17
Hallucinations and the Transparency of Perception
Abstract
This paper examines the way in which concepts of a low-level classificatory kind occur in different kinds of experiences, and what happens when subjects of deceptive musical hallucinations reassess their experiences and come to realize that they are hallucinating.
Drawing on this account, it is shown how it is possible for subjects to adopt different conceptual stances with respect to veridical perceptual experience. The issue of the seeming “transparency” of perceptual experience is explored, and it is argued that the sense in which perceptual experience is transparent is compatible with the critical realist version of the causal theory of perception: hallucinatory phenomena show how perception can be intentionally direct, yet causally mediated. Transparency does not provide any special support for direct realist or disjunctivist theories of perception.
In this chapter, I argue that an examination of some important features of hallucinations throws light on the phenomenon of transparency and shows what is right and what is wrong with the idea that we see and hear physical objects directly. According to the direct realist theory, in veridical perception the subject stands in a relation of direct awareness to a mind-independent physical object. Like many descriptions of views in this area of philosophy, this brief summary needs to be further elaborated and explained to be made completely clear. Direct realist theories of perception can take various forms. For the purposes of this paper, what matters is that those theories concur in denying that veridical and hallucinatory experiences form a common metaphysical kind: the experiences I have when I perceive something veridically do not share a common content with those that occur when I hallucinate. Many versions of disjunctivism uphold this negative claim.1
When I hallucinate, I am aware of phenomenal qualities present to my consciousness; yet there is no outer physical object of the relevant kind to which such qualities relate. A plausible view is that in hallucinating, I have an inner experience that I can reflect on and examine. This is what enables me to distinguish a visual hallucination of something red from one of something blue, and to distinguish both experiences from a hallucination of the sound of a piano. However, according to direct realism, normal perception involves a quite different notion of experience. One reason often cited in current work on perception for preferring the direct realist position to alternative views is the thought that perception is “transparent.” When I veridically perceive some physical object, my experience of it is direct. If I try to introspect my veridical experience, I am not aware of any mediating inner state or entity; all I seem to be aware of is the external mind-independent object I perceive. If I look at an apple in normal conditions, I see the apple directly, and I am not aware of any inner state that mediates my perception of it. I also find that when I try to examine the nature of my own experience, I find nothing further present to consciousness other than the very apple I see out there in the world. Hallucinatory and veridical experiences should be analyzed very differently, and they share no common experiential content.2 According to this view, it is difficult to understand how normal perceptual experiences can be inner states.
In the alternative causal theory of perception, veridical and hallucinatory experiences do belong to the same ontological kind. Both are inner states of some (perhaps complex) kind and differ only in their causal ancestry. One argument for the inner experience view is the causal-scientific argument advanced by Jerry Valberg (1992) and also by Howard Robinson (1994). There is a second, related argument: the direct realist is hard put to provide a coherent positive account of the relation of direct acquaintance that is assumed, on the theory, to hold between my experience and the specific object I perceive in having my experience.3 For such reasons, the causal theorist holds that veridical perception should be analyzed by reference to the causal relation (of the appropriate kind) that holds between the subject and the specific external object perceived. When I see an apple, the experience I have is distinct from the actual apple I see, although the external physical object normally forms the focus of my perceptual awareness.
If perceptual experiences are regarded as inner states, there is a prima facie problem for the causal theorist in responding to the transparency objection. If the transparency argument—which I will elaborate in more detail subsequently—is correct in holding that introspection of a veridical experience cannot reveal anything distinct from the outer physical object perceived, then it is difficult to understand how experiences could be inner states. I begin by summarizing, without argument, the critical realist version of the causal theory, so as to provide a framework for assessing the claims of transparency. Critical realism derives from the writings of Wilfrid Sellars, and I have defended it against a number of commonly advanced objections in other work.4 The main vindication of the theory in the present paper is that it provides a way of reconciling the causal analysis of perception with facts about the phenomenology of perceptual experience. If direct realist theories are to seek support, they need to do so without appealing to claims arising from our intuitions about the transparency of perception. The critical realist version of the causal theory is not in any way undermined by the phenomenon.
2 The Critical Realist Model of Perceptual Experience
According to the critical realist account, in perception two distinct components combine to form an inner experience. When I see an object such as an apple, what makes my state of consciousness a visual or auditory episode is the distinctive sensory component in virtue of which I am aware of a range of phenomenal qualities. But in addition, perception has a cognitive aspect. Usually when I perceive an object of a familiar kind in normal conditions, I directly recognize it for what it is, and classify it by employing some (perhaps low-level) concept or category. I am only conscious of seeing anything because I am aware of what I see as of a certain kind. For experiences to be reasons for perceptual beliefs, the experiences themselves must contain classificatory concepts of some low-level kind. Even if I do not recognize the object in front of me to be an apple, I am only perceptually conscious of it because I see it as some sort of reddish, round object, and so on; necessarily, I exercise concepts in experience. I use the term “concept” and its cognates here to cover the exercise of the kind of ability that Peacocke associates with proto-propositional content.5 The exercise of low-level concepts by a subject does not imply that the subject must be capable of self-awareness or have other higher-level cognitive capacities.
The distinctive claim of critical realism is that in normal circumstances, the concepts I exercise in perception relate directly to the physical objects or events in my surroundings. They are caused by my inner phenomenal states but do not focus on them. Normally, in seeing a friendly black retriever dog bounding toward me, I am not concerned with my subjective experience of the dog, or which surfaces it presents from my viewpoint; my concepts refer to how the dog is objectively. I take it to be a unified, solid, independent thing moving in front of me. Sellars refers to the conceptual component of experience as the subject’s “perceptual taking.” Since I do not, in the usual case, form concepts referring to the phenomenal aspect of my experience, there is no process of inference to the exercise of concepts in the perceptual taking: the connection is causal.
The low-level conceptual element in experience accounts for the phenomena of “seeing as” and “hearing as.” In looking at the dog, I see it directly as a dog; seated by a swimming pool, I directly hear a sound as the splash of someone jumping into the water. The concepts exercised in seeing as and hearing as are integrated with the phenomenal aspect of experience through the exercise of the imagination, as Kant, Sellars, and Strawson have argued: in seeing the colored shape in front of me as a person, or as a dog, and in hearing a sound as a musical instrument, or as someone talking, I am implicitly prepared for different future possible patterns of experience.6 The concepts I exercise in this way relate to natural kinds and individuals that I am able to recognize directly, without inference. They also apply to broad categories of things. Even if I do not know which specific class a perceptual object belongs to, I may still see it as an animal, or a tree, and so on.
One important point, which will be relevant to the argument later on, is that the concepts exercised automatically in perception are, in part, determined by the subject’s general knowledge and awareness of the overall context. Depending on my beliefs about the circumstances, I may spontaneously categorize what I see at a high or low level. The untutored observer may see only a line of condensation in a laboratory cloud chamber; the trained physicist looking at the same apparatus may immediately take what he or she sees to be an alpha ray. In undergoing a hearing test, I may directly classify some sound as of a clarinet, or I may simply classify what I hear as the occurrence of a faint sound.7 Various different factors will come into play, and they all influence what is perceived. The concepts directly exercised in experience will result from the subject’s general prior learning, background beliefs, and attentional set, as these interact with a nonconceptual awareness of phenomenal qualities caused by sensory input. A subject such as a painter, who adopts some nonstandard attentional set, may come to conceptualize experience in very different ways from normal.
Both phenomenal and low-level conceptual states are necessary for perceptual experience. However, in view of the complex origins of our perceptual takings, it is plausible to argue that the two components of experience have a degree of independence from each other. Depending on my background beliefs, and so on, the same phenomenal state could lead to my exercising different sets of concepts in responding perceptually to my surroundings. So the same physical input might lead me on one occasion to see the object on a nearby chair as a black sweater, and on another occasion as my cat. Similarly, I might hear the same faint hissing sound either as the background in an old musical recording, or as a garden spray, or as the frying of onions. The idea that some slippage between phenomenal and conceptual aspects of experience is coherent I dub “the logical independence thesis.”
Writers such as McDowell argue that we cannot make sense of differences at the preconceptual level—in “receptivity”—that are independent of our concepts.8 The assumption is that our phenomenal states are solely identified by reference to the concepts and beliefs that the perceiver entertains. But to argue in this way is to ignore the contribution of the role played by the overall physical surroundings of the perceiver to our concept of experience when we make attributions about a subject’s phenomenal experience from an external perspective. External factors, which include the precise nature of the sensory input from the objects perceived, contribute to the identification of what the subject is aware of in the phenomenal sense, and thus support claims about unattended complexity at the phenomenal level. Sperling’s experiments on the selective recall of multiple figure displays and change-blindness experiments can be cited as further evidence for the independence thesis, as philosophers such as Dretske and psychologists such as Rensink have argued.9 The logical independence thesis is not advanced on the spurious basis of a comparison between our concepts and what is supposed to be revealed in experience by direct inspection. The idea of such a comparison would not be coherent. Instead the thesis gains its credence because it is the best way of making sense of a whole variety of different perceptual phenomena. It is also consistent with the very fact that we come to be able to refer to other people’s inner states on the basis of an understanding of the input to their perceptual systems, as well as on the basis of their subsequent verbal and other behavior.
3 Reconceptualizing Experience
For the purposes of this paper, what is important is the nature of the process that takes place when the same underlying phenomenal state is reconceptualized by the subject. Consider again the example of my seeing an object in the corner of a room as a black sweater on a chair. By attending more to what I am looking at, I may come to reconceptualize my visual experience. Or without concentrating harder, I may come to notice a faint purring sound, or remember that my sweater is actually upstairs. As a result, a phenomenological change occurs in my visual experience. I shift from seeing the object as a sweater to seeing it as a cat.
In reconceiving the object on the chair, I am considering it objectively. I am in part causally guided by how the object appears, in the subjective sense, in my consciousness. But I normally pay no attention to how things seem to me subjectively; a fortiori, I am not conceptualizing the phenomenal experience as such. My focus is on how things are, objectively, in my surroundings. What matters is how I should classify the independent physical object I am seeing. It is true that a phenomenological change accompanies my reconceptualization of what I take there to be on the chair; but that is adequately accounted for by the shift in my perceptual takings, together with changes in the implicit expectations I form about the future behavior of the objects I perceive.10 In considering the sweater or cat as having objective existence, I understand it to be an entity on which various different perspectives are possible, apart from my own.
Let me give another example, this time from the modality of hearing. I hear what I take to be the sound of loud noise emitted by a distant piece of machinery, such as an electric saw. I then realize that what I hear is the nearby whine of a relatively quiet mosquito, close to my ear. I do not change the concepts I apply because I make an inference from a prior state of pure unconceptualized phenomenal awareness. Such a view would not be coherent, as Sellars pointed out in criticizing what he termed “the myth of the given.” Instead an awareness of the overall context, perhaps accompanied by a slight variation in the phenomenal character of the experience, directly produces a new perceptual taking. The actual process of reconceptualizing experience is essentially causal. I am prompted to reconceive the objective nature of what is physically present in my surroundings.
One need not deny that intellectual processes can sometimes be involved in leading to a change in the concepts I apply. From my belief that a cat is purring, I infer that there is a cat in the room, and this process is part of what causes me to see what is on the chair as a cat. But then the fact that I become aware of the sound I hear as the purring of a cat needs to be accounted for; I am not first aware of a neutral sound from which I infer the presence of a cat. Nor, when I come to see the shape on the chair as a cat, do I first form a nonconceptual awareness of a purely phenomenal state of some kind and reason from that awareness to apply the concept of a cat to what I see. When I hear the sound as that of a mosquito, my experience becomes classifiable because I am causally prompted to exercise concepts integrated with the phenomenal qualities of a high-pitched whining noise I immediately experience. As Wittgenstein emphasizes in his discussion of seeing as, when I directly perceive something as of a certain kind, a concept forces itself on me. This essentially causal process is clearly demonstrated by the striking phenomena investigated by Johansson, in his examples of the awareness of kinds of physical movement based on bare perception of moving spotlights attached to hidden human figures. Subjects who were shown a restricted visual display from spotlights attached to the limbs of moving persons would immediately and spontaneously see the pattern of moving spotlights as people walking, running, and dancing, and so on, despite being unable to see the shapes of the human figures supporting the lights.11
This model of the way that we can reconceptualize our experiences has applications to hallucinations. Recent surveys have provided evidence that full-blown, complex visual and auditory hallucinations are not uncommon. These arise most often, but not exclusively, in subjects who have suffered some kind of sensory deprivation, or brain damage triggered by a stroke.12
According to Oliver Sacks:
In most cases of musical hallucination, there is sudden onset of symptoms, then the hallucinatory repertoire expands, becoming louder, more insistent, more intrusive . . .
. . . The experience for someone with musical hallucinations is not mere imagery, but often physically loud, as-if-heard “actual” music. (Sacks, 2007, 73)
Patients who suffer frequent musical hallucinations often use expressions such as “my inner radio” or “my inner iPod” to label them.
One case described by Oliver Sacks concerns a Mrs. C who had been suffering progressive deafness and had recently been placed on medication. She was awakened suddenly one night by what she described as “dreadful noises . . . like trolley cars, bells clanging.” Thinking she was hearing a fire engine, she rushed over to the window but, looking out, saw that the street was empty. She then realized that the noise was “in her head” and she was hallucinating. After about an hour, the clanging was replaced by tunes from The Sound of Music (and from other sources). “I was well aware that there was no orchestra playing, that it was me,” she emphasized.
Sometimes, however, subjects do not realize immediately that they are hallucinating. In another case described by Sacks, the subject thought for about two weeks that she was hearing the noise of a neighbor’s tape recorder and only gradually came to believe that the music might somehow be in her mind.13
When subjects realize that their hallucinations have been deceiving them, the phenomenal aspect of the experience remains largely unchanged, but the way they conceptualize their experiences alters. Subjects of auditory hallucinations realize that the noises are independent of the environment, and that the sounds they experience are in some way dependent on themselves. What is initially taken to be an objective aspect of the environment is then recognized as something subjective. It is this shift in the conceptualization of a hallucinatory sound that I want to focus on.
Let me first say something about how we normally hear the sounds we take to come from physical objects or events in our surroundings. Our conception of sound, as Strawson argued, is essentially linked with our concept of space.14 To think of a sound as an objective, mind-independent feature of the world is to think of it as having a position, but aside from this there is some latitude in the categories we apply to sound. A sound particular may be considered as an event located in some roughly determinate region of space. Examples would be the sound from a piano on a stage or a roll of thunder down a valley. In this sense, if two musicians play the same D minor chord on two different pianos in different rooms, there are two sound particulars, two sound events belonging to the same general type, in the different rooms. Alternatively, sounds might be considered as properties of objects; we can speak of the sound, the look, and the feel of a bell as all different properties of the same individual thing. In this sense, the sound can be considered as instantiating an identifiable causal source and occupying some fairly specific position.
The differences between these various conceptions of sounds are interesting in their own right, but they do not affect the argument of my paper. What matters here is the essential connection between sound and space. A sound or sound source, considered as a feature of the objective world, is conceived as locatable in the physical space surrounding the perceiver. As well as being loud or soft, sounds can be near or distant; they can be clear or muffled. If conceived as spread out through space, they have an approximate center. Sounds can also have a loose identity through time: I can speak of the continuing noise of a running tap, the same distant music that I hear when a door is intermittently opened and closed, and so on. In this loose sense, I can recognize a sound as “the same again” as it goes in and out of my experience. In doing so, I may be attributing a continuing property to the physical source of the sound.
What is important is the claim that, in conceiving of sound in this way, we think of sounds as distinct from ourselves. They are physical events of some kind, features of the external world, which exist independently of our experience. Perceptual language is dominated by visual terms, yet we can still apply auditory analogues of a perspective, in thinking of an objective sound as something on which we can have an auditory “point of view.” When I think of a sound as an objective physical phenomenon, I take it to be at a certain position relative to my body.
A further feature of sounds needs to be accounted for. In understanding what it is for me to hear a certain sound, I also conceive of it as somehow related to my consciousness. I distinguish my own experience of that sound from the experiences other people might have of it. I can understand what it is for a sound to continue while something prevents me from hearing it, so that it exists while no longer in my experience. Other people might hear the same sound when it is not in my experience, because I am too distant from the sound’s location. A complete description of our conception of sound must account for this further feature and explain how the same sound can somehow be part of different centers of consciousness that are necessarily distinct from each other.
5 The Recognition of Hallucination
These remarks about our conceptualization of sound and about the reconceptualization of experience can help to explain what happens when a subject begins to realize that the musical sounds she hears are not real but hallucinatory. The process involves a shift in the concepts exercised by the subject and comprises three essential stages.
Stage 1: Deceptive Phase of Hallucination
Initially the subject hears the sound of melodies being played and takes her experiences to be veridical. For a subject like Mrs. C to think of the music as objective involves taking the sound to be something mind independent. It is taken to involve some external physical object such as a neighbor’s tape recorder, and to have an approximate spatial position—coming from across the hallway. The experience of hearing the sound as louder or softer is assumed, for example, to be the result of the neighbor moving the tape recorder around or turning the volume up or down. The source is taken to be something that can physically be acted on, which changes independently of the subject’s attention. At this stage, the music is heard as an externally produced sound. In a manner that is analogous to sight, the sound is conceived to be something on which the subject has what I shall describe as an auditory “perspective”: how the sound is experienced is assumed to depend in part on the relative spatial position of subject and sound source. The focus of the subject’s conceptualization is the assumed mind-independent sound event, not the subject’s own experience of the sound.
Stage 2: Transition
When subjects experience sustained hallucinations of musical tunes, they gradually come to realize that none of the normal connections obtain between their bodily movements and their auditory experiences. They also find that their reports of the music fail to tally with those made by others. The tunes apparently heard can no longer be attributed to any independent external source, yet the sound experience still pervades consciousness. As the subject realizes that the tunes she hears are not based on any external spatially located source, she begins to abandon the conception of the sound as something objective and independent from herself. There is no longer a sense that it is possible to obtain a different perspective on the source of the sound.
The subject therefore ceases to form perceptual takings that refer to some external, mind-independent sound event. Despite this, the subject is still able to identify the types of sounds she seems to hear by reference to the kinds of instruments (or other objects) that would normally cause the experiences she has.
Stage 3: Realization of Hallucination
When the subject comes to accept that she is not hearing any objective sound, she recognizes that the sound is not independent of her consciousness; it is not something that she can change her auditory perspective on. There are no hidden aspects to the phenomenal sounds that are immediately present in consciousness; there is nothing else to the sound except what is present in consciousness. Nothing she does by direct action on physical objects can alter the awareness of phenomenal sounds—only the indirect action by taking drugs, or the change of attention through an effort of concentration, can have any effect on her overall conscious experience as of musical sounds.
The sound experience needs to be reassessed, but in the hallucinatory case no external object or feature is available in terms of which the sounds can be reconceptualized. This means that the framework of objective sounds no longer has any application to experience. Yet the hallucinatory sounds still exist as a feature of consciousness. So in this case the subject switches the focus of her attention and conceptualization to what is immediately present in consciousness: there is a change not only of concepts but also of focus from the outer world to the inner subjective states of the self. The subject can attend to the intrinsic character of the phenomenal qualities in her experience—the intrinsic qualities of pitch, volume, and timbre.
But to attend in this way to aspects of her own experience, the subject has, necessarily, to employ concepts. Normal subjects will already have a range of concepts relating to inner states; they can attend to and think of their own sensations and other mental states and episodes. Hence they are in a position to recategorize the sound experiences as a states of themselves, as in some way belonging to the same general category as other kinds of inner states, such as pains and other sensations (both hallucinatory music and pains, etc., are conceived of as intrusive and annoying). Hallucinatory states such as sounds (and also other forms of hallucination, such as migraine-related visual aura) therefore become conceptually absorbed into the general category of inner conscious states and episodes: dependent items essentially tied to the consciousness of a single subject, such that no alternative perspective is available on them.
Note that even if a subject employs concepts to pick out elements of her experience in a “direct” fashion, as advocates of the phenomenal-concepts strategy advocate, the concept so exercised will still need to locate such elements in the appropriate conceptual space: the concept has to specify the sortal, and more general categorial features, of the phenomenal aspect of experience that is attended to by the subject.
None of these claims is intended to deny that there are interconnections between concepts relating to the inner and the outer. But such connections are complex. There are familiar arguments for the claim that for inner sensations to be identified requires that there be some outer criterion for such inner states. This claim is consistent with the thought that, in a specific case, a token of some inner state can occur, and be directly identified by the subject, in the absence of awareness of any external feature of the kind normally used as the criterion.15
There is therefore a significant shift in the concepts exercised by the subject when she realizes that a certain note or tune that she hears repeated over and over again is in fact a hallucination. She becomes unable to conceive of the sounds she hears as having objective features and an independent existence. Despite this, she is still left with the awareness of something that she is able to conceptualize in quite a different way. The concepts she starts to exercise relate to her own experiences and their intrinsic qualities. The subject may lack the sophisticated terminology of cognitive scientists, but she still thinks of the musical sounds as her own and conceives of how things are in her experience.
When the subject hears the same irritating tune begin again—for perhaps the twentieth time in a row—she will directly conceptualize it as her own subjective state, and not as anything independent. That subjects speak of their hallucinations in terms such as “my inner radio,” and so on, reflects this understanding. Such experiences are conceived of as inner states that can be attentively examined, or else, through an effort of will, ignored, or even, in some cases, controlled: some subjects, for example, are reported to be able to change the hallucinatory tunes they hear, even when they are unable to make the sounds cease altogether.16
6 Hallucinations and the Transparency of Experience
As I noted at the outset, many of those who advocate direct realist theories of perception appeal to claims about the nature of perceptual experience in support of their view. Some philosophers assume direct realist views to be superior to causal theories because they believe only direct realist theories can do justice to the transparency of normal veridical experience. The analysis of what goes on when subjects realize that they are hallucinating can be used to clarify the phenomenology of ordinary perception and to help us assess what is right and what is wrong with such claims.
The notion of transparency involves two linked intuitions about perceptual experience. There is, first, a claim about the directness of experience: the perception of ordinary things is unmediated, so that when I perceive the world, I am aware of the physical objects I see and hear directly, without inference. Nothing intervenes between me and the objects I see and hear. Second, there is a claim about the contents of experience: when I turn my attention inward, I don’t find any extra entities in my consciousness, in addition to the very objects I perceive to be in the external world; there is no change at the phenomenal level in my experience, no difference in the qualities I am immediately aware of.17 Martin summarizes these two claims in stating:
Introspection of one’s perceptual experience reveals only the mind-independent objects, qualities and relations that one learns about through perception. (Martin, 2002, 378)
But the account of deceptive hallucination sketched in the previous section suggests that introspection is a more complex affair, and not as straightforward as many writers suppose. A comparison between veridical experience and cases of deceptive hallucination shows how it is possible to do justice to the foregoing intuitions while still analyzing perception as involving inner experiences.
When I am deceived about what is going on in having an experience that I do not realize is hallucinatory, I take there to be some external physical object in the surroundings that I am seeing or hearing directly, such as a radio playing loudly across the hallway. My concepts focus on the physical objects or events I think I am perceiving; they are guided by the inner phenomenal states that form part of my hallucination. But my concepts do not refer to inner experiences I have; instead they refer to outer objects that I take to exist around me. Yet even after I discover that I am hallucinating, there is still something going on in my consciousness. It may be that even if I no longer believe there is a radio nearby, I retain an inclination to think there is one. More importantly, as well as any inclination that persists, I also have an awareness of phenomenal qualities of some kind—certain tonal qualities that have a definite volume and pitch. Because I am aware of the subjective sound continuing, I am prompted to attend to my own inner phenomenal state, and I recognize it for what it is, a state of myself. When, in this manner, I introspect my hallucinatory experience, what happens is that I reconceptualize it and focus on it directly in the manner described earlier.
Hallucinatory experiences can be indiscriminable from veridical ones, as the clinical evidence cited earlier clearly shows. If the causal theory of perception is correct, this is because they both involve inner experiences of the same ontological kind. But if perception can be analyzed causally in this way, then it is obvious that the same process of reconceptualizing experience is also available to the subject in the veridical case. That is, it is open to the subject to take up two different conceptual stances in relation to the veridical visual and auditory experiences she has. She can either respond to her experiences in a perceptual mode, by forming concepts relating to the objective features of the world that she assumes belong in her surroundings, or she can respond in the very different mode of introspection.
When responding in the perceptual mode, the subject does not exercise concepts that refer to what is immediately present in inner experience, as we have noted. Hence she is not conceptually aware of any entity intervening between her and the things she sees or hears. The concepts exercised spontaneously in experience refer directly to the type of outer object assumed to be present—an apple, a dog, a piano, someone talking, and so on. There is no process of inference; the exercise of concepts is causally guided by the phenomenal aspect of experience. This phenomenal aspect of experience is not, in the ordinary case, the object of the perception. The perceptual object is some physical thing (or event) that the perceiver takes to be out there in the external world. So we can respect the first intuition of transparency, that perception is a direct process. It is cognitively direct, because the concepts employed apply only to the external objects perceived, objects taken to be immediately present.
No concepts are normally applied directly to our inner experiences in veridical perception. The concepts exercised in veridical experience relate to external objects and situations that transcend the phenomenal states given immediately in experience, and are guided by such states, just as they are in the deceptive hallucinatory case. But those phenomenal states exist and are immediately present in consciousness. According to the causal theory, they are inner subjective states that can be conceptualized in a direct fashion, if the subject is prompted to reflect on his or her phenomenal consciousness. I am able to think about my experience for what it is, a subjective state of mine, containing phenomenal qualities that normally prompt perceptual thoughts focused directly on the external world. I might, for example, as a consequence of some head injury suffer double vision. I would then have a reason for attending to and conceptualizing my own visual experience as an inner state that only corresponds to the external world in a distorted manner. So it is open to the subject to switch conceptual stances even in the perceptual case, and to adopt different ways of conceptually responding to what is immediately present at the phenomenal level.
Because just the same continuing phenomenal state is involved, it is correct to say that when I introspect my perceptual experience of hearing a sound, I don’t discover anything new; the phenomenal state was a part of my conscious experience all along. No change in my experience occurs at the phenomenal level when I introspect. I do not become immediately aware of any additional phenomenal qualities. Hence there is some truth to the second intuition of transparency. There is no new entity, no further qualities waiting to be noticed.
Yet the second intuition is also in an important sense misleading. There is a significant difference in the concepts I exercise when I introspect: at the cognitive level, I can become aware of something I did not appreciate before, because I can now alter the concepts I exercise with respect to experience. I am able to form additional beliefs about the full nature of what is going on in perception, about how things are subjectively. In the hallucinatory sound case, this reconceptualization is forced on the subject when she discovers that concepts applicable to objective events have no external application. She then finds she has a range of different concepts relating to her own inner states, which she can coherently apply to make sense of her continuing experience of sound.
But because of the subjective indiscriminability of hallucinatory and veridical experiences, a parallel possibility of reconceptualizing one’s experience is fully available to the subject in the veridical case. It is perfectly coherent to envisage an alteration in the way that one’s normal veridical experience of objects is conceptualized. I am able to reflect on my inner experience qua experience, rather than focus attention on the objective world around me. I can think of that experience as an inner state. Whereas in the hallucinatory case the change in conceptualization is forced on the subject by the circumstances, in normal perception the change would be an additional option open to the subject, so that upon reflection the subject can come to understand more about the nature of her inner experience. (It is always possible, of course, that the subject might have theoretical motives for resisting the opportunity to reflect in this way.)
So the idea is that even in normal veridical perception, subjects can come to have additional thoughts about the way that an object is perceived, relating to their own subjective experiences of perceiving. The conceptual framework for considering subjective experiences themselves is to be contrasted with the framework we employ when we consider the objective world about us.
Thus suppose that I and my companion listen to the same musical concert performance. I can conceptualize my experience in different ways. First, I am able to reflect on the objective performance as a musical event I hear that is going on independently of me. But I may also become aware that, in an important sense, my experience of hearing the music is distinct from the experience of my companion. This need not be just because I am considering differences at the emotional or cognitive level. I can also be aware that, at the level of phenomenal consciousness, my subjective experience of hearing the sounds is distinct from my companion’s experience. Both subjective experiences are different events from the objective events that constitute the musical performance itself.
I don’t find anything new at the phenomenal level when I reflect in this way. But I become newly aware that, in an important sense, the fact of my having a phenomenal experience is a further fact about the overall situation, in addition to the objective facts about the musical performance. According to the causal theory of perception, this further fact is a fact about an inner experience.
The upshot is that a careful consideration of the complex nature of hallucinations can provide us with some insight into the nature of veridical perception. It indicates that the claim that veridical experiences involve inner states is not in conflict with the phenomenology of perception. The inner-state thesis respects the two intuitions relating to claims about the transparency of perceptual experience noted at the start of this section: perception is direct, and introspection reveals no new entity in experience. That perceptual experience is in some sense transparent does not rule out the causal theorist’s claim that perception involves inner experiences.
I have not tried to show here why direct realism is mistaken, and why we should embrace a critical realist version of the causal theory of perception as a more plausible alternative. Rather, I have sought to defuse a commonly advanced objection to the causal theory and to show how the analysis of perception as involving inner experiences makes clear sense. To hold that experiences are distinct inner states and the bearers of phenomenal qualities is not to imply that perception is indirect. My perception of the world is direct, because nothing at the conceptual level comes between my experiences and the objects I perceive. As the examples of deceptive hallucinations demonstrate, inner phenomenal states can cause and guide our perceptual takings, and also our expectations about the future course of experience, without themselves necessarily becoming objects that we focus on.
To end on a slightly concessive note, in one sense the disjunctivist is correct to say that there is no highest common factor to all cases of hallucination and veridical perception. To be precise, no experience is common to both the situations where a person perceives and the situations where a person knowingly hallucinates if we understand experience in the inclusive sense, to comprehend the whole of what goes on in consciousness. Experience, understood in this broad sense to include both phenomenal and conceptual components, is not something shareable, because the conceptual aspect differs significantly. The concepts exercised that are essential to being conscious in each case will relate to quite different categories. Nevertheless, if critical realism is correct, the remaining component—the phenomenal aspect of veridical and hallucinatory experience—does indeed belong to a common ontological kind. The transparency of perceptual experience is therefore perfectly compatible with the critical realist claim that the distinctive phenomenal component of experience is an inner state, of the kind that is shared by both hallucinations and perceptions. The critical realist theory of perception, a theory that recognizes that experiences are inner states, and also the key role played by concepts in experiences, is not undermined by any phenomenological considerations relating to transparency.
References
Campbell, J. (2002). Reference and consciousness. Oxford: Clarendon Press.
Coates, P. (2007). The metaphysics of perception: Wilfrid Sellars, perceptual consciousness, and critical realism. Oxford: Routledge.
Dretske, F. (2006). Perception without awareness. In T. S. Gendler & J. Hawthorne (Eds.), Perceptual experience (pp. 147–180). Oxford: Oxford University Press.
ffytche, D. H., & Howard, R. J. (1999). The perceptual consequences of visual loss: “Positive” pathologies of vision. Brain, 122, 1247–1260. http://dx.doi.org/10.1093/brain/122.7.1247.
Haddock, A., & Macpherson, F. (2008). Introduction: Varieties of disjunctivism. In A. Haddock & F. Macpherson (Eds.), Disjunctivism: Perception, action, knowledge (pp. 1–24). Oxford: Oxford University Press.
Johansson, G. (1973). Visual perception of biological motion and a model for its analysis. Perception and Psychophysics, 14, 201–211.
Johnston, M. (2006). The function of sensory awareness. In T. Gendler & J. Hawthorne (Eds.), Perceptual experience (pp. 260–290). Oxford: Oxford University Press.
Lowe, J. (1995). Locke on human understanding. London: Routledge.
Manford, M., & Andermann, F. (1998). Complex visual hallucinations. Brain, 121, 1819–1840.
Martin, M. (2002). The transparency of experience. Mind and Language, 17, 376–425.
McDowell, J. H. (1986). Singular thought and the extent of inner space. In P. Pettit & J. H. McDowell (Eds.), Subject, thought, and context (pp. 137–168). Oxford: Clarendon Press.
McDowell, J. H. [1994] (1996). Mind and world. Cambridge, MA: Harvard University Press.
Peacocke, C. (1992). A study of concepts. Cambridge, MA: MIT Press.
Rensink, R. (2000). When good observers go bad. Psyche, 6. http://psyche.cs.monash.edu.au/v6/psyche-6-09-rensink.html.
Robinson, H. (1994). Perception. London: Routledge.
Sacks, O. (2007). Musicophilia. London: Picador.
Sellars, W. (1956). Empiricism and the philosophy of mind. In H. Feigl & M. Scriven (Eds.), Minnesota studies in the philosophy of science (Vol. I): The foundations of science and the concepts of psychology and psychoanalysis (pp. 253–329). Minneapolis: University of Minnesota Press.
Sellars, W. (1975). The structure of knowledge: Perception. In H. Casteñeda (Ed.), Action, knowledge, and reality: Critical studies in honor of Wilfrid Sellars (pp. 295–316). Indianapolis: Bobbs-Merrill.
Sellars, W. (1978). The role of imagination in Kant’s theory of experience. In H. W. Johnstone (Ed.), Categories: A colloquium (pp. 231–245). Pennsylvania: Pennsylvania State University.
Sellars, W. (1982). Sensa or sensings: Reflections on the ontology of perception. Philosophical Studies, 41, 83–111.
Smith, A. D. (2002). The problem of perception. Cambridge, MA: Harvard University Press.
Snowdon, P. (1981). Perception, vision, and causation. Proceedings of the Aristotelian Society, 81, 175–192.
Snowdon, P. (1990). The objects of perceptual experience. Proceedings of the Aristotelian Society, 64 (Suppl. Vol.): 121–150.
Strawson, P. F. (1959). Individuals: An essay in descriptive metaphysics. London: Methuen.
Strawson, P. F. (1970). Imagination and perception. In L. Foster & J. W. Swanson (Eds.), Experience and theory (pp. 31–54). Amherst: University of Massachusetts Press.
Tye, M. (2000). Consciousness, color, and content. Cambridge, MA: MIT Press.
Valberg, J. J. (1992). The puzzle of experience. Oxford: Clarendon Press.
1. There is an unfortunate lack of clear agreement over the use of many expressions used in connection with the nature of perception and experience. By the expression “direct realism,” I refer to perceptual theories opposed to the causal theorist’s claim that experiences form a common metaphysical kind. My use of “direct realism” is intended to include disjunctivist accounts. Disjunctivist theories of various forms can be motivated by different considerations, but most disjunctivists accept the “no common content” claim; see, e.g., Snowdon (1981, 1990), McDowell (1986), and Martin (2002); cf. also what Campbell (2002) terms “the relational view” and Smith’s (2002) rejection of indirect realism. On the different forms of disjunctivism, see the useful introduction to Haddock and Macpherson (2008).
2. The transparency argument is widely accepted in current work on perception; see in particular Tye (2000) and Martin (2002).
3. A particular problem for the direct realist view, connected with the second argument, consists in providing a principled basis for distinguishing between the following set of cases: (i) where S sees a given apple X and not a similar apple Y situated nearby; (ii) where S hallucinates as of an apple in the presence of an apple; and (iii) where S is caused, via a nonstandard causal chain from an actual physical apple, to hallucinate an apple. The direct realist’s central notion of an “awareness relation” threatens to become parasitic on a causal analysis of what it is to perceive a particular object, and from the causal theorist’s perspective is an idle posit. I discuss the ramifications of this argument in chapters 3, 4, and 6 of Coates (2007).
4. See Sellars (1956, 1975, 1982) and Coates (2007).
5. See Peacocke (1992) and also Smith (2002, chap. 3).
6. It is because of the implicit expectancies, or preparedness, with respect to the future likely patterns of experience that are set up when I directly exercise a concept that the two components of experience are unified. Thus the critical realist view advocated here should not be confused with what Johnston (2006) criticizes as “the wallpaper view”; see Sellars (1978), Strawson (1970), and Coates (2007, chap. 9).
7. These claims about the comparative richness of the concepts that can be directly applicable in experience are compatible with the thesis that the range of qualities immediately present at the nonconceptual phenomenal level is restricted. At the phenomenal level, it is arguable that my awareness is of inner phenomenal states that have a limited set of counterpart qualities and are analogous to the colors and shapes of the objects I can see, or to the pitch, volume, and timbre of the sounds I hear. Since this claim about counterpart properties does not affect the main thesis of this paper, I shall not pursue it further here. It is defended in Sellars (1956) and elsewhere, and more recently in Lowe (1995).
8. McDowell (1996, 51).
9. See Dretske (2006) and Rensink (2000).
10. See Coates (2007, chap. 9) and also Sellars (1978).
11. Johansson (1973).
12. This is the Charles Bonnet syndrome. For evidence of the relatively widespread incidence of hallucinations in subjects who suffer some kind of perceptual deterioration, see Manford and Andermann (1998) and ffytche and Howard (1999).
13. Sacks (2007), 72, n. 13.
14. Strawson (1959), chap. 2.
15. I should also emphasize that nothing in these claims commits me to the view that inner experiences are objects, in the metaphysical sense advocated by sense-data theorists, which implies a possible independence from the subject. Rather, experiences should be thought of as complex states of the subject’s consciousness, which can be treated adverbially. The only sense in which they are like objects is that they can be the objects of the subject’s self-reflective thoughts and attention; see Coates (2007, chap. 9).
16. Sacks (2007, 65, 75).
17. For a more detailed discussion of these two intuitions, see Coates (2007, chap. 8).