Philosophy of Perception Research Paper

1. The Perceiver–World Relation

1.1 Direct Realism, Indirect Realism, And Phenomenalism

As I look out of the window, I see objects, such as buildings, and I see their properties, such as the green color of their roofs. Presumably, the objects and their properties continue to exist even when I am not looking at them. (I can also be said to see events, such as the passing of pedestrians along the sidewalk, and facts, such as that there is traffic on the street, but we will ignore these matters here.) It seems to me that my acquaintance with the objects I perceive is an intimate one. I have no sense that anything stands between me and the buildings I am looking at. I do not feel that I am looking at a representation of the buildings as I might if I were looking at a photograph of them, nor that I am receiving information about them that has to be decoded. I merely open my eyes, and there they are.

This common sense view of visual perception is called direct realism. Direct realism holds that the objects one perceives are the physical objects of the external world and that these objects are perceived directly, that is, without the aid of any intermediary object or representation. A question arises about the status of the properties of objects when they are unperceived. Naı e direct realism holds that all of an object’s perceived properties persist when it is unperceived; scientific direct realism holds that only the properties countenanced by science—such as shape— persist when an object is unperceived (see Dancy 1981, Martin 1998). Object color, for example, is thought by some to be a mind-dependent property and does not, on this view, persist when an object is unperceived. Scientific direct realism holds, therefore, that color is not perceived directly.

Despite its apparent obviousness, philosophers have challenged direct realism. An important objection to the view is expressed by the argument from illusion. Suppose I am having a hallucination of Beaver Lake. What is the object of my experience? Since I am not perceiving a physical object, I must be experiencing some other kind of object, perhaps a mental object. Further, since the experience of hallucinating Beaver Lake is, we may suppose, identical to the experience of actually perceiving Beaver Lake, the object in each case must be identical. In all true perception, therefore, there must be some (mental) object that acts as an intermediary or representation in the perception of the external world. The view that one is directly aware of intermediaries has been expressed as the claim that there is a given in perception—a nonconceptual state that forms the basis for full-blooded perception which requires the possession of concepts by the perceiver.

Various strategies have been adopted to resist the argument from illusion. First, the adverbial theory of perception denies the inference from the fact of hallucination to the claim that there is a mental entity that is the object of that hallucination. Instead, the theory holds that when I hallucinate a red triangle, for example, I am ‘experiencing redly and triangularly’, where ‘redly’ and ‘triangularly’ qualify the manner of experiencing, and does not presuppose the existence of a mental object. Because no inner mental object has to be posited even in cases of hallucination, it is possible to reclaim the common sense view that in true perception it is physical objects that are directly perceived.

A second strategy is to deny the claim that the experience of true perception is indistinguishable from that of hallucination. The disjunctive conception of perceptual experience (see Snowdon 1990) holds that these experiences are constituted by different mental states. Thus, even if hallucinatory experience involves a mental representation, true perception need not, and could be direct.

A third strategy is to treat perception as a special kind of belief (or inclination to believe) rather than a sensory state (Armstrong 1961). On this view, to see a red triangle is to believe that there is a red triangle before one’s eyes. A hallucination of a red triangle is no more than a false belief and does not, therefore, call for the introduction of a mental representation any more than the false belief that the local hockey team will win the Stanley Cup requires a mental object to be the thing believed. Some support for this view derives from the phenomenon of blindsight (Weiskrantz 1986), a state in which subjects with particular forms of brain damage are able to make crude perceptual judgements about the properties of stimuli in the absence of any conscious experience. If one is willing to accept the claim that blindsighted subjects are perceiving, then blindsight undermines the common sense view that all perception involves conscious experience. Since the perception-as-belief view does not make conscious experience central to perception, it is thus strengthened.

A fourth strategy for resisting the argument from illusion has been developed by information-theoretic approaches to perception which explicate the transaction between the world and the perceiver as an exchange of information (see Dretske 1981). For example, the computational approach to vision (Marr 1982) investigates how the information about properties of the environment contained in patterns of retinal stimulation can be extracted by the visual system.

Whatever the success of these strategies, direct realism also has trouble accounting for a different problem—that of perceptual error. If one’s experience of the external world is mediated by a mental representation, then error can be explained as a mismatch between the representation and what it represents. But if there is no such representation but only a direct perception of the physical objects themselves, how is misperception possible?

These arguments support indirect or representative realism according to which external physical objects are perceived, but only indirectly, by means of the direct perception of intermediary entities (see Dancy 1981, Jackson 1977). These intermediaries are usually thought to be mental representations and have traditionally been called ideas, appearances, sensibilia, and sensa, among others. In the twentieth century they have been referred to, rather notoriously, as sensedata. Much of the investigation into the nature of the relation between the perceiver and the world has focussed on two distinct questions: first, whether perception is in fact mediated in this way; and, second, whether one is consciously aware of the intermediaries. This investigation has been made more difficult by the fact that these quite different questions have often been conflated (Crane 1992).

Indirect realism also faces important objections, however; in particular, that it seems to lead to scepticism about the external world. If I am aware only of mental representations, then for all I know there is no external world at all. One response to this objection is to claim that it misconstrues the claims of indirect realism. The theory does not say that we have no access to external objects but rather that our access is by way of mental intermediaries. A second response embraces the sceptical possibility. Bishop Berkeley (1685–1753) famously argued that there is no external world but only perceptions, or, at any rate, that the external world is no more than collections of actual or possible perceptions, a view known as phenomenalism or idealism. Despite the importance of Berkeley’s arguments for the theory of perception, however, the radical nature of his claims about the world have prevented them from being widely held.

Most philosophers of the first half of the twentieth century were indirect realists of the sense-datum variety. The early part of the second half of the twentieth century saw a complete rejection of sense-datum theory and indirect realism, and many contemporary philosophers are direct realists. Nevertheless, the tendency to think that there is some aspect of experience that involves a perceptual intermediary or that there are features of perceptual experience that cannot be construed as the direct perception of the external world has proved very strong. Some aspects of sensedatum theory have emerged again in the debate about qualia (see below), and even outright sense-datum theories are no longer anathema.

1.2 The Causal Theory

The causal theory of perception takes up a second important issue in the study of the relation between perceiver and world. The causal theory holds that the transaction between the perceiver and the world should be analyzed primarily in terms of the causal relation underlying that transaction (Grice 1961). One version of the causal theory claims that a perceiver sees an object only if the object is a cause of the perceiver’s seeing it. Traditionally, indirect realism was associated with the causal theory, but not all recent versions of the causal theory are indirect theories (see Vision 1997). Indeed, the causal principle seems relevant to any view of perception, whether direct or indirect. This is illustrated by the case of a desert traveler who hallucinates an oasis. His eyes are closed, so nothing outside of his own mind is causing the experience. However, it happens that an oasis is in fact in front of him, and, coincidentally, the hallucination matches the oasis perfectly. Despite the match between the experience and the environment, however, the absence of a causal relation moves us to deny that the case is one of perception (Pears 1976).

It seems clear, nevertheless, that an object’s being a cause of a perceptual state is not sufficient for perception. For example, the state of perceiving an object typically will be the effect of a long causal chain, but typically only one of the links in the chain will be the object perceived. Suppose that a cup is illuminated by a flashlight. The flashlight is a cause of the perceptual state just as the cup is, but it is not the object perceived, and it is hard to see how appeal to causation alone will capture this obvious fact (Pears 1976).

1.3 Color Perception And The Primary–Secondary Quality Distinction

A different question about the relation between the perceiver and the world is the question of how successful the process of perception is at informing the perceiver about her environment. This question has been explored primarily in the debate about color perception, the most actively investigated topic in the area since the mid-1980s. Although philosophical interest in the perception of color is as old as philosophy itself, the contemporary debate has been invigorated by significant developments in color science and the introduction of these developments into philosophy (see Byrne and Hilbert 1997).

The perception of color is related to the success of visual perception because there is some reason to doubt whether objects actually possess colors. If it could be shown that objects do in fact fail to have colors, then visual perception would be seen to be mistaken in its representation of objects as colorbearers. It is generally assumed that the only way in which colors could be properties of external objects would be if they were physical properties—properties countenanced by science. The epistemological question whether color vision is reliable, therefore, is often recast as the metaphysical question whether there is a scientific property that is sufficiently close to our conception of color to deserve the name. If there is no such property, then doubt is cast on the idea that color is indeed an ordinary property of objects, and this in turn raises doubt about the accuracy of vision quite generally.

Two facts suggest that colors are not physical. First, color does not appear in the basic scientific description of the world. Second, colors seem to be variable in a way that ordinary physical properties are not. For example, a yellow object against a red background will appear greenish, but the same object against a green background will look reddish, a phenomenon known as simultaneous contrast. Because of these facts, some philosophers have concluded that color is a secondary quality, that is, a property the existence of which is dependent on a perceiving mind. Secondary qualities are contrasted with primary qualities, such as shape or mass the existence of which is not dependent on the perceiving mind.

The philosophical investigation of color has led to two opposing camps. Color realists accept as an obvious fact that colors are properties of external objects. The task is to identify the physical property that is identical to color and show that the variability of color, such as that due to simultaneous contrast, does not undermine the idea that colors are properties of external objects. Color antirealists take the variability of color to be central to its nature and infer that color, as we understand that concept, does not exist. They hold that while some of the physical properties of the environment cause viewers to have color experience, there is nothing in the environment that is correctly represented by those experiences.

A special sort of realism attempts to accommodate both camps. The dispositional theory of color (usually associated with John Locke (1632–1704)) acknowledges both the variability of color and its external reality. An object is navy blue, on this view, if it has the disposition—a real property of the object itself—to cause normal perceivers to have experiences of navy blue in standard conditions. Colors are identical to these dispositions and, therefore, are properties of external objects. Color is variable because objects have at least as many color dispositions, and therefore colors, as there are color experiences.

2. The Perceiving Mind

2.1 Qualia

The debates concerning the nature of the mental states involved in perception recapitulate some of those dealing with the mind-world relation. Closely related to the color debate is the question about the existence of qualia (singular: quale, pronounced kwale). The term is sometimes used to refer to the properties of the external environment, such as object color, that are experienced in perception. It is also used in more philosophically radical ways to refer to the properties of experiences themselves that give experiences their ‘felt’ qualities. This radical understanding of qualia thus takes them to be mental entities. Think, for example, about orgasms. There is a distinct feel to the experience of an orgasm that is absent in the mere belief that one is having an orgasm. After all, others can also believe that you are experiencing an orgasm without experiencing its characteristic feel. There is, as it is sometimes put, something it is like having certain kinds of experience. According to some, this is because these kinds of experience possess qualia and beliefs do not.

Because qualia in the radical sense are mental items, the qualia debate has focused in part on whether the existence of qualia shows that physicalism—the view that the only entities, properties, and relations that exist are those countenanced by the physical sciences—is inadequate. (Philosophers who believe that it does have come to be called qualia freaks.) A central argument here is the knowledge argument (Jackson 1982): Mary is born and raised in an entirely black and white room. She becomes a scientist and, we suppose, learns everything a complete science can teach about perception and the world perceived. She then emerges from her room and, seeing a tomato for the first time, learns something new, namely, what it is like to see something red; that is, she experiences a red quale. But if Mary knows everything that a complete science can teach about the world and the perception of it, then she has learnt something that lies outside of the realm of the scientific, and physicalism is false.

The existence of qualia has also been taken to show that the dominant theory of mental states, called functionalism, is false. Functionalism holds that mental states are to be analyzed solely in terms of their relations to other mental states, relevant sensory input, and the behavioral output to which they lead. How these states are embodied in an animal, or indeed in a computer, is taken to be irrelevant to the nature of the states. Being in pain, for example, is to be analyzed as the state that is typically caused by damage to one’s body; that causes one to withdraw from the source of the damage; that is related to feelings of anger or regret; and so on. But now consider which is made possible by the existence of qualia spectrum inversion. Inverted perceivers have systematically different color experience from normal perceivers because the qualia associated with their perceptual experience is inverted: what normal perceivers call ‘red’ they see as the color normal perceivers call ‘green’; what normal perceivers call ‘yellow’ they see as the color normal perceivers call ‘blue’; and so on. Despite their differences from normal perceivers, however, spectrally inverted perceivers could never be identified. After all, inverted perceivers would learn the same language as normal perceivers: they would call grass ‘green’ and tomatoes ‘red’ even if their experience of these objects was different from that of normal perceivers. If there were inverted perceivers, therefore, nothing in their behavior would reveal that they were inverted. In particular, nothing about the functional relations among their mental states revealed in their behavior would distinguish inverted from normal perceivers. By hypothesis, however, their experience is different. This argument—the argument from spectrum inversion— seems to show, therefore, that there has to be more to perceptual states than their functional relations.

It has been argued, however, that the very idea of qualia as properties of experience is based on a confusion between the properties experience represents the environment as having and the properties experience itself possesses (Harman 1990). When one introspects to examine these putative properties of experience, one finds only the properties of the objects one is experiencing and not the properties of experiences. However, the former are not mysterious in the least; they are just ordinary properties of the physical world.

2.2 Perceptual Content

The qualia debate has also provided a way of investigating the nature of the representational content of perceptual states (see Villanueva 1996). Philosophers have asked whether there is something more to perceptual experience than what is represented by that experience—its representational content (or, simply, ‘content’). The view that denies that there is anything more is sometimes called the intentional theory of perception (Harman 1990) (where ‘intentional’ refers to what a mental state is about). On the intentional view, once the objects and properties in the scene perceived are specified, the content of the perceptual state has been exhausted. The qualia freak denies the intentional theory. He believes that there is more to perceptual states than their content, namely, the felt qualities of perceptual experiences. Consider orgasms again. Whatever the representational content of that state—perhaps that a particular physiological process is occurring in one’s body—the felt quality of the experience does not seem to be exhausted by that content.

In the realm of visual experience, intuitions are less clear. Suppose I am standing on a road on which there are two trees of the same height, one of which is at a distance of 100 meters and the second of which is at a distance of 200 meters. While I might judge the two trees to be of the same height, it is also apparent to me that the nearer tree takes up more space in my visual field. Since experience cannot correctly represent the trees both as equal in size and as different in size, it has been suggested that there is an entity, the visual field itself, that is the bearer of the latter aspect of the experience: the trees are represented as being the same size but appear in the visual field to have different sizes. There are thus two orders of properties: those of the visual field and their sister properties in the external world. On this view, the properties of the visual field are not part of the representational content of the perceptual state (Peacocke 1983).

A related question about perceptual content is whether it is conceptual or nonconceptual. The mainstream position is that all perceptual content is conceptual; the perceiver must possess the concepts for the entities and properties perceptually represented by her. For example, the ability to see a dog as a dog requires that one possess the concept ‘dog.’ This mainstream position evolved in part as a rejection of sense-datum theory and its commitment to a nonconceptual given in perception. Nevertheless, some of the considerations that led to sense-datum theory have emerged again as relevant, and it has been argued that some perceptual representation does not require the possession of concepts. One such consideration is the epistemological view that conceptual thought has to be grounded in something that is nonconceptual (see Sedivy 1996). A second consideration is that perceptual states seem to represent the environment in a much more fine-grained fashion than a perceiver has the concepts for. For example, it is sometimes thought that the colors we see are more varied and subtly different from each other than we have concepts for. A third consideration is the desire to link the conceptual nature of thought with the contents of the computational states posited by the computational theory of vision. The computational theory of vision describes some of the content of a visual state in terms of various mathematical concepts, but perceivers do not have to possess those concepts in order to see.

One way of fleshing out nonconceptual content is in informational terms. A second way of doing so is by means of scenarios. A scenario is a representation of the way the physical space around the perceiver is filled out. It is specified by identifying within a perceiver-relative frame of reference the locations of object surfaces and their visible properties. The scenario itself, rather than any mental representation, is part of the content of the perceptual state, and it captures how the world looks to the perceiver independently of the perceiver’s conceptual resources (Peacocke 1992).

A third question about perceptual content has focussed on the computational theory of vision, in particular as it has been presented by David Marr (1945–80). Theorists have asked whether the contents of computational states have to make essential reference to the external world or whether they can be fully specified in terms of the perceiver’s states alone. The former view is one (but not the only) version of externalism; the latter is a version of internalism or Individualism. (In philosophical jargon, externalism is committed to wide content whereas individualism is committed only to narrow content.) Externalism and internalism are positions that have been important in recent philosophy of mind, and the debate about computational vision has been pursued largely as an instance of the wider debate about what cognitive theory says about mental content quite generally (Burge 1996).

2.3 Spatial Perception

The scenarios mentioned above in connection with perceptual content are essentially spatial and reflect a concern in the philosophy of perception about space. The importance of this topic has been recognized since Immanuel Kant’s (1724–1804) Critique of Pure Reason tied spatial perception to the concept of an objective world. One reason for thinking that spatial perception is necessary for objectivity is that the perception of space seems essential to the formation of a concept of something that persists over time. The notion of persistence, in turn, is necessary to form a concept of an external object and, thereby, an objective world. Imagine a space-less world constituted entirely of sound. In such a world it seems impossible to articulate the distinction between, for example, two different C sounds occurring at different times and the very same occurrence of C being heard, disappearing, and then reappearing. It is argued, therefore, that without the perception of space, it is impossible to distinguish between the occurrence of multiple fleeting objects of the same type and a single persistent object (Strawson 1959). And without a notion of persistence, it is impossible to develop a concept of an objective world. A closely related idea is that the perception of space—or possibly the ability to attend selectively to different parts of space—is required in order to entertain thoughts about particular physical objects.

A second issue regarding space perception is that of the concept of a surface and its role in ordinary language and theories of perception. Under typical conditions, I only see part of the surface of an object. Does this mean that I see only part of an object? Or do I see a whole object but only part of it directly? And does that imply that I see the whole object indirectly? Investigation of the nature of surfaces thus leads to a distinction between direct and indirect perception that is rather different from the traditional one and does not depend on a distinction between the physical and the mental (Stroll 1988). The philosophical investigation of surfaces is important also because the computational theory of vision is in large part about how visual perception computes the surface properties of objects.

A third issue has come to be known as Molyneux’s question or Molyneux’s problem (see Degenaar 1996). It was formulated by the Irish scientist William Molyneux (1656–98) and discussed by John Locke with whom Molyneux corresponded. In a letter to Locke on July 7, 1688, Molyneux formulates the question: ‘A Man, being born blind, and having a Globe and a Cube, nigh of the same bigness, committed into his Hands, and being Told, which is Called the Globe, and which the Cube, so as easily to distinguish them by his Touch or Feeling; Then both being taken from Him, and Laid on a Table, let us suppose his Sight Restored to Him; Whether he Could, by his sight, and before he touch them, know which is the Globe and which the Cube?’ This question was addressed by almost every major theorist of perception following Molyneux and has been of sufficient historical interest to have led Ernst Cassirer to claim that it was the focus of eighteenth century epistemology and psychology as well as the stimulus for all of Berkeley’s philosophy. Molyneux’s question continues to receive some attention from philosophers as a way of exploring a number of issues including the relation between different perceptual modalities, especially sight and touch; the existence and nature of the concept of space; and whether or not certain perceptual abilities are learned or innate.

Finally, the question of whether some quasiperceptual representations have significant spatial properties has been examined in the debate about mental imagery, the quasivisual experiences of imagination (see Block 1981). Philosophers and psychologists have debated the claim that mental images are essentially spatial as against the view that they are no more than the language-like representations thought by some to constitute the mental representations underlying all of cognition. Because there is now evidence that mental images are produced by the same neural machinery that produces visual experience, further investigation into imagery may shed light on the nature of visual representations, and research into vision may produce a better understanding of imagery.

Bibliography:

1. Armstrong D M 1961 Perception and the Physical World. Routledge & Kegan Paul, London
2. Block N (ed.) 1981 Imagery. MIT Press, Cambridge, MA
3. Block N J, Fodor J A 1972 What mental states are not. Philosophical Review 81: 159–81
4. Burge T 1986 Individualism and psychology. Philosophical Review 95: 3–45
5. Byrne A, Hilbert D R (eds.) 1997 Readings on Color. MIT Press, Cambridge, MA, Vol. 1
6. Crane T 1992 Introduction. In: Crane T (ed.) The Contents of Experiences: Essays on Perception. Cambridge University Press, Cambridge, UK
7. Dancy J 1985 An Introduction to Contemporary Epistemology. Blackwell, Oxford, UK
8. Degenaar M 1996 Molyneux’s Problem [trans. Collins M J]. Kluwer, Dordrecht, Holland
9. Dretske F I 1981 Knowledge and the Flow of Information. MIT Press, Cambridge, MA
10. Grice H P 1961 The causal theory of perception. Proceedings of the Aristotelian Society 35: 121–52
11. Harman G 1990 The intrinsic quality of experience. In: Tomberlin J E (ed.) Philosophical Perspectives 4. Ridgeview, Atascadero, USA
12. Jackson F 1977 Perception. Cambridge University Press, Cambridge, UK
13. Jackson F 1982 Epiphenomenal qualia. Philosophical Quarterly 32: 127–36
14. Marr D 1982 Vision. Freeman, New York
15. Martin M G F 1998 Perception. In: Craig E (ed.) Routledge Encyclopedia of Philosophy. Routledge, London
16. Peacocke C 1983 Sense and Content. Clarendon Press, Oxford, UK
17. Peacocke C 1992 A Study of Concepts. MIT Press, Cambridge, MA
18. Pears D 1976 The causal conditions of perception. Synthese 33: 25–40
19. Sedivy S 1996 Must conceptually informed perceptual experience involve non-conceptual content? Canadian Journal of Philosophy 26: 413–31
20. Snowdon P 1990 The objects of perceptual experience. Proceedings of the Aristotelian Society 64: 121–150
21. Strawson P F 1959 Individuals. Methuen, London
22. Stroll A 1988 Surfaces. University of Minnesota, Minneapolis, MN
23. Villanueva E (ed.) 1996 Perception. Ridgeview, Atascadero, Holland
24. Vision G 1997 Problems of Vision. Oxford University Press, Oxford, UK
25. Weiskrantz L 1986 Blindsight: A Case Study and Implications. Clarendon Press, Oxford, UK