Searle against the world:

how can experiences find their objects?

KENT BACH

Department of Philosophy

San Francisco State University

kbach@sfsu.edu

Here's an old question in the philosophy of perception: here I am, looking at this pen [I hold up a pen in my hand]. Presumably I really am seeing this pen. Even so, I could be having an experience just like the one I am having without anything being there. So how can the experience I am having really involve direct awareness of the pen? It seems as though the presence of the pen is inessential to the way the experience is.

Traditionally, this question was used to raise skeptical worries about perceptual experience and to motivate the sense-data thesis, according to which perceptual experiences, even veridical ones, are directly of mental or private objects and only indirectly of their physical objects. I addressed these epistemological and ontological problems in my dissertation nearly thirty years ago, but it's not yet time for me to take them up again. What interests me today is what might be called the semantic problem of perceptual experience. It concerns the fact that experiences are directed at objects. This is what Searle and many other philosophers call the Intentionality of experience.

Here I am looking at this pen. What makes it the case that it is this pen that I am experiencing, that it is the one that appears to me of such-and-such shape, size, and color? The problem here is that there is nothing in the character of my visual experience to distinguish this pen from any other pen that would look just like it. If this pen were replaced instantaneously by another, I wouldn't and couldn't tell the difference. Accordingly, the usual answer to the question of why this pen is the one I am experiencing is that it is the one that causes (in a certain characteristic way) my visual experience. Had another pen been in its place, the other pen would have been the one I am experiencing. Had this pen been replaced instantaneously by another, the other pen would have immediately become the one I am experiencing.

John Searle finds this answer objectionable. It's not so much that the answer is incorrect as that it is given from the "third-person point of view." It's the kind of answer that would be given to the question, what would make a photograph of a pen a photograph of this pen? In that case, a straightforward causal answer is clearly the right one. To be the pen "in" the photograph, a pen would have had to reflect light into the camera and on to the exposed film, etc., etc. Had a different pen been the one that did this, it would have been the one in the picture. Appealing to the analogous causal fact in the case of visual experience "fails," according to Searle, "to answer the question as to how this fact gets into the Intentional content" (1983, 63). Searle calls this a "first-person internal" question.

Searle clearly appreciates that in visual (and tactual) perception we are aware of particular things in the world. We are aware of this pen, not just that there exists a thing of a certain sort. In contrast, if we see a shadow or a footprint, we are not aware of the thing that casts the shadow or the person that left the footprint. Seeing a shadow enables us to think of the thing that casts it only under a description like 'the thing that is casting this shadow.' Seeing a footprint enables us to think of the person that casts it only under such a description as 'the person who left this footprint.' Similarly, if we saw things only in photographs and films, our cognitive access to them would be equally indirect. We would be in a position even worse than Chauncey Gardiner's, the hero of Jerzy Kozinski's novel Being There, whose access to things other than those in his room is only via television. At least he got to see the things in his room.

If actually seeing things were like seeing things in pictures and films, the connection between things in the world and our experience would be merely causal. I have nothing against causal relationships but, as William Alston says, "Causality is no substitute for awareness" (1997, 12). Searle of course would agree. When he sees an object such as his magnificent yellow station wagon, he says that the experience he has "is directly of the object. It doesn't just 'represent' the object, it provides direct access to it. The experience has a kind of directness, immediacy and involuntariness which is not shared by a belief I might have about the object in its absence" (1983, 45). That is why he prefers to call perceptual experiences "presentations" rather than "representations" (1983, 46). The question, though, is whether his official account of the contents of perceptual experience does justice to fact that experiences are directed at particular things. We will get to that account shortly.

1. The puzzle of experience and the problem of particularity

Alston mentions the line of argument which goes from the premise that "hallucinatory experience can be [subjectively or qualitatively] indistinguishable from the real thing" to the conclusion that "we can't regard it as intrinsic to perceptual experiences that there is a direct awareness of objects" (1997, 4). After all, so the argument goes, how can an object be presented to us in experience if a qualitatively indistinguishable experience did not involve the presence of any object? Even when the object is there, it seems inessential to the experience, for the experience could have been just as it is without the object being there.

On the other hand, it does seem that in ordinary perceptual experience we are directly aware of external objects-they appear to us, they are present to us. So we have a dilemma, what J. J. Valberg calls "the puzzle of experience" (this is the title of his 1992 book): how can a perceptual experience be a direct awareness of an object if it is indistinguishable from some possible hallucination? What puzzles Valberg is that the object seems "potentially irrelevant." To make the puzzle vivid, he goes so far as to imagine that during the course of an experience of a certain object, say this pen, God intervenes, eliminating the pen in an instant but preserving our experience. Accordingly, so the argument goes, the experience could be just as it is without the presence of the object. (Arguably, this is merely a doxastic possibility, not a metaphysical one, at least on the plausible supposition that a given event (token), such as an experience, could not have had a different cause from its actual cause, but that doesn't affect Valberg's point-we can always compare a veridical experience with an hallucination.)

Is it really true, then, that this pen "presents" itself to us? Is appearing really the direct relationship that Alston says it is? The phenomenological support for its directness seems to be offset by simple causal considerations, e.g., that one's experience of the pen could be just as it is even if the brain state underlying it were caused by a drug, an evil demon, or a mad scientist who has immersed one's brain in a bubbling vat. Inveterate internalist that he is, Searle wants to say that he "can have exactly that [experience] even if it is a hallucination, even if the [pen] does not exist" (1991, 239). In this regard he distinguishes the content of the experience from its object, and maintains that "the content can be exactly the same even if no object exists." He agrees with traditional epistemologists that there is nothing about the character of a perceptual experience that precludes the possibility that it does not have the object-or the cause-it appears to have.

Now in order to make sense out of all this, it is not enough to pay lip service to the phenomenological fact that experiences seem to have objects. We cannot do justice to the fact that ordinary perceptual experiences involve the immediate presence to us of physical objects if we cannot reconcile this with the fact that experiences with the same character could occur in the absence of any physical objects.

Related to the puzzle of experience is the problem of particularity. This problem is posed by Searle's question, "What is it about this experience that requires that it be satisfied by the presence of [this pen] and not just by any [pen] with such and such characteristics type identical with [this one]?" (1983, 63). Unfortunately, Searle's explicit discussion of the problem of particularity (1983, 62-71) does not address this question. As John McDowell has pointed out, it is limited to cases where particularity relates to a prior identification" (1991, 223). And, as Alston notes, "this is not the problem initially advertised-how the propositional content can be singular in form-but rather the problem of reidentification-how the particular referred to can be picked out as the same one as previously picked out. … It should be obvious that the latter problem presupposes the former" (1997, 18). So let me repeat the question, as Searle himself formulates it, that poses the problem of particularity.

(PoP) What is it about this experience that requires that it be satisfied by the presence of this pen and not just by any [pen] with such and such characteristics type identical with this one?

It is easy to see what the problem is: there seems to be nothing about the experience itself, no feature of its character or content, that makes it be an experience of this pen in particular. It seems that the only fact relevant to that is the causal relation between the pen and the experience, but citing this fact, says Searle, does not solve "the first-person internal problem." Then what does?

2. Searle's inherently descriptivist solution

It appears that we are faced with a dilemma. We can try to characterize the content of a visual experience either in terms that make reference to the object of the experience or in terms that do not. If we characterize it in a way that does make reference to the object, then we have violated the requirement that "the content can be exactly the same even if no object exists." Call this the "existence-independence" requirement. On the other hand, if we try to characterize the content in a way that meets this requirement, it seems we can do so only in general terms. The dilemma is analogous to a dilemma about proper names in the philosophy of language. There is a strong intuition that proper names "directly refer" to their bearers, that they do not have Fregean senses and are not equivalent to Russellian definite descriptions. On the other hand, it seems that names without bearers, like 'Pegasus' and 'Santa Claus,' are perfectly meaningful. It is not only meaningful but true to say, for example, that Santa Claus does not exist. In general, it seems that a proper name's having a meaning does not depend on its having a reference. On the other hand, it seems to many philosophers that sentences containing proper names express singular or object-involving propositions. But how can this be in the case where the name does not stand for any actual thing? I can't go into the problem of proper names here, but I should note that both Searle (1983, ch. 9) and I (Bach 1981, 1987, chs. 7 & 8) have offered solutions that reject the terms of the dilemma. I mention it just for purposes of analogy. The analogous dilemma is that a visual experience can have its content without having an object and yet it seems to be directly of its object.

Searle rejects the view that the contents of experiences are singular propositions. I totally agree with him on this, because such propositions contain nothing about the way in which the object is (re)presented (Bach 1987, 14-15). He rejects any view on which the contents of experiences are "object-dependent," such as the Evans (1982)-McDowell (1984) theory of de re senses. Also, he has a long-standing aversion to the notion of de re belief. He objects, for example, to Tyler Burge's (1977) view that a de re belief or, for that matter a visual experience, is one "whose correct ascription places a [person] in an appropriate nonconceptual contextual relation to the [relevant] object" (1977, 346). Searle maintains that a de re belief or a visual experience for that matter can be fully characterized in terms of its content. Moreover, he rejects Burge's assumption that to be fully statable the content must be purely conceptual, and completely analyzable in general terms. He thinks "there is a third possibility, … forms of Intentionality which are not general but particular and yet are entirely in the head, entirely internal" (1983, 211). To realize this possibility he invokes the ingenious idea of "causal self-referentiality," a resource that is overlooked on a Fregean or a Russellian approach. That sort of approach, on which the object of the experience is whichever thing satisfies a certain general condition laid down by the experience, would leave us with materials insufficient to determine any one object as the one to which the experience is directed. As John McDowell describes it, Searle's strategy is to exploit "the possibility of anchoring the particular-directedness of the [experiences] in question to the particularity of the relevant experiences themselves" (1991, 216). That is signalled by the phrase 'this visual experience' in Searle's specification of the satisfaction conditions of a visual experience (1983, 48; 1991, 228):

(CVE) Vis Exp (that there is a yellow station wagon there and the fact that there is a yellow station wagon there is causing this visual experience)

The idea here is that any visual experience of the same character determines a condition of satisfaction of this form, but since the condition contains a reference to the particular (token) experience, the requirement that each such experience lays down is distinct-it is that very experience, not just any experience like it, which something must cause to be the object of the experience. CVE is satisfied only if there is an object of the sort the experience represents there to be and if there being such an object causes the experience.

Searle's view is not just a new version of the old causal theory of perception. On his view the causation is "Intentional causation" and the experience is "self-referential": how the experience is caused enters into the conditions of satisfaction that its content determines. (At one point Searle (1983, 49) describes both the content and, three lines later, the experience as self-referential; he has also (1991, 231) described the condition of satisfaction as self-referential, but I won't quibble about this here.) The question is whether this "causal self-referentiality" is enough to solve the problem of particularity. Does it do justice to the phenomenological fact that in visual experience objects appear to us, are "present" to us? And, does it do justice to the psychosemantic fact that visual experience actually puts us in a position to think of and refer directly to particular things in the world, and to do so not merely under a description, e.g., 'the thing that causes this visual experience' or 'the thing that looks thus-and-so'? In short, has Searle solved the problem of particularity?

In my opinion, he has not. I am not worried by what worries Burge, that the self-referentiality required on Searle's account is "too complex or too sophisticated" (1991, 198). Searle is careful to point out that he does not mean that the experience literally refers to itself, i.e. in the description "the cause of this experience." Obviously small children and animals have visual experiences of particular things, but Searle claims that they can do so without benefit of the concepts of cause or experience. I am inclined to agree that we can experience causal relationships-between objects, between objects and ourselves, and, when we do things, between ourselves and objects-without having the concepts necessary for articulating these experiences. And I am not worried about the skeptical objection that one can seem to see causal relationships which do not actually obtain. This is no more worrisome than the fact that one can seem to see objects that are not really there.

What worries me are two other features of Searle's account. First, the formulation given by CVE does not accommodate the case of misperception. Suppose I am having an experience as of a yellow station wagon but what I am looking at is not a yellow station wagon but a hearse painted orange and seen in yellow light. I am experiencing that hearse, but CVE requires that there be a yellow station wagon causing this visual experience. So CVE, at least as it stands, is too restrictive. Searle needs to find a way, in giving the form of the content of experience, to isolate the representation of the causal role of the object from what the object is represented as. His formulation needs to separate what determines the object of the experience from what features the object is required to have if the experience is to be veridical. I'm not sure how such a formulation would go, but I take this first problem to be technical, not fundamental.

There is a deeper problem that worries me about Searle's account, as encapsulated in CVE. As CVE has it, the awareness of the station wagon is only "under a description." I don't mean by this that the experience is in words, but only that the person is aware of the station wagon in a way expressible by a definite description. A definite description, of course, is a phrase that denotes one thing uniquely (if it denotes anything at all). Simple examples include 'the discoverer of X-rays' and 'the king of France.' Notice that a sentence containing a definite description can be meaningful and express a determinate proposition even if nothing satisfies the description. So, for example, the sentence 'the king of France is bald' is meaningful and expresses a certain proposition even though France has no king. As Russell demonstrated, even when the description is satisfied, as with the sentence 'The discoverer of X-rays was bald,' the proposition expressed does not involve the individual who discovered X-rays. That proposition is made true by a fact about that individual, but the proposition itself does not contain that individual as a constituent. This proposition is a general, existential proposition, or what I call a "uniqueness proposition," to the effect that some one individual discovered X-rays and that this individual was bald. It is important to appreciate that phrases containing indexicals or demonstratives, such as 'your mother' or 'the author of that book,' can still qualify as definite descriptions. Sentences containing such descriptions, such as 'Your mother is bald' and 'The author of that book is bald,' although made true by facts about particular individuals (your mother or the author of that book), do not express propositions involving those individuals-they express general, uniqueness propositions.

So you can guess what my worry is about Searle's account. As CVE has it, the condition of satisfaction of a visual experience is a general proposition. Just look at the language: the clause 'that there is a yellow station wagon there' obviously expresses a general, existential proposition, and so does the clause, 'the fact that there is a yellow station wagon there is causing this visual experience.' The condition of satisfaction essentially involves the visual experience in question and it does not essentially involve the station wagon. So Searle's account in terms of CVE does not solve the problem of particularity.

Searle rightly claims that visual experience is "not general but particular." The problem is that CVE does not capture this fact. As CVE has it, the content of one's experience of the yellow station wagon is general, albeit in a way that adverts to the particular visual experience. The content is particular with respect to the experience but it is general with respect to the station wagon. CVE represents one's awareness of the station wagon as being only under a description, and has the content of the experience as being to the effect that there exists (a unique) something of a certain sort, namely a yellow station wagon. So, just as Alston says "causation no substitute for awareness," I say description is no substitute for awareness either.

In saying this, I agree with Burge who, despite his misplaced contention that Searle's account is too complicated and too sophisticated, rightly comments that "knowledge of the causal relation between visual experiences and physical objects seems posterior to the experience (and knowledge of the experience) of seeing physical objects" (1991, 206). And I agree with McDowell, despite my aversion to his de re senses, when he says, "In Searle's picture, 'externally' directed particularity in content is secured indirectly, by way of contents targeted on 'internal' items. … [This] inverts the priority that Searle's discussion suggests of 'this visual experience' … over 'that ]pen]'" (1991, 224); "It is as if Searle has here forgotten, or perhaps never quite took the measure of, the 'direct realism' that he resolutely espouses" 1991, 222-3). Although Searle accepts the naive view that perception is "direct" (1983, 46 & 57)-he is anything but a sense-data theorist-his account of the content of a visual experience in terms of CVE suggests that this relation is indirect: one is aware of the object as just the cause the experience. I don't mean to suggest that Searle's descriptivism is solipsistic in the way he thinks phenomenalism is solipsistic (1983, 60), but his view does seem to imply that we are aware of physical objects only by proxy. Even though the satisfaction condition of an experience requires that an object of a certain sort stand in a certain relation to the experience, the condition does not specify which object it is. How could it? How could the satisfaction condition, insofar as it is determined by the internal content of the experience, do more than determine what sort of state of affairs must obtain for the experience to be veridical and, in the process, what sort of object(s) must be involved in that state of affairs? The object itself cannot enter into the satisfaction condition because that condition is what it is independently of the existence of that object (or the obtaining of that state of affairs). It seems, then, that Searle's account does not capture the intentionality or direct(ed)ness of visual experience.

3. Why content can't determine conditions of satisfaction

A central feature of Searle's view of content in general is that the content of an Intentional state determines its conditions of satisfaction (although many philosophers use 'narrow content' or 'internal content' to mean what Searle means by 'content' simpliciter, and use 'content' simpliciter to mean conditions of satisfaction, I will follow Searle's usage). I agree with Searle's internalism about content and with the existence-independence requirement, that is, that the content of a state is what is independently of how the world is (Bach 1982). A brain in a vat could have states with just the contents that ours have. However, I think the problem of particularity cannot be solved unless we give up the supposition that the content of an experience determines its conditions of satisfaction. For the content of the experience cannot determine its object. Even if the condition makes reference to that very experience, as in CVE, it does not require that the object which in fact causes that experience must be the object that causes it. The requirement is only that there be some object or other that causes it.

Solving the problem of particularity requires explaining how the object gets into an experience's conditions of satisfaction. For Searle "it is essential … that facts [about the presence and features of the object causing the experience] enter into the Intentional content" (1983, 48). In his view, the experience represents its object as standing in this causal relation, and it is not sufficient simply that there in fact be a object in this relation, represented or not. Searle thinks that those

who rely on perceptual and indexical [states] … correctly see that there is a class of [states] that cannot be accounted for in purely general terms [and] that these [states] depend on contextual features. [But] they then mistakenly suppose that these contextual features cannot themselves be entirely represented as part of the Intentional content. (1983, 214)

Searle claims that the element of causal self-referentiality does the trick:

Once you understand that the visual experience [of a man wearing a red cap] has a causally self-referential propositional content you don't need to worry about 'describing' or 'conceptualizing' anything in words in order to individuate the man: the Intentional content of the visual experience has already done it. (1983, 212).

But it can't do this, not if it is as Searle formulates it:

(There is a man there causing this visual experience and that man is wearing a red cap.)

This is a blatantly general, existential proposition, the condition on whose truth requires merely the existence of a man of a certain sort. Once again Searle's actual formulation of the content does not do justice to the fact that the experience is of a particular individual, one who figures essentially in the experience's condition of satisfaction. The contents of experiences are not general propositions.

It seems to me that Searle is constrained to specify experiential contents as general propositions because, first, he recognizes that they can't be singular propositions but, second, because he can't shake the assumption that contents determine conditions of satisfaction. The contents of experiences can't be general propositions, but they can't be singular propositions either, because such propositions contain nothing about the way in which the object is (re)presented. But if the contents of experiences are not singular propositions and they are not general propositions either, what are they? In my view, they are not propositions at all, at least not complete propositions. They are indexical. We can make sense of this idea once we give up the supposition that the content of a perceptual experience determines its condition of satisfaction.

Before trying to make sense of it, let me raise one other problem for this supposition. What is to prevent a state other than a perceptual experience, say a vivid fantasy (not an hallucination), from having the same condition of satisfaction as a perceptual experience? If the content of the experience determines its condition of satisfaction, then why wouldn't a fantasy with the same content have the same condition of satisfaction? Searle might object that a fantasy couldn't have the same content, because it doesn't contain any self-referential causal element. But one could turn the point around and say that its not a difference in content that makes the difference in conditions of satisfaction between the perceptual experience and the fantasy but rather, as I wish to suggest, their difference in psychological mode.

4. The indexicality of experience: why the relation between object and experience is not part of content

Whereas for Searle an experience must represent its object as standing in a certain causal relation, in my view it is sufficient simply that there in fact be an object in this relation, represented or not. One is visually aware of the object not because one represents the relation between oneself and the object but because one is in that relation-the condition is imposed simply because the state in question is a visual experience. It is not the content of a visual experience that determines that its cause is its object. Rather, it is the psychological mode. It is the fact that the experience is perceptual that determines that its object(s) is that which causes, in the appropriate way, that very state. (Analogously, as argued in Bach 1978, it is the psychological mode of what Searle calls an "intention in action" that determines that the change that fulfills its condition of satisfaction is the one that it causes in the appropriate way.)

On the indexical view, the causal self-referentiality of an experience, or what I prefer to call its token-reflexivity, resides not in its content but in its condition of satisfaction. In this respect, its content does not fully determine its condition of satisfaction. The content determines only a schema for the condition of satisfaction of any particular experience with that content. Specified schematically, the content includes a slot, represented by a free variable, for the object that is the cause of the experience. In the case of seeing a yellow station wagon, it takes the form,

(IVE) Vis Exp (yellow station wagon there (x))

The object that fills that slot is whichever object (if any) is in the relevant causal relation to that experience. A different experience with the same content can have a different object or even no object at all-not because the content is a general, existential proposition but because the content is schematic. It is filled in by the object of the experience.Notice that IVE does not include any information about how the object of the experience is determined. The fact that the psychological state in question is a visual experience makes it the case that its object is whatever is in the appropriate causal relation to that experience.

I wonder if the above is not what Searle really believes. In a very revealing passage he writes, "the sense in which the visual experience is self-referential is simply that it figures in its own conditions of satisfaction. The visual experience itself does not say this but shows it" (1983, 49). He seems here to be holding on to the claim that the content of the experience determines its condition of satisfaction while at the same time while recognizing that there is nothing in the content of experience that is equipped to do the job. Indeed, he has since conceded that it may have been inadequate to describe the self-referentiality which determines an experience's conditions of satisfaction as "shown but not stated by the visual experience," and suggests that it might "have been better to put the point without using the notion of reference, but e.g. that of token-reflexivity" (1991, 238).

In my view, an experience is token-reflexive not because it refers to itself but because, being perceptual, its content is context-sensitive. The condition of satisfaction is experience-relative. To experience a certain event or state of affairs is to experience it from a certain point of view. That is, you experience it in a spatio-temporal relation to your point of view. But though you experience it in an experience-relative way, you do not experience it under an experience-relative description. It is by having the experience at a certain time and place, with a certain orientation, that the time and place of the event or state of affairs experienced can be represented as being when and where it is. But you do not have to represent the time and place of the experience in relation to which you represent the time and place of the event or state of affairs. Your experience is essentially perspectival.

On the indexical view, the experience (qua type) is the indexical. Even if you are aware of this pen as the cause of your experience (token), or even if, in what amounts to the same thing, you are aware of it as the thing that looks thus-and-so to you now, it is not the fact that you experience it as such that makes it the object of your experience. It is not the object of your experience because it satisfies a condition determined by the content of your experience. It does satisfy the description, 'the object of this visual experience,' but it does so independently of there being anything in the experience that imposes this condition. The object of the experience is, as I like to say, determined, not "satisfactionally" but "relationally" (Bach 1987, 12).

By treating experiences as themselves indexicals, we avoid the difficulty that McDowell raises for Searle's account, namely, of locating the particularity in the wrong place. Why, McDowell asks, does Searle allow a element of content of the form 'this experience' but not one of the form 'this pen' (1991, 218)? If an experience can refer to itself, why can't it reach out to the world? After all, Searle has always been a direct realist. He rightly thinks that we see pens and pigs, not sense-data of pens and pigs (and that we see pens and pigs not by way of seeing sense-data of pens and pigs). So he should not insist that an experience can "refer" to itself but not to this pen. Saying that the experience refers to the pen, i.e., that it is directed at the pen, does not imply the pen is part of the content of the experience-it is the object of the experience.

Nor is it implied that the content of the experience is a singular proposition. After all, as Searle says, experience (and thought, for that matter) is always under an aspect. Visual experience, for example, is always from a point of view, from which certain portions of an object are visible and not others, and selective, insofar as only certain of the object's properties figure in how it appears. But it is the experience itself that provides this perspective on the object.

On the indexical view you can be aware of this pen and be in a position to refer to it as "this pen" just because it is in the right relationship to you. What makes your experience an experience of this pen is not that it meets a condition laid down by the content of your experience but that it enters into a certain causal relationship to your experience (such that its features affect what you experience and changes in its features or in which features are accessible to you differentially affect what you experience). Searle's descriptivist view gets the "direction of fit" backwards (if I may borrow one of his favorite phrases) between the object and the content of the experience. The object does not have to satisfy any element of the content in order to be the object-it is the object because of its causal relation to the experience.

5. Afterthoughts

I want to finish up with a brief comment on Alston's dispute with Searle, followed by a speculative hypothesis about the dependence of thought on experience.

In Searle's view, "the content of [a] visual experience, like the content of [a] belief, is always equivalent to a whole proposition. Visual experience is never simply of an object but rather it must always be that such and such" (1983, 40). I have argued that the content of a visual experience is equivalent only to an incomplete proposition, which is completed when its objectival slot is filled by the object causing the experience. Alston disagrees with Searle for quite a different reason. He maintains that visual experience is a matter of "direct, unmediated awareness of … objects," that objects "are simply 'displayed' or 'presented to awareness." It seems to me that Searle agrees with this. After all, he grants that visual experiences have objects and that these are physical objects, not private objects like sense-data. And yet he holds that "the whole content [of an experience] must be propositional (even though phenomenologically it seems a matter of just perceiving an object)" (1991, 235). I think Searle's view can be reconciled with Alston's if, first, instead of saying that "visual experience is never simply of an object but rather it must always be that such and such," Searle were to say that it must always be of an object experienced as such and such. In other words, seeing is always seeing-as, not seeing-that. Secondly, except when explicitly stating his theory, Searle could avoid the cumbersome language of causal self-referentiality and just say that the content of an experience is, e.g., "this yellow station wagon looks old." He could do this by arguing that the causal self-referentiality is built into the word 'looks.' After all, to say that this station wagon looks old is just another way of saying that one sees a certain thing as an old station wagon.

I would like to end on a friendly note by offering a hypothesis that indirectly supports Searle's (1992) controversial "Connection Principle," which says that every mental state either is conscious or is at least potentially or in principle conscious. I won't try to defend the CP directly, though it would be helpful to try to pin down the force of the expressions 'potentially' and 'in principle' so that the CP does not turn out to be trivially true. Presumably Searle does not want to count as mental a state involved in visual edge detection or a state involved in the phonological parsing of a heard sentence. The problem is to explain in precisely what sense such states are not potentially or in principle conscious.

Anyway, let me tentatively offer a different though related principle. I call it "Experience First." It says that only beings that enjoy phenomenal states are capable of being in Intentional states at all, as opposed to mere informational states. In other words, only beings with perception and sensation are capable of thought. If this is so, there can be no such thing as a pure thinking machine. Nothing qualifies as a thinking thing unless it qualifies as a sentient one. A robot, no matter how sophisticated its sensors and its processing of the information received from them, does not qualify as a thinking thing, a conscious being, unless it has experiences.

The rationale for Experience First is this: you can't have general thoughts unless you have singular thoughts, and having singular thoughts depends on being in perceptual relations with particular things. Moreover, thoughts about things involve concepts of physical properties and having such concepts requires having experiential or recognitional concepts. Having such first-person concepts requires having, or having had, perceptual experiences. In short, no concepts without percepts. Sound familiar?

* * * * * *

To conclude, I have argued that Searle's account of the content of visual experience is, despite its element of causal self-referentiality, essentially a descriptivist theory. As such, it does not solve the problem of particularity. To solve that problem I have suggested an indexicalist alternative to the various de re account that Searle criticizes. No doubt he will criticize this account too, but, as he says in reply to his other critics, "I do not know of a demonstrative argument to show that visual perception includes a causal component in such a way that the analysis of the content must contain the causal self-referentiality that I allude to. … In the end, perhaps, it is one of those points in philosophy where you either see it the way I do or you don't" (1991, 236).

References

Alston, William (1997), "Searle on Perception," manuscript.

Bach, Kent (1978), "A Representational Theory of Action," Philosophical Studies 34:361-79.

_____ (1982), "De Re Belief and Methodological Solipsism," in Andrew Woodfield, ed.,

Thought and Object. Oxford: Oxford University Press.

_____ (1987), Thought and Reference, Oxford: Oxford University Press.

Burge, Tyler (1977), "De Re Belief," Journal of Philosophy 74:338-62.

_____ (1991), "Vision and Intentional Content," in LePore and Van Gulick.

Evans, Gareth (1982), Varieties of Reference, Oxford: Oxford University Press.

LePore, Ernest and Robert Van Gulick, eds. (1991), John Searle and his Critics, Oxford: Blackwell.

McDowell, John (1984), 'De re senses,' in Crispin Wright, ed., Frege: Tradition and Influence, Oxford: Basil Blackwell.

_____ (1991), "Intentionality de re," in LePore and Van Gulick.

Searle, John (1983), Intentionality. An Essay in the Philosophy of Mind, Cambridge, Eng.: Cambridge

_____ (1991), "Response: Reference and Intentionality," in LePore and Van Gulick.

_____ (1992), The Rediscovery of the Mind, Cambridge, Mass.: MIT Press.

Valberg, J. J. (1992), The Puzzle of Experience, Oxford: Oxford University Press.University Press.