something it is like to have or to undergo a psychological state or experience constitutes the conscious character of the state or experience—see Nagel 1974. But, he goes on to say, there is more to the subjectivity, i.e., conscious character, of a conscious psychological state or experience

than the fact that what it is like to perceive a green square is subjectively distinct from what it is like to perceive a blue circle. … What it is like to perceive a green square is different from what it is like to remember or imagine a green square. Moreover, … in perceiving or imagining an object consciously, the object appears in a determinate manner to ourselves. Whereas the object might be termed the accusative of the perceiving, the subject is the dative. (Zahavi 2009, pp.554-555; underlining added)

The ‘argument’ thus appears to be as follows: what constitutes a psychological state or experience as conscious is there being something it is like to be in this state or to undergo this experience. In other words, consciousness is what-it’s-likeness. But the what-it’s-likeness of a conscious psychological state or experience does not exhaust itself in the phenomenal characteristics the state or experience has in virtue of its content.6 It also includes the phenomenal aspects of what Husserl would call its Akt-Charakter, Searle its representational mode. There is thus an element of what-it’s-likeness which consists in direct, phenomenal awareness of the conscious psychological state or experience itself. Furthermore, since awareness of a psychological state or experience implicates awareness of the subject whose state or experience it is, the conscious character of a psychological state or experience is awareness not just of the state or experience itself but also of the subject of this state or experience. Finally, the awareness of Akt-Charakter or representational mode at issue here is phenomenal, so to speak awareness from the inside. Consequently, it is both awareness of the state or experience as mine and awareness of the subject as me. Therefore consciousness, because it consists in there being something it is like to be in the relevant psychological state or to undergo the relevant psychological experience, is a form of self-awareness.7

This is clearly a bad argument. Let us grant that the consciousness of a conscious psychological state or experience consists in there being something it is like to be in this state or to undergo this experience. Let us also grant that the what-it’s-likeness of some conscious psychological states or experiences does not exhaust itself in phenomenal characteristics possessed by the state or experience in virtue of its content but also involves phenomenal characteristics possessed in virtue of act-character or representational mode. Even so, from neither of these two things does it in any obvious way follow that the subject of a conscious state or experience is aware of either the latter or of itself. The ‘argument’ begs the question in claiming, on the basis of the two preceding claims, in particular, the second, that what-it’s-like-ness consists in part in direct, phenomenal awareness of the conscious psychological state or experience itself.

The fact that Zahavi has no better argument than this for the thesis that consciousness just is a form of self-consciousness intimates that he is trying to unite what cannot be united. We should therefore expect the position of Zahavi and Sartre simply to mirror the incoherence of the position of Brentano and the neo-Brentanians. This is indeed the case: whereas Brentano and the neo-Brentanians claim that the self-awareness in which consciousness consists is something representational and referential which is not like anything paradigmatically representational and referential, Zahavi and Sartre claim the inverse: this self-awareness is not something representational and referential which is like something paradigmatically representational and referential.

Nor do either higher-order conceptions or Brentano and the neo-Brentanians fare any better in arguing for the thesis that consciousness is a form of self-consciousness. This deficiency in all three approaches forces us back to the drawing board. We need an account of the consciousness of conscious psychological states and experiences which does not entail that consciousness is a form of self-consciousness. At the same time, this account must explain why so many have thought that this thesis is true. Lastly, we need an account which explains a curious feature of Zahavi’s presentation: its orientation towards conscious perceptual experience. In particular, the notion of givenness to which Zahavi appeals involves more than mere what-it’s-like-ness. For strictly speaking, only perceptual experience involves the givenness of something to something. This suggests the following hypothesis: in some way conscious perceptual experience is a foundational form of consciousness—foundational for what Sartre calls reflective self-consciousness, i.e., explicitly first-personal thinking.8 And from this one can derive a further condition of adequacy by satisfying which one can confirm the hypothesis: an account of consciousness in general must show conscious perceptual experience to be foundational for reflective self-consciousness. In what follows, I will attempt to develop an account of the consciousness of conscious perceptual experience which meets these three conditions of adequacy.

Section 2: Back to the Drawing Board

There is something right in Zahavi’s remark that in the examples of conscious intentionality he adduces—conscious remembering of the quasi-perceptual recollecting kind, imagining and in particular perceiving itself—the subject occurs as a dative to which the object is given.9 At the same time, the remark needs considerable reconstruction if we are to understand it correctly, in particular, in a way which does not entail that consciousness is a form of self-awareness. To this end, I want to return an old issue which one might at first think to be only distantly related to the issue of consciousness.

Section 2.1: Revisiting the Supermarket

Consider the following well-known example:

I once followed a trail of sugar on a supermarket floor, pushing my cart down the aisle on one side of a tall counter and back the aisle on the other, seeking the shopper with the torn sack to tell him he was making a mess. With each trip around the counter, the trail became thicker. But I seemed unable to catch up. Finally it dawned on me. I was the shopper I was trying to catch.

I believed at the outset that the shopper with a torn sack was making a mess. And I was right. But I didn’t believe that I was making a mess. That seems to be something I came to believe [once, namely, I realised that I was the shopper I was trying to catch]. And when I came to believe that [namely, that I was making a mess], I stopped following the trail around the counter, and rearranged the torn sack in my cart. My change in beliefs seems to explain my change in behavior. (Perry 1979, italics and bracketed additions added, p.27)

Perry concludes from this example that in order accurately to characterise the change in belief, in particular, the belief which is causally responsible for his ceasing to follow the trail and his rearranging the torn sack in his shopping trolley, he must use the first person pronoun “I”. He writes, “My beliefs changed … in that I came to have a new one, namely, that I am making a mess … .” (Perry 1979, p.27) The occurrence of the word “I” in Perry’s expression of what he came to believe is, he claims, essential. For “(w)hen we replace it with other designations of me, we no longer have an explanation of my behavior and so, it seems, no longer an attribution of the same belief.” (Perry 1979, p.27)

But how correct is this? It seems very clear, and above all it is certainly possible, that Perry goes through the following process of thought leading to action10: he explicitly and self-consciously recognises that he is identical with the shopper he was trying to catch. That is, he explicitly and consciously judges, “I am (identical with) the shopper I am trying to catch.” This engenders the belief causally responsible for his subsequent behaviour, the belief, namely, that he is making a mess. Thereupon he acts—without ever explicitly thinking, “I am making a mess.”11 It seems that in this process only one phase is strictly or genuinely first-personal, such that one must use the first person in characterising it: what Perry describes as its dawning upon him that he is the shopper he is trying to catch. This does seem very plausibly to be a case of explicit, first-personal thinking, for which reason, one must, even from the third person perspective, use the first person if one wishes to capture what exactly happened. One must say, “John Perry thought to himself, “I am the shopper I am trying to catch.”

Now in his own characterisation of the case Perry does not explicitly say that he consciously thinks, “I am making a mess.” He speaks of the importance of the word “I” in the expression of what he came to believe. But this just makes the obscurity and imprecision of the example all the more evident: the word “I” can only be essential to Perry’s expression of what he came to believe if it is essential to capturing the structure of what he came to believe. But we have just seen that it is possible, even likely, that he did not explicitly or consciously judge, “I am making a mess.” So use of the word “I” is not essential to his expression of the intentional state which causes him to rearrange the torn bag of sugar in his trolley. For this reason, it would be perfectly acceptable as an explanation of Perry’s behaviour for anyone, Perry included, to say, “John Perry believes that he is making a mess, this is why he is rearranging the torn bag of sugar.”12 In fact, the details of the case suggest that if the use of the word “I” is essential at any point in the linguistic characterisation of the process of thought Perry goes through, then only at that point at which one comes to describe Perry’s judging that he is the shopper he is trying to catch. For only this judging appears to be explicitly first personal. Consequently, expressing this exactly requires attributing to Perry the first personal mental judgement, “I am the shopper I am trying catch.”

This points to two questions which Perry’s example and his general presentation of the issue tend to elide: Firstly, in characterising, for purposes of explanation, Perry’s various intentional states and experiences can we capture the causal efficacy of these states and experiences if we use merely singular terms in the standard sense—names, definite descriptions and even indexicals or demonstratives13to characterise the references these states and experiences make to Perry? Secondly, can we capture the causal efficacy of these states and experiences if we do not characterise them as a matter of Perry’s thinking in explicitly first-personal terms? In effect, there are two problems here, the problem of whether the kind of anaphoric construction exemplified by “Perry believes that he is making a mess” is essential; and the problem of whether the first person is essential, as it appears to be when it dawns on Perry that he is the shopper he is trying to catch.14

Evidently, to insist that there are two questions here is to insinuate that a subject can believe that he, she or it is Φ without either explicitly thinking, “I am Φ,” or even believing this sentence to be true in the sense of being disposed to assert it. In other words, it insinuates that so-called attitudes de se differ not only from attitudes de dicto; they can also differ from attitudes de me—differ objectively, that is, in a fashion which is not simply one of reporting perspective. Can, then, a being can have solely anaphorically characterisable intentional states and experiences without these states or experiences being explicitly self-conscious, first-personal thoughts or their subject’s being capable of such thoughts at all? In other words, can there be a kind of intentionality whose content is only anaphorically expressible—what one might call, speaking loosely, anaphoric intentionality?15

Section 2.2: The Concept of Anaphoric Intentionality

Consider another example:

S sees a large dog aggressively charging towards it and runs away.

Evidently, this could be the explanation of the behaviour not of a linguistically competent, self-conscious human being but of another dog. I take it that if S were a dog, it would be problematic to attribute to S the explicitly first-personal judgement, “A large dog is aggressively charging towards me.” Even so, the anaphorically used pronoun “it” is essential in the sense that one cannot replace it by any co-referential singular term and still capture the causal connection between S’s seeing and its behaviour. In order to see this, let us replace the pronoun by another singular term ‘Sʹ’ which is co-rereferential with ‘S’. Clearly, the sentence “S sees a large dog aggressively charging towards Sʹ and runs away” only does the same explanatory work for an explainer of S’s behaviour if the explainer knows that Sʹ is identical with S. The pronoun ‘it’ cannot be replaced salva explanatione by any co-referential term other than S.

Nor can it be replaced by the singular term ‘S’ itself. Use of a singular term in the characterisation of the content of the intentional experience ascribed to S is not justified simply because S’s being the referent of this term, or bearing attributes in some way identified by its being the referent of this term, is causally relevant. Let us assume that S is John Perry’s dog, which is notoriously fearful, hence inclined to run away from any dog, no matter how large or small.16 This could be relevant to the determining of S’s behaviour. Even so, this does not justify using the singular term ‘John Perry’s dog’ in a characterisation of the content of the intentional experience ascribed to S. Rather, any such merely objectively causally relevant fact must be appended to the explanation as a whole, as an addition to it. In order for any fact about S to be justifiably mentioned in the characterisation of the content of S’s intentional experience, S’s believing this fact must be causally relevant. So the anaphorically used pronoun can only be replaced by the singular term ‘S’ itself if S’s believing that it is S is causally relevant, which in this case it is not.

With this, we see an important general point: the real issue driving the kind of example we find in Perry, Castañeda and others has nothing to do with self-consciousness in the sense of explicit, first-personal thinking—with attitudes de me. Sometimes, of course, the agent in such examples will be explicitly and self-consciously judging it itself to be thus and so and in this case the third person pronoun anaphorically used in any explanation of the agent’s behaviour will track explicit, first-personal thinking.17 Even so, this kind of example only reveals, not essential indexicality, but essential anaphora: there is a kind of belief or mental judgement, indeed of intentional state or experience generally, which can only be reported by the use of pronoun occurring within the intensional context, a pronoun which derives its reference anaphorically from that of the subject-expression in the report—and yet the state or experience reported is not first-personal, is not a so-called ‘I’-thought.

Relatedly, there is a use of the pronouns ‘he’, ‘she’ and ‘it’ which stands so to speak in between standard indexical and extensionally anaphoric uses and the use which Castañeda marks with his special symbol ‘he’; this use in between occurs in reports for which Castañeda’s ‘he’ is not appropriate—not appropriate because what is reported is not an ‘I’-thought. One thus cannot say, as Castañeda does, that when a third person pronoun occurs in a psychologically intensional context and refers “through referring back to the expression” which occupies subject position of the psychological verb, it is “used to attribute first-person references to” this expression’s referent. (Castañeda 1968, p.90) Sometimes this is the case but it is not always so.

Section 2.3: Trapped by the Doctrine of Propositions: Perry on Agent-Relative Knowledge

Interestingly, Perry himself later recognises, if only in round-about fashion, that essentially anaphoric reports do not necessarily attribute first person thought. But what they do report when they do not report first person thought? Perry gives the example of his seeing an apple directly there in front of him which he reaches out, picks up and eats—see Perry 1998, p.329. Evidently, if one were observing Perry and sought to characterise, for purposes of behavioural explanation, the perceptual experience which leads Perry to pick the apple up and eat it, one would use precisely something like the kind of anaphoric construction Perry himself uses in giving the example. One might say, for example, “John Perry sees an apple there in front of him.” But what kind of thought is one attributing in so doing? Perry says that at least typically it will be what he calls agent-relative knowledge. And agent-relative knowledge is, he says, a matter of selfless thought, that is, knowledge or belief concerning ourselves which does not involve representation of ourselves. (Perry 1998, pp.326-329) “This kind of knowledge is,” he claims, “self-knowledge, in that it embodies knowledge of the relations things stand in to the agent … . But it does not require that the agent have an idea of self or a notion of itself.” (Perry 1998, p.329) For this reason, it is not explicitly self-conscious or first-personal.

Perry acknowledges that this might seem strange. Surely he needs know such things as “from whom the apple is a certain distance and direction.” (Perry 1998, p.327) After all, if the apple were this distance or direction from someone else, then he would not have behaved or indeed perceived as he did. Perry’s seeing must either be or convey knowledge about himself—in which case it must surely involve some notion, some representation, of himself.

But Perry rejects this as misleading. “A natural way,” he says,

for me to report what I saw would be to simply say:

That’s an apple


There is an apple there

There is nothing in this remark that refers to me. And after all, why should there be? I didn’t see myself, I saw an apple. (Perry 1998, p.327)

These remarks are true enough. They are, however, completely irrelevant. We wish to understand how perceptual experience and belief can inform its subject of certain crucial relations in which things stand to this subject—‘in front of’, ‘behind, ‘to the right’, ‘far away’, etc. In particular, we wish to understand how, as Perry claims, perceptual experience and belief can do this without representing or even referring to the subject. For this we need an account of what Perry’s seeing and believing are and not what Perry actually gives, namely, an account of how Perry himself would report what he saw or came to believe. Nothing Perry says here entails that it is not perfectly accurate to describe either Perry’s perceptual experience as from the outset a matter of seeing an apple in front of him, or the belief he acquires through it as from the outset a matter of believing there to be an apple in front of him. Both descriptions clearly attribute reference to Perry.

Perry makes another attempt to explain how perceptual experience or belief can inherently constitute or convey information about something without representing or referring to it: imagine a child who can tell time for all local practical purposes but has acquired no understanding of time zones, hence of the relativity of times to a time zone. If the child says, “It is now 7 pm,” it makes a statement which is true if and only if it is 7 pm in whatever time zone the child is in. Nonetheless, there is nothing in the child’s statement or in its thinking generally which articulates or reflects that fact that the child is in a particular time zone since ex hypothesi the child’s life is practically restricted to its immediate location, whatever this happens to be, and so it has no need to represent or refer to time zones. In consequence, there is no representation of or reference to a time zone in the child’s utterances about times of day even though these utterances objectively concern a time zone in the sense that there is essential reference to it in the biconditional specifying the utterance’s truth condition—see Perry 1998, p.328.

This, thinks Perry, provides a model for understanding how, both in the specific case of the apple and more generally, his perceptual experience and belief can be agent-relative: across all the cases of very basic, action-guiding perceptual experience and belief of the kind which the example of the apple illustrates, because the subject of this knowledge remains the same, there need be no representation of or reference to this subject even though the biconditional specifying when its agent-relative perceptual experiences and beliefs are veridical or true does refer to it. The child’s concept of being 7 p.m. has a relativity to current location packed into it which makes the child’s judgement non-representationally and non-referentially sensitive to time. So, too, perceptual experience and the low-level belief it gives rise to have a relativity to the subject packed into them which makes them able to be or convey information to the subject about itself without any representation of, or reference to, this subject.

Another example helps to make clearer where Perry is going with all this:18

A child who is unconcerned about and even unaware of the weather anywhere but where he is, can treat the issue of whether it is raining or not as a property of a time, rather than a relation between times and places. He says, “It is raining now” rather than “It is raining here now.” (Perry 1998, p.328)

This child, says Perry, “is right when he judges “It is raining now,” if it is raining where he is. (Perry 1998, p.328) It seems, then, that the child’s concept of weather and in particular of what it is to rain is to be understood as containing an implicit ‘here’. The child’s judgement can therefore be parsed as “It is raining-here now.” If this is so, then in the child’s concept of weather there is no argument place for any kind of reference to a spatial location and its judgements about the weather contain no such reference. So, too, for our first child and its judgements about times of day. Just as the concept employed by the second in its judgement, “It is raining now,” is to be understood as the single-place concept ‘raining-here’, the first child’s judgements about times of day are to be understood as involving the single-place concept ‘time-of-day-here’. Its judgement about its now being 7 p.m. is thus to be parsed as “It is now 7 p.m.-here.”

With this, we see, at least as best we can, what Perry is getting at in suggesting that when one describes him from the third person perspective as believing that there is an apple in front of him, it does not follow that he is referring to himself. When the child says, “It is 7 p.m. now,” we must describe the child as believing that it is 7 p.m. EST. But it does not follow from this that the child is positively representing or referring to the time zone labelled ‘Eastern Standard Time’; rather, it may just believe that it is 7 p.m. here, a belief which is of course true if and only if it is now 7 p.m. EST. So, too, in the case of Perry’s belief: although we must report this belief by using the third person sentence “Perry believes that there is an apple in front of him,” it may be the case in that Perry believes simply that there is an apple in front, i.e., in-front-of-him.

Yet this account of what Perry is getting at simultaneously highlights a significant problem. The child does not represent or refer to a time zone because in its judgement that it is now 7 p.m. it is not wielding a concept of a time-of-day which involves the concept of a time zone; indeed, the child does not have this latter concept. Something similar must therefore apply to Perry and his belief that there is an apple in front of him: in his perceptual experience and belief he is not wielding the genuinely relational concept of something x’s being in front of something y but the simpler monadic one of something’s x being in front-of-him (or -Perry, or -the-man-who-once-made-a-mess, and so on, for all things true of Perry). So either Perry does not possess the relational concept at all or he has two concepts, one wielded in thinking and talking reflectively about what is front of what, the other in simple perceptual experience and the low-level, action-guiding belief it gives rise to. The first disjunct is clearly false and the second is either also false or at the very least grossly implausible.

So why does Perry puts us through these hoops? The clue lies in a crucial equivocation in the notion of agent-relative knowledge: is agent-relative perceptual experience and belief selfless in the sense that it involves no representation of self? Or is it selfless in the sense that it involve no reference to self? It is by no means obvious that these are the same or even mutually entailing theses yet Perry does not distinguish them. In the essay we have been discussing, Perry writes,

When we perceive how the world is around us and act upon it, we need to judge what distance and direction things stand relative to ourselves. But we do not need to keep track of who it is that we are judging things to be in front of or to the left or, at least as long as we are basing our actions on simple perceptual knowledge. In this case, our knowledge concerns ourselves but need not involve an explicit representation of ourselves. (Perry 1998, p.329; italics added)

But in another essay he concludes from the same observations about perception and action that “(t)here is no need for a self-referring component of our belief … .” (Perry 1986a, p.182; italics added)19 Indeed, this equivocation between self-reference and self-representation is present in the essay we are discussing. For in this essay Perry says that in all the reports he would naturally give of his perceiving an apple there is no reference to him.

So why is there this equivocation in Perry? Or, if the charge of equivocation is too strong, why does he treat self-reference and self-representation as inseparably linked? Presumably because he believes there to be no reference without representation. By this I simply mean that there is no reference to an entity in the content of an intentional state or experience such that this reference cannot be captured—precisely represented—by a free-standing linguistic expression, either a name, a definite description, or even an indexical like ‘I’.20 This immediately commits one to treating as eliminable the kind of anaphoric construction we have been considering, in which a third person pronoun refers from inside an intensional context through anaphoric binding to an appropriate singular term outside of the context. So the reason why one maintains that there is no reference without representation is the conviction that intentional contents must be free-standing items which can, to use Perry’s word, correspond to, hence be individuated in terms of, sentences, even sentences containing indexicals—see Perry 1979, p.39. For only when an intentional content is free-standing in this sense can intentionality be construed as a relation to this content—as a propositional attitude, at least in the case of those intentional contents which are propositional.

Note that to regard the kind of anaphoric construction we have been considering as marking a distinctive and ineliminable form of intentionality is to deny the first and most fundamental of what Perry calls the three main tenets of the doctrine of propositions—see Perry 1979, pp.29-30—, namely, that belief, hence mutatis mutandis intentionality in general, “is a relation between a subject and an object, the latter being denoted, in a canonical … report, by a that-clause.” (Perry 1979, p.29) Perry does not deny this first tenet, not even when, having developed his official position on what the objects of belief and, mutatis mutandis, other intentional states and experiences are, he countenances

(a) more radical proposal [which] would do away with objects of belief entirely. … Rather than saying I believed in the de re proposition consisting of me and the open proposition, x is making a mess, we would say that I stand in the relation, believing to be making a mess, to myself. There are many ways to stand in this relation to myself, that is, a variety of belief states I might be in. And these would be classified by sentences with indexicals. On this view de dicto belief, already demoted from its central place in the philosophy of belief, might be seen as merely an illusion, engendered by the implicit nature of much indexicality.” (Perry 1979, p.41)

Evidently, this is not to do away with the idea of intentionality as a relation of the subject to something but merely with the idea of it as a relation of the subject to something other than the subject.

Section 3: The Essential Unity of Spatial Demonstration and Spatial Orientation in Perceptual Experience

The idea of anaphoric intentionality, of intentionality de se which is not de me, is important not merely for a general philosophical understanding and explication of intentionality. It helps us, I think, to develop an account of the conscious character at least of conscious perceptual experience. For it is the key to seeing what is right in Zahavi’s claim that conscious psychological phenomena possess a givenness which accords a dative status to their subject. In particular, it is the key to understanding how this dative status need not implicate any form of self-awareness.

One cannot, however, see this simply by considering the idea of essential anaphora alone. One needs to link it up with another idea, one which can be developed out of Perry’s deliberations on essential indexicality and which concerns perceptual experience specifically. In his essay Perry argues that not merely the first person pronoun but also the demonstrative adverbs of time and place, ‘now’, ‘here’ and ‘there’ are essential. Here, I believe, with regard specifically to demonstrative adverbs of place, we do find a truly essential indexicality or rather demonstrativeness. This essential demonstrativeness, when taken together with essential anaphora, permits an account of the what-it’s-like-ness of perceptual experience, hence of what it is for conscious perceptual experience to be conscious, which meets the conditions of adequacy listed above.

Consider the following example from Perry:

[A hiker] stands in the wilderness beside Gilmore Lake … . He desires to leave the wilderness. He believes that the best way out from Gilmore Lake is to follow the Mt. Tallac trail … . But he does not move. He is lost. He isn’t sure whether he is standing beside Gilmore Lake, looking at Mt. Tallac, or beside Clyde Lake looking at Jack’s peak … . Then he begins to move along the Mt. Tallac trail. If asked, he would have explained the crucial change in his beliefs this way: “I came to believe that this is the Mt. Tallac trail and that is Gilmore Lake.” (Perry 1979, p.28; abridged by me)

Once again we have a complex example in which a lot of explicit deliberation is going on. Very plausibly, then, the hiker explicitly thinks, “This, i.e., this trail here, is (identical with) the Mt. Tallac trail.”

This explicit demonstrative identity judgement is to be compared with the simpler judgement an experienced hiker who is not lost might make. This latter would not typically spend much time explicitly deliberating. If the experienced hiker made any kind of explicit judgement at all, this would simply be, “Here’s the Mt. Tallac trail.” More likely, however, the experienced hiker would simply see this, without any explicit thought at all.

Now from the third-person perspective one could express the experienced hiker’s perceptual experience by saying, “The experienced hiker sees the Mt. Tallac trail there in front of him.” Note the naturalness of using the demonstrative adverb ‘there’ in this third-person characterisation.21 Does this naturalness indicate that some demonstrative adverb of place, either ‘here’ or ‘there’, is essential?22 Or are such adverbs in principle eliminable? Somehow the location of the trail must be given to the hiker, hence referred to in the content of his experience. It would seem that the only plausible non-demonstrative candidate for replacing the demonstrative would be some such thing as being about 10 metres in front of him: “The experienced hiker sees the Mt. Tallac trail about 10 metres in front of him.” Will this do as explanatorily complete, hence adequate third-person depiction of the hiker’s perceptual experience?

Surely not. Perceiving subjects can certainly see how far things are away from them. But they do so only in seeing them here or there; the former is founded in the latter, as Husserl would say. So the hiker only sees how far away the Mt. Tallac trail is, namely about 10 metres, in seeing it there. The same applies, of course, for his seeing the Mt. Tallac trail in front of him; this, too, he sees only in seeing it there. Both relational propositional functions, ‘… is about 10 metres away from him’ and ‘… is in front of him’, are, in a sense, perceptible as applying to the trail, but only in seeing the trail there. Spatial demonstrativeness is a necessary feature of perceptual intentional content.23

It is, however, not sufficient. The hiker does not see the Mt. Tallac trail as nothing more than there. If he only saw the trail as simply there, he would be disorientated. But ex hypothesi the hiker is not disoriented: upon having his perceptual experience the hiker directly and confidently strides along the Mt. Tallac trail. From the outset, then, his perceptual experience oriented him such that he could act as he did. It not only informed him of the trail’s presence there but also placed the trail in spatial relation to him, thereby orienting him towards it. The hiker therefore did not see it as simply there, he saw it as there in front of him. This and only this explains why, as soon as he sees the Mt. Tallac trail, he heads off directly along it, without pausing and looking around.

There is therefore a difference between the propositional function ‘… is about 10 metres away from him’; and the propositional function ‘… is in front of him’. The latter is a relation of spatial orientation, alongside ‘…is to right of one’, ‘… is above one’, and so on. Such relations of spatial orientation are more fundamental than properties such as being about 10 metres away from one. In fact, they are as necessary to the perceptual experience of well-oriented perceiving and purposively behaving subjects24 as is spatial demonstrativeness. The one complements the other, so both constitute an essential unity.25 Importantly, these relations of spatial orientation essentially involve what, from the third person perspective, one would express by means of an anaphorically used pronoun. So because in normal, action-guiding perceptual experience spatial orientedness complements spatial demonstrativeness, so, too, does anaphorically expressible reference back to self.

There is another way in which the essentiality of spatial demonstrativeness can be confirmed. This way is particularly important because it works by showing precisely the necessary unity of such demonstrativeness with spatial orientation, hence with anaphoric reference back to the subject of experience. Let us consider once again the case of a being clearly incapable of explicitly demonstrative and explicitly first personal thinking.26 S, which we have assumed to be a dog, sees a large dog charging aggressively towards it. Where the dog is attacking from is crucial to determining S’s behaviour. Depending on whether it is charging towards S from on the other side of a river, on the other side of a window pane of the house in which S is safely ensconced or whether where it is charging from gives the dog a clear run on S, S will behave differently. It is thus essential to capturing the causal efficacy and role of S’s perceptual experience that one attribute to it a content which involves this element of where-fromness or whenceness, in which both spatial demonstrativeness and spatial orientation are inseparably combined. The dog is, after all, attacking from out of there; its movement towards S is always already perceived by S as a movement-towards-it-from-there. This whenceness entails the spatial orientation of S towards what it perceives, an orientation which is present in the content of S’s perceptual experience.

This inherent spatial orientation of the subject is present even in the ‘static’ cases of the hiker’s seeing the Mt. Tallac trail there in front of him or of Perry’s seeing an apple there in front of him. For as Perry points out in his clumsy notion of agent-relative knowledge, in seeing the apple there he also learns how the apple stands to him, namely, in front of him. Indeed, given that Perry directly and immediately reaches out and grasps the apple, he sees it as something more, namely, as graspably there in front of him. After all, the apple could be at all sorts of different places and still count simply as in front of him. Given this and the fact that no non-demonstrative specification of place will do, we see once again the essentiality of spatial demonstration of the apple’s location. At the same time, we also see why it is essential: it exists in essential unity with the spatial orientation of the subject towards what is perceived, an orientation one can only capture from the third person perspective via an anaphorically used pronoun. This unity constitutes the oriented character of the perceiving subject as able to act vis-à-vis what it perceives as there—as able to comport itself towards entities, d.h., sich zum Seienden verhalten zu können. Spatial demonstration and spatial orientation, hence essentially anaphorically expressible reference to the subject of experience, come as an essential unity in the structure of normal, action-guiding perceptual experience.

Section 4: The Consciousness of Conscious Perceptual Experience

Let us now turn to explicate the character of conscious perceptual experience as conscious. The strategy will be to use the thesis that perceptual experience is an essential unity of spatial demonstration and anaphorically expressible spatial orientation of the subject in order to develop an account of conscious perceptual experience which satisfies the conditions of adequacy listed above. Specifically, we will attempt to account for character of what-it’s-like-ness as always for a subject without thereby construing it as entailing any kind of self-awareness.

Section 4.1: What-it’s-like-ness as the How of Perceptual Appearance

Imagine that on his hike through the wilderness Perry’s hiker sees a jagged mountain range extending left and right there before him.27 When later asked what the mountain range was like,28 he says that it was sharply jagged, that is to say, acutely rather than obtusely so, that it looked smaller than it actually is due to the distance, that there was pure white snow on the mountain tops which, however, was yellowishly tinged because of the afternoon sun, that the afternoon sun also made the forest below look a darker green than it was and that perhaps the leaves were turning autumn red but he could not tell accurately due to his slight colour blindness. Importantly, he could also just draw a picture which, should he be skilled enough as an artist, would be a much better way of conveying to his interlocutor both what the mountain range was like and what it was like to see it.

Evidently, what the hiker is doing is describing (or depicting) how what he saw appeared to him as what he saw, namely, a jagged mountain range extending left and right there in front of him. So perceptual experience must have a certain structure. In particular, it must be contentful in two interrelated, mutually dependent senses. Firstly, it is intentionally contentful, i.e., contentful in that sense which, to use traditional language, constitutes directedness at an object. But secondly it is contentful in the sense of containing that information about the object perceived, its context and indeed the subject perceiving itself which constitutes how the object perceived appeared as it was perceived, that is, as what it was represented to be by content in the first sense. Evidently, content in this second, so to speak adverbial sense is determined by three factors: (a) the specific way in which the object perceived is objectively what it is—the mountain range is acutely rather than obtusely jagged—; (b) the specific nature of the context—it is late afternoon, the sun hangs low in the sky, etc.—; and (c) the specific nature of the subject perceiving—the hiker is slightly colour blind.

Now the mode or way in which something perceptually appears as thus and so obviously has cognitive value; this is shown by what the hiker is retrospectively able to recollect about the mountain range and his experience of it. But of course the defining functional role of this mode of appearance is practical rather than epistemic. In particular, the mode of appearance is there in order to guide action in ongoing, self-regulating and -revising fashion. When the hiker sees the Mt. Tallac trail there in front of him, the trail can appear to him in numerous ways, e.g., leading off to his left, inclining upwards, partially blocked by a large boulder around which he must walk, crossing a stream, and so forth. Each of these differences in the way in which the trail appears will make rationally appropriate slightly different behaviours. Yet they will do so completely non-cognitively; all the hiker needs to perceive is the Mt. Tallac trail as there in front of him, without necessarily forming any beliefs about how more specifically the trail is, what its context is like and the limitations of his perception of it.

A simpler case illustrates how mode of appearance operates completely unreflectively in the generation of behaviour and in its self-regulation across time: I see the book I need lying over there on the table in front of me and I move towards the table in order to grasp it. When I reach out, I do not from the outset open my hands to whatever objective width I might in some sense register the book to have, then thrust my arm out to grasp it. Certain kinds of robot may do this but humans utilise a much more flexible, self-regulating method which, while perhaps not always as quick and accurate as robotic movements under tightly controlled conditions, is better in dealing with unexpected idiosyncrasies of the situation: as I walk towards it, I progressively raise my arm and open my hand in such a way that the width of my opened hand corresponds to the width of the book just at the moment my hand touches the book.

Obviously, I am guided in the raising of my arm and the opening of my hand not by the objective location and width of the book alone but rather by the way the objective location and width appear to me as I move myself and my hand towards it. It is for this reason that what I need to see is simply the book I need lying there on the table in front of me as the book I need lying there on the table in front of me. For to see this just is for this width to appear perceptually to me in a way which would systematically vary as I move myself and my hand towards it, in so doing adjusting my body and my hand, in particular, the width of my hand, in response to these variations. Notice, however, that what I see the book as exhibits an essential unity of spatial demonstrativeness and spatial orientation, which latter involves essentially anaphorically expressible reference to self. This is the structure an intentional content must have if it is to be embeddable in a mode of appearance, that is, if it is to be the intentional content of a perceptual experience. For only so can it play its part in constituting perceptual experience as able to guide action in the ongoing, self-regulating, feedback fashion manifest in my moving towards and picking up the book.29 Similarly, the hiker was only able to move along the Mt. Tallac trail as he did because his perceptual experience presented the trail to him in a certain mode of appearance to which he could unthinkingly respond. At the same time, this responsiveness, for all its ‘thoughtlessness’, was a matter of intelligently adjusting movement in the light of how the trail appeared to the hiker as there in front of him as he moved along it. As in the example of my seeing the book I need, so, too here: intentional content must display both spatial demonstrativeness and spatial orientation, hence essentially anaphorically expressible reference back to the subject of experience.

We may now simply identify what it is like to undergo a perceptual experience with its mode of appearance and thereby its conscious character with this mode. This identification is plausible because we would expect this character to be something which contributes to the capacity of the perceiver to behave intelligently. Moreover, this identification captures well the character of what-it’s-like-ness, hence of consciousness, as something ‘subjective’ or qualitative, a matter of how things look from the inside. For on this account the what-it’s-like-ness of perceptual experience must be primarily and non-derivatively accessible only from the first-person perspective—assuming, of course, that what undergoes the perceptual experience can adopt the first-person perspective in the first place. This accomplishment is due to the essential spatial demonstrativeness the account ascribes to perceptual experience: perceptual experience just is a matter of things there showing themselves to a subject as thus and so in a certain mode or how of appearance.

Finally and in this context most importantly, this identification captures well what is right in Zahavi’s claim that the what-it’s-like-ness, hence conscious character, of perceptual experience imparts to the subject of such experience a dative status. For the account given here has persistently characterised the what-it’s-like-ness of perceptual experience in such a way that it must be essentially for, or given to, the subject of this experience. This accomplishment is due to the only anaphorically expressible reference back to the subject of experience which this account ascribes to perceptual experience. As such, the presence of this reference in perceptual intentional content entails nothing de me, no kind of self-awareness at all. The account thus not only plausibly describes how conscious perceptual experience practically orients its subject in the world. And it not only shows that the conscious character of perceptual experience matters—matters in the intelligent way one would expect this character to matter. It also clearly fulfils the first condition of adequacy. At least in the case of conscious perceptual experience we need not maintain that consciousness is a form of self-consciousness.

Section 4.2: Is All Perceptual Experience always Conscious?

There is an important objection one could now make to the account of conscious perceptual experience just given: it tacitly assumes that all perceptual experience is conscious since all perceptual experience must have, even to qualify as perceptual experience, the character of presenting something in a certain mode of appearance. It thereby rules out the possibility of unconscious perceptual experience, which is surely wrong.

The first thing to note in reply to this objection is that the account does not rule out unconscious perceptual experience in absolutely every sense of the word. One can say, of the objects and states of affairs of which, in a conscious perceptual experience, I am merely peripherally aware, for example, the fact that the peaks of the mountain range in the example of the hiker previously considered are topped with white snow, that one unconsciously perceives them. The second thing to note is that it is actually not at all implausible to maintain that this is the only sense in which one can unconsciously perceive things. The phenomenon of blind sight is not at all a counterexample to this. Blind sight is a consequence of damage to the visual apparatus. It is therefore a fundamental mistake to think of it as just a form of perceptual experience alongside others, the only difference being that, because of the damage, it is not conscious, as is shown by the fact that it cannot be recalled in the manner of the hiker or of witnesses who recall unconsciously perceived details of a crime scene by re-enacting their actions at the time. The fact that blind sight cannot be recalled by distinctively self-conscious subjects is a defect. Consequently, one may reasonably expect blind sight to be defective with regard to its capacity to guide action and to integrate itself with other things the subject learns through perceptual experience about the world. For this reason, blind sight, precisely because of the way it is unconscious, is not a full or non-defective form of perceptual experience at all. One must therefore resist the temptation to think of blind sight as a model for the kind of perceptual experience enjoyed by non-self-conscious beings such as dogs and cats. Indeed, if one accepts, as so many do, that the conscious character of conscious perceptual experience consists in there being something it is like to undergo such experience, then one should not be surprised that perceptual experience in the full sense of the word is essentially conscious and at most only peripherally unconscious.

Section 5: Towards an Account of Self-Consciousness

The account of consciousness given in the previous section fulfils, for the specific case of conscious perceptual experience, the first condition of adequacy. For evidently even a completely non-self-conscious like a dog perceives things as thus and so in some objectively, contextually and subjectively conditioned way; this way or mode of appearance constitutes the what-it’s-like-ness of its perceptual experience without which it could not behave intelligently at all. So if the consciousness of conscious perceptual experience just is there being something it is like to have such experience, then a being can be conscious without any kind of self-awareness. But what about the second and third conditions? Does our account permit an explanation of why so many have assumed that consciousness is a form of self-consciousness? And can it explain the connection Sartre sees between consciousness and explicit, first-personal thinking? 30 Because the second issue is more important and because addressing it enables one to deal with the first issue quite shortly, I will take the second issue first.

Section 5.1: What is Reflective Self-Consciousness?

It is clearly not possible here to draw out all the implications of this account of consciousness for the concept of self-consciousness—by which I mean explicitly first-personal thinking, as when one thinks, “I am tired,” or indeed “(T)his proposition ‘I am, I exist’ is necessarily true whenever it is stated by me or conceived in my mind.” (Descartes 1641/2003, p.24) Nonetheless, we need at least to explain the sense in which self-consciousness presupposes consciousness, specifically, conscious perceptual experience. Consciousness displays great diversity, from pain and sensation through conscious perceptual experience to conscious belief, desire and ultimately emotion. One should therefore not expect a “one-size-fits-all” account for all cases of consciousness. Yet despite this diversity there are presumably conceptual connections between the various forms which any adequate, hence complete account of consciousness must reconstruct. In particular, the paradigmatic status which many give to conscious perceptual experience when, having followed Nagel in associating the conscious character of such experience with what-it’s-like-ness, they go on to associate the latter with givenness. This paradigmatic status must be explained—explained by showing how precisely self-consciousness depends on conscious perceptual experience.

In fact, once one has accepted the ineliminable status of what I have been calling, admittedly loosely, anaphoric intentionality, it is quite easy to meet this demand. We may stipulate that by conscious thought or belief the following is meant:

S has the occurrent conscious thought that a is Φ if and only if S thinks, “a is Φ.”

By thinking a sentence I mean giving it its literal interpretation. Obviously, there may be other senses in which one can think a sentence, e.g., one can imagine what it would be like to pronounce it. But however many other senses there are, thinking a sentence in the sense of giving it its literal interpretation is the primary sense. So to think a sentence in the sense intended here is a matter of imagining oneself speaking in the locutionary sense of expressing, as one is wont to say, the proposition determined by the meaning of the sentence. Then we may say:

S thinks, “a is Φ,” if and only if S imagines itself saying (in the locutionary rather than illocutionary sense) that a is Φ.

And this yields in turn our analysis:

S has the occurrent conscious thought that a is Φ if and only if S imagines itself saying (in the locutionary rather than illocutionary sense) that a is Φ.

The idea behind this is that conscious thinking is the soul’s internal dialogue with itself. As such, it intimates the surely plausible thesis that internal conscious thinking, of which dogs and cats are, I take it, not capable, presupposes external speaking and therefore presumably also communicating.

One might make two interrelated objections at this point: firstly, consciously thinking is not just a matter of imagining oneself merely locuting; it is surely often a matter of imagining oneself asserting, proposing, discussing, debating, etc. This latter is, of course, perfectly true but here to locute means to perform what Austin called a locutionary act. As such, in imagining itself locuting S might be imagining itself doing all these serious things. Equally, however, S might be simply entertaining the thought literally expressed by the sentence “a is Φ.” Indeed, we must say that S is primarily imagining itself asserting, proposing, discussing, debating or simply entertaining a thought because the illocutionary act is conceptually prior to the locutionary act, if not in the sense that all locution is illocution, then at least in the sense that locution must be as a rule illocution. (One might add that, for this very reason, locution must also as a rule be communication.)

In response to this, one might now secondly object that if conscious thinking were simply a matter of imagining, then it could not be ‘serious’ in the manner required for serious conscious thinking. For surely to imagine is simply to phantasise. But this second objection is just plain wrong. As the first objection rightly assumes, it is perfectly possible to be seriously imagining oneself asserting, proposing, discussing, debating, etc. S may be imagining itself saying what the sentence “a is Φ” expresses in order to think up or think through an argument. As such, S’s imagining is accomplished with perfectly serious intent.

Evidently, this conception of conscious thinking displays a dependence of such thinking on perceptual experience. For the conception of imagination is related to that of perceptual experience: to imagine doing such and such involves imagining what it is like to do such and such, indeed what it is like to do such and such in the manner of perceptual experience.31 When, for example, I imagine myself saying something, I can ‘hear’ myself speaking, hence register, or rather, imagine myself registering, the sound of my voice, the rise and fall of my speech, etc. And this would clearly not be possible had I not actually experienced myself speaking or heard the sound of my own voice. In this sense, perceptual experience is a presupposition of conscious thinking. At this point, we may extrapolate from our general account of conscious thinking to that particular form of it which is explicit, first-personal thinking—what Sartre calls reflective self-consciousness. This is a simple matter:

S has the occurrent self-conscious or first-personal thought that it is Φ if and only if S thinks, “I am Φ.”

This yields:

S thinks, “I am Φ,” if and only if S imagines itself saying (in the locutionary rather than illocutionary sense) that it is Φ.

Which obviously gives us:

S has the occurrent self-conscious or first-personal thought that it is Φ if and only if S imagines itself saying (in the locutionary rather than illocutionary sense) that it is Φ.

This is by no means trivial. For one very obvious thing, it ties self-conscious, first-personal thinking, via the concept of imagination, to linguistic mastery, thereby explaining our strong intuition that such thinking is always in and of a language—typically, of course, one’s mother tongue but not necessarily. For another and less obvious thing, it crucially relies for its informativeness and non-circularity on the idea that the anaphoric construction of which it makes use does not necessarily report occurrences of self-conscious or first-personal thought. It is clear that when a subject says that it is Φ, it must use the first person. But as we have seen, it is just wrong to think that when the subject perceives the Mt. Tallac trail to be there in front of it or comes to believe that it is making a mess, it is performing acts of explicit, self-conscious or first-personal thinking. This intimates a further reason for accepting what I am calling anaphoric intentionality as basic and ineliminable: not only does one do justice to the intentionality of the merely sentient,32 one also does justice to the intentionality of what is not just sentient but also sapient.

Section 5.2: Why have so Many thought that Consciousness is a Form of Self-Consciousness?

Once one appreciates just how poorly grounded the claim is that consciousness is a form of self-consciousness; and once one sees how to accommodate the intuition about the dative status of the subject in the what-it’s-like-ness of conscious perceptual experience: it becomes easy to explain why people have convinced themselves of the claim. We have already touched upon this: the idea that there can be no reference without representation, i.e., no reference to an entity which cannot be represented by a free-standing linguistic expression, up to and including indexicals like ‘I’. And, as we have also seen, one maintains that there is no reference without representation because one believes that intentional contents must be free-standing items corresponding to sentences, even sentences containing indexicals. Lurking behind this is the doctrine of propositional attitudes, or rather, the first and most fundamental tenet of this doctrine, namely, that intentionality is a relation between a subject and some entity, for this requires that intentional contents be free-standing

But how precisely does someone who believes there can be no reference without representation convince themselves that consciousness is a form of self-awareness? Note that if one accepts both the point made by the kinds of example given inter alia by Geach, Altham, Castañeda and Perry; and that there is no reference without representation, then the only expression remaining for expressing what is thought in these examples is the first person pronoun. Here we encounter the reason why Castañeda at least regards the anaphoric interpretation of “The Editor of Soul believes that he is a millionaire” as reporting an explicit first-person judgement and introduces the expression ‘he’ in order to distinguish it from other uses of the pronoun ‘he’. For he regards intentional contents as items towards which one can have an attitude. He is, of course, not alone in this conviction; Perry33, Lewis and indeed most other philosophers apparently endorse it.34 More importantly, here we encounter the reason why so many have taught themselves to find plausible what is actually implausible. When Zahavi and others reflect on the diverse range of phenomena which count as conscious, they are seeing through a glass but darkly something which they can only accommodate by regarding it as somehow first-personal. What, however, they are really glimpsing is essential anaphora or rather intentionality de se which for want of the requisite resources they are unable to distinguish from intentionality de me.


Altham, J. (1980) “Indirect Reflexives and Indirect Speech,” in Intention and Intentionality: Essays for G. E. M. Anscombe, edited by Cora Diamond and Jenny Teichmann, Cornell University Press, pp.25-37

Castañeda, H-N. (1966) “‘He’—A Study in the Logic of Self-Consciousness,” in Castañeda 1999, pp.35-60

Castañeda, H-N. (1968) “The Phenomeno-Logic of the I,” in Castañeda 1999, pp.89-95

Castañeda, H-N. (1999) The Phenomeno-Logic of the I: Essays on Self-Consciousness, edited by James G. Hart and Tomis Kapitan, Indiana University Press, first edition

Christensen, C. B. (2007) “Nichts Neues unter der Sonne: Bewußtsein und Selbstbewußtsein bei Paul Natorp,“ in Kant-Studien, Vol. 98, No. 3, pp.372-398

Christensen, C. B. (2008) Self and World: From Analytic Philosophy to Phenomenology, Walter de Gruyter, Berlin

Christensen, C. B. (2013) “The Horizonal Structure of Perceptual Experience,” in Logical Analysis and History of Philosophy/Philosophiegeschichte und Logische Analyse, Vol. 16, pp.109-141

Descartes, R. (1641/2003) Meditations and Other Metaphysical Writings, translated and introduced by D. M. Clarke, Penguin Books, London

Evans, G. (1984) The Varieties of Reference, edited by J. McDowell, Oxford University Press, Oxford

Geach, P. (1957) “On Beliefs about Oneself,” in Analysis, Vol. 18, No. 1, pp.23-24

Heidegger, M. (1987) Zollikoner Seminare: Protokolle–Gespräche–Briefe, herausgegeben von Medard Boss, Vittorio Klosterman Verlag, Frankfurt am Main

Kriegel, U. (2003) “Consciousness as Intransitive Self-Consciousness: Two Views and an Argument,” in Canadian Journal of Philosophy, Vol. 33, No.1, pp.103-132

Lewis, D. (1979) “Attitudes De Dicto and De Se,” in The Philosophical Review, Vol. 88, No. 4, pp.513-543

Nagel, T. (1974) “What is it Like to be a Bat,” in The Philosophical Review, Vol. 83, No. 4, pp.435-450

Peacocke, C. (1992) A Theory of Concepts, The MIT Press, Cambridge, Massachusetts

Perry, J. (1977) “Frege on Demonstratives,” in Perry 2000, pp.1-26

Perry, J. (1979) “The Essential Indexical,” in Perry 2000, pp.27-44

Perry, J. (1983) “Castañeda on He and I,” in Perry 2000, pp.77-100

Perry, J. (1986a) “Thought without Representation,” in Perry 2000, pp.171-188

Perry, J. (1998) “Myself and I”, in Perry 2000, pp.325-339

Perry, J. (1999) “Self-Notions,” in Logos: Philosophical Issues in Christian Perspective, Vol. 11, pp.17-31

Perry, J. (2000) The Problem of the Essential Indexical and Other Essays, expanded edition, CSLI Publications, Stanford, California

Stoljar, D. (2013) “The Semantics of ‘What it’s like’ and the Nature of Consciousness,” unpublished manuscript, Australian National University, Canberra

Thomasson, A. (2000) “After Brentano: A One-Level Theory of Consciousness,” in European Journal of Philosophy, Vol. 8, No. 2, pp. 190–209

Zahavi, D. (2004) “Back to Brentano?,” in Journal of Consciousness Studies, Vol. 11, No. 10-11, pp.66-87

Zahavi, D. (2009) “Is the Self a Social Construct?,” in Inquiry, Vol. 52, No.6, pp.551-573


  1. See Zahavi 2004, p.82.

  2. For Zahavi these include, in addition to the account given by Brentano himself, also the more recent accounts given by Thomasson (2000) and Kriegel (2003). Zahavi calls these latter neo-Brentanians but objects to this label being applied to the phenomenological conception—see Zahavi 2004, p.75.

  3. Sartre gives his account across three texts: his first publication in philosophy, The Transcendence of the Ego of 1936, his main work Being and Nothingness from 1943, and the article “Self-Consciousness and Self-Knowledge” of 1948.

  4. Sartre is quite explicit about this: “(C)onsciousness,” he writes, “is self-consciousness. It is this same notion of self which must be studied, for it defines the very being of consciousness” (Sartre, 1976, p. 114). It is to be noted that the later Heidegger explicitly rejects this view: see Heidegger 1987, H 284-285. There can be no doubt that Heidegger’s remarks here are just as much directed against Sartre as against the philosophers he explicitly mentions, namely, Husserl and Descartes. The later Heidegger is thus not to be counted as an adherent of the allegedly phenomenological conception of consciousness. If one could maintain that the later Heidegger could no longer be regarded as belonging to the phenomenological tradition, then one could still maintain that the Sartrean conception of consciousness is also the phenomenological one. Perhaps indeed Zahavi thinks this, i.e., perhaps he thinks that the phenomenological tradition is defined by the kind of view one finds in Sartre and that therefore there is a radical difference between the later and earlier Heidegger. In fact, I think that both these theses are false. If this is so, then a different interpretation must be put on the passages from the early Heidegger which Zahavi adduces as showing the early Heidegger’s commitment to the allegedly phenomenological conception of consciousness. The account of anaphoric intentionality, essential spatial demonstratives and the structure of perceptual experience developed here provide resources for such an alternative interpretation.

  5. See Zahavi 2004, p.70, p.74, p.79, p.80, p.81, p.85 and p.86.

  6. Note that there appears to be implicit in, perhaps even inherent to, this argument a restriction to those psychological states and experiences which are genuinely contentful, i.e., intentional.

  7. Note that in order to get the position which distinguishes both Brentano and the neo-Brentanians, and Sartre and ostensibly the phenomenologists, from higher-order conceptions, one has only to maintain, not implausibly, that if the character of a psychological state or experience as conscious is a matter of there being something it is like to be in this state or undergo this experience, then it must be intrinsic to the state or experience since the what-is-likeness of a psychological state or experience is indeed intrinsic to it.

  8. Note that to assert the foundational status of conscious perceptual experience is not deny that other forms of consciousness, in particular, forms of sensation such as pain, may be equally foundational. This I believe to indeed be the case—see Christensen 2013.

  9. See Zahavi 2009, p.555. I suspect that to acknowledge that there is something right in the idea that the phenomenal character of a conscious psychological state or experience—its what-it’s-likeness—is always for the subject of the state or experience is implicitly to acknowledge that the ‘for’ in such sentences as “There is something it is like for Alice to have a toothache” does more than merely mark the fact that, as Stoljar points out—see Stoljar 2013, p.5—, the name ‘Alice’ is the covert subject of the infinitive verb phrase ‘to have a toothache’. The ‘for’ has a more-than-grammatical or -syntactical function, something Stoljar himself acknowledges but does not explain when he introduces his hypothesis H2—see Stoljar 2013, p.14.

  10. The description given in the passage just quoted is consistent with a number of different processes of thought Perry possibly goes through.

  11. In other words, Perry’s coming to believe that he is making a mess is what Kriegel calls, following Malcolm, “a non-conscious occurrent mental state.” (Kriegel 2003 p.116)

  12. It would certainly be odd, at least relative to the conventions of everyday English discourse, for Perry to use this anaphoric construction rather than the first person. Even so, it is still perfectly acceptable as an explanation—as is in fact shown by the possibility that by the time he comes to express his beliefs and the behaviour to which they gave rise Perry might have forgotten that he is John Perry, that he was the shopper in the supermarket who was making a mess, who he was thus trying to catch, etc.

  13. Castañeda points out that it is possible to understand the pronoun ‘he’ in sentence “The Editor of Soul believes at time t that he is a millionaire” demonstratively, as when a speaker utters this sentence with the clear intention that the pronoun ‘he’ refer to a person he is pointing to—see Castañeda 1968, p.91. Clearly, it is equally possible to use the pronoun ‘it’ in this demonstrative way in the case of the sentence “S sees a dog aggressively charging towards it.” Indeed, this possibility, combined with the fact that such demonstrative use very obviously does not capture the causal efficacy of S’s perceptual experience even when the explainer is pointing to S itself, just drives the point home that the anaphoric use is essential. Incidentally, it would also seem possible to use the pronouns ‘he’ and ‘it’ anaphorically without binding them to the singular terms ‘The Editor of Soul’ and ‘S’ but rather to some other singular term wielded in some other part of the discourse in which utterance of these sentences is embedded. Here, too, it is clear that the causal efficacy of S’s perceptual experience is not captured, as is shown by the fact that the only way such an explanation could reveal why S behaved as it did would be if it were clear that this other singular term referred to S.

  14. Castañeda does recognise the first as a distinct question: on p.42 of Castañeda 1966 he concludes that names and descriptions cannot replace occurrences of pronouns occurring within psychological contexts which are tied anaphorically to the subject of the psychological verb.

  15. I first introduce this notion and terminology in Christensen 2007.

  16. This points to an important but for current purposes incidental point: the fact that S is a King Charles Spaniel may well be causally relevant since such dogs are typically small and small dogs typically run away from larger ones. With this, we see that the explanation given of S’s behaviour is incomplete in that the dog doing the charging has been inadequately characterised: we need to say that S sees a larger dog agressively charging towards it. (Obviously, this incompleteness is irrelevant to the issue of whether the anaphorically used pronoun is replaceable salve explanatione by some co-referential singular term, even one specifying S as a small or smaller dog.)

  17. If, however, the intentional experience or event captured in such anaphoric use of a third person pronoun is the agent’s thinking that it is identical with so and so or the such and such, then this thinking, this identity-judgement, may well have to be explicitly self-conscious and first-personal. My dog hears footsteps at the front and begins to bark, thinking that someone is there. Then it sees I, its master, am there and so stops barking. Do we need at any point to attribute to my dog the identity judgement that the person there is its master? Surely not. First the dog believed that someone was there and behaved accordingly, then it believed that its master was there and behaved accordingly. This suggests that an identity judgement of the indicated kind is needed only when change in belief and behaviour is due to recognition of error, hence is under the rational, self-conscious control of the believer. In other words, such identity judgement arises when change in belief and behaviour is a matter of rational self-regulation—in which case the identity judgement will always have to be explicitly first-personal.

  18. At one point, Perry seems to get a bit muddled. He writes, “The child [in his time zone example] says, “It is now 7 p.m.,” treating being 7 o’clock p.m. as a property of the present time, rather than a relation between times and places.” (Perry 1998, p.328) Surely this entails that the child has a concept of time such that if it is 7 p.m. anywhere, it is 7 p.m. everywhere. Then, however, all the child’s judgements about times of day would be false because all sorts of different places, even across different time zones, can share the same moment of time; by definition, no moment of time is one time of day across different time zones. So Perry cannot mean what he might initially seem to mean. The example of the child who only cares about the local weather shows that he does not in fact mean this.

  19. Indeed, in this latter passage, Perry immediately juxtaposes to the lack of need for self-reference a lack of need for self-representation: quoted fully, the passage reads, “(t)here is no need for a self-referring component of our belief, no need for an idea or representation of ourselves.” (Perry 1986a, p.182; italics added)

  20. I thus do not mean that there is no reference without Fregean sense, which, given the usage employed here of the term ‘representation’ is merely a specific way of construing reference as representation.

  21. Indeed, it is natural to use the demonstratively adverbial phrase ‘immediately there’, for this phrase captures, from the third-person perspective, the fact that the trail is here for the hiker: we are assuming the Mt. Tallac trail to be nearby, hence here for hiker but further away, hence there for the describer.

  22. It is clear that, as a matter of brute linguistic fact and convenience, one may leave these demonstrative adverbs of place out of sentences used in describing the experienced hiker’s belief and perceptual experience. Similarly, of course, one may leave the demonstrative adverb of time out: in Perry’s exampe of the academic who has to go to a meeting at 12 noon—see Perry 1979, p.28—the academic comes to believe simply that the meeting is starting while Perry’s hiker comes to believe simply that the Mt. Tallac trail is in front of him, or again, sees simply the Mt. Tallac trail in front of him. But the permissibility of leaving these demonstrative expressions out does not mean that the demonstrativeness the expressions mark is not there in the intentional states or experiences described by sentences from which they have been left out.

  23. Implicit in this argument is another reason why no non-demonstrative characterisation of the hiker’s belief will do, i.e., why spatial demonstration is necessary. The hiker has no need, either from the outset or at any point in his movement towards it, see how far away the trail is. Rather, he just regulates and adjusts his movement in the light of the changing way the trail appears there in front of him as he moves towards it. So all he needs to see or believe, across his entire movement towards it, is its being there in front of him. For this reason, the propositional function ‘… is about 10 metres away from him’, unlike the propositional function ‘… is in front of him’, is not in the intentional content of the hiker’s perceptual experience. Yet only by putting some such propositional function into the content could one hope to provide any alternative to a demonstrative characterisation. So the attempt to deny spatial demonstrativeness is inherently distorting. The same cannot be said, of course, of the propositional function ‘in front of him’ because this is relevant to how the hiker’s perceptual experience shapes his subsequent behaviour. It fails, however, to specify any particular location for the object perceived and so, because no other propositional function can combine with it to accomplish this, it must work in tandem with spatial demonstrativeness.

  24. If a perceiving subject is disoriented, it might well see something simply as there and have then to reflect upon its experience in order to determine the orientation of what it sees to it. But if a subject is disoriented, it cannot automatically and directly act.

  25. One can think of the spatial demonstrativeness—its thereness or hereness—as analogous to the radial coordinate in a pair of polar coordinates which must, of course, in order to do its job, be accompanied by the angular coordinate, to which the orientation of the object perceived to the subject is analogous. (Note how this suggests that the depth dimension is the most important of the three dimensions of perceptual and actional space—contrary to what much modern philosophy and psychology have assumed.)

  26. Part of the problem with Perry’s argument is precisely that all the examples involve very sophisticated beings engaged in sophisticated activities, capable of the first person and indeed of language. This complexity constantly obscures what is essential to the intentional states and experiences being described; and what is essential to the description of these intentional states and experiences.

  27. This example is derived from Peacocke—see Peacocke 1992. I have used this example elsewhere—see Christensen 2008—, admittedly not in connection with either Perry himself or his example.

  28. The hiker thus recalls his perceptual experience not in the sense in which a school pupil might remember that the Battle of Hastings took place in 1066 but rather in the sense of remembering what it was likeboth what the perceptual experience was like and what the mountain range was like.

  29. Perry’s notion of agent-relative knowledge shows that he is entirely focussed on (what he takes to be) the strictly intentional, directedness-securing content of perceptual experience. This is a mistake; unless one sees that perceptual experience is contentful also in the sense of possessing a certain how or mode of appearance in which its strictly intentional content appears as satisfied, one will get no adequate account of what its action-guiding agent-relativity is.

  30. It is worth noting that this account clearly meets what one might well regard as a fourth condition of adequacy, namely, that it gives the conscious character of conscious perceptual experience something to do. For it is clearly crucial that consciousness have some functional role which renders it more than a mere epiphenomenon. There being something it is like to undergo perceptual experience enables the subject of experience to respond intelligently, precisely as a conscious being would, to what it perceives things objectively to be.

  31. There is indeed a crucial difference between the what-it’s-like-ness of perceptual experience and that of sensation. In the former case, for example, in vision there is something qualitative to perceptual experience but in no sense do I feel anything. There is something it is like to undergo visual and aural experience but I am not affected them in the way or sense I am affected by, e.g., a burning flame applied to my skin or even the pressure of someone’s hand when they shake mine. This latter example indicates that there is one form of perceptual experience in which both kinds of what-it’s-like-ness are combined: touch. In this regard, Husserl has some interesting observations to make about the phenomenological of tactile experience.

  32. Note that this account of explicitly self-conscious, first-personal thinking eschews any suggestion that such thinking is in any way a matter of forming beliefs about beliefs. This is an immensely important advantage because one simply cannot explain the behaviour of higher animals like dogs, beings which nonetheless are in capable of the Kantian “I think,” without attributing to them beliefs about beliefs, indeed a cut-down or simpler version of what Lewis and Schiffer call common and mutual knowledge respectively.

  33. In order to deal with the problem raised by indexicals and demonstratives for a Fregean doctrine of sense, Perry distinguishes between what Frege understands by sense (Sinn) and by the thought (Gedanke). In consciousness. This leads him at one point to suggest that intentional states and experiences of the kind exemplified by belief be construed as consisting in two interconnected relations to something. In the first instance, they are relations to relativised propositions, which Perry defines as “abstract objects corresponding to,” hence individuated in terms of, “sentences containing indexicals.” (Perry 1979, p.39) This permits them to be relations to a second kind of object which is defined in terms of the first: an intentional state or experiences of the kind exemplified by belief is a relation to the ordered n-tuple consisting of a relativised proposition “shorn of its indexical elements” (Perry 1979, p.40) and the referents of the sentence to which this relativised proposition corresponds.

    Lewis criticises this proposal on the grounds that it is overly complex—see Lewis 1979, p.537. He proposes an alternative conception of belief (and, mutatis mutandis, of other intentional states and experiences) which is simpler in that it postulates only one object for belief to be an attitude towards while dealing just as well with the problems which motivates Perry to propose his conception. But there seem to me to be at least two significant difficulties with this alternative conception.

    Firstly, it is not clear to me at all that Lewis’ proposal is coherent. Lewis distinguishes between belief de dicto and belief de se (and also, mutatis mutandis, for all other intentional states and experiences). Belief de dicto is of course belief in which no indexicality is involved. What Lewis calls belief de se is belief of the essentially anaphoric kind of which Geach, Altham, Castañeda, Perry and others have made so much. Now belief de dicto, says Lewis, is self-locating in logical space in the sense that it is, as belief in a certain proposition in the traditional sense, always at the same time a matter of self-ascribing the property of inhabiting a possible world in which the proposition is true. Belief de se is self-locating in ordinary space and time in the sense that it is a matter of demonstratively identifying either who, when and/or where one is—see Perry 1979, p.29. Lewis now suggests that there is a commonality here: “(b)elief de dicto is self-locating belief with respect to logical space” while “belief irreducibly de se is self-locating belief at least partly with respect to ordinary time and space, or with respect to the population. I propose that any kind of self-locating belief should be understood as self-ascription of properties.” (Lewis 1979, p.522) This permits us to take the properties self-ascribed in these two kinds of belief, which are, of course, the only kinds of belief there are, as the objects of belief. This gives us uniform objects of belief and uniform objects of belief “ … facilitate systematic common-sense psychology,” (Lewis 1979, p.514) i.e., enable systematic explanation of the role played by belief and the other ‘attitudes’ in causing behaviour.

    Now by self-ascription Lewis must surely mean a particular species of ascription in the normal sense. Indeed, Lewis himself says, “Self-ascription of properties is ascription of properties to oneself under the relation of identity,” (Lewis 1979, p.543) which clearly involves appeal to ascription in the standard sense—the sense which permits one to say that when S asserts that a is Φ, S ascribes to a the property Φ. But if self-ascription is a case of ascription in the standard sense of a property to something, then whenever S asserts that a is Φ, S ascribes to itself the property of inhabiting a possible world in which a is Φ; and ascribes to itself the property of ascribing to itself the property of inhabiting a possible world in which a is Φ; and so on, ad infinitum.

    Secondly, how would one develop, on this conception of belief and intentionality generally, a plausible account of animal intentionality, that is, the intentionality of beings not capable of explicitly self-conscious, first-personal thought? For what kind of self-ascription is self-ascription? Can this view avoid entailing the clear falsehood that one can only have belief and other forms of intentionality if one can have them self-consciously, i.e., if one can quite literally self-ascribe them—what Kant means when he speaks of attaching an “I think” to one’s representations? For how else are we to understand what Lewis means by self-ascription except as the capacity explicitly to judge, “I inhabit a possible world in which a is Φ”?

  34. Searle is an exception. He, too, rejects the idea that intentional states and experiences are propositional attitudes, i.e., a relation of a subject to an entity called a proposition. He does not, however, reject the idea that intentional contents are fundamentally propositional.