Tye on Phenomenal Character and Color

by Barry Maund e-mail

Department of Philosophy
The University of Western Australia
35 Stirling Highway
Crawley W.A. 6009

Tye’s principal aim is to defend what he calls ‘strong representationalism’, a view that aims to tell us precisely what the phenomenal character of our experiences is: it is the same as representational content (that meets further specifiable conditions). More precisely, it is given as the theory that runs by the title PANIC: phenomenal character is one and the same as Poised, Abstract, Nonconceptual, Intentional Content. It is nonconceptual in the sense that the subject need not possess any of the concepts that we, as theorists exercise when we state the correctness conditions for that content. It is abstract in that it is content into which no particular concrete object or surface enters. What is crucial to phenomenal character is explained in terms of the distinctive functional role the experiences and feelings, qua bearers of phenomenal character, play. They arise at the interface of the nonconceptual and conceptual domains and they stand ready and available to make a direct impact on beliefs and desires.

One of the important applications of Tye’s Representationalist thesis is to the vexed question of colors and color experiences. The thesis is used to argue that we can explain the phenomenal character of color experiences without admitting any phenomenal, subjective qualities of the experiences (without admitting qualia in the strong sense).

His setting up the framework of the book in the first two chapters is impressive. Tye provides an excellent analysis of Jackson’s Knowledge Argument concerning Mary, the all-knowing black and white vision expert. Tye is persuasive in bringing out the strengths and weaknesses of Lewis’s Ability Hypothesis (in Ch.1), as an explanation of what Mary discovers when she first experiences color.

In criticising the hypothesis, Tye says that when Mary has her first experience of red and remarks “so, this is what it is like to experience red”, Mary does discover something that the Ability Hypothesis cannot handle. She now knows what it is like to experience red. On the Ability Hypothesis, she has acquired some know-how. But, argues Tye, she retains that know-how, even after she stops having any experience of red. Intuitively “while she attends to her experience, Mary has knowledge-that she didn’t have before, knowledge that this is the experience of red”. I take it that by ‘this’ what is meant is ‘this experience’. Tye proceeds to argue that Mary can be thought to acquire, through having an experience, a phenomenal concept, where a phenomenal concept is one that plays a certain distinctive functional role. That functional role, it is further argued, does not provide a threat for physicalism. Tye’s account of phenomenal concepts also plays an important role in Tye’s treatment of the well-known Explanatory Gap. He argues that the latter is an illusion, one that can be handled, once we have a full understanding of phenomenal concepts.

In my criticisms, I shall concentrate on two chapters, Ch.3 where he describes the representationalist thesis in some detail, and Ch.7, in which the main topic is color. Since I disagree with his thesis, I shall be critical, but I should acknowledge the merits of the book. Tye is to be congratulated for spelling out and defending the theory in great clarity, and with strong arguments.

1.1 Transparency of Experience

One of the powerful motivations for the Representationalist thesis is provided by the transparency of experience. “Focus your attention on the scene before your eyes and on how things look to you. You see various objects by seeing their facing surfaces.” (p.45)  In seeing these surfaces, he holds, you are immediately and directly aware of a whole host of qualities. You experience these qualities as qualities of the surfaces. You do not experience any of these qualities as qualities of your experience.

So far, this account is non-controversial. What is problematical, however, is what exactly these qualities are, and what exactly is the sense in which one is aware of the qualities. Tye has a brief argument to dispose of two hypotheses: (i) that these qualities are qualities of one’s (visual) experiences; (ii) that they are qualities of some inner object. The best hypothesis is that they are qualities of external surfaces if they are qualities of anything (and since there are hallucinations and illusions, he holds, they may not be the qualities of anything.)

To suppose that the qualities of which perceivers are directly aware in undergoing ordinary visual experiences, are really qualities of the experiences would be to convict such experiences of massive error, But this, Tye says, is just not credible (it is totally implausible.) Accordingly, the qualities of which you are directly aware, in focussing on the scene before your eyes and how they look, are not qualities  of your visual experiences. Moreover, when you introspect your experience, you are not aware of any inner object or thing. “The only object of which you are aware are the external ones making up the scene before your eyes.” (p. 47.)

Since you are not directly aware of any qualities of your inner experiences, your experience is transparent to you. But when you introspect you are certainly aware of the phenomenal character of your visual experience. “Via introspection you are directly aware of a range of qualities that you experience as being qualities of surfaces at varying distances away and orientations; and thereby you are said to be aware of the phenomenal character of the experience.” (p. 47) By being aware of the external qualities, you are aware of what it is like for you.

What then is the visual phenomenal character? One possible hypothesis, Tye remarks, is that it is a quality of the surface experienced. But that hypothesis is intelligible, he adds, only if it is assumed that the surface is an immaterial one of the sort the sense-datum theorists posited. This hypothesis has already been dismissed. The best hypothesis he suggests, is that visual phenomenal character is representational content of a certain sort: content into which certain external qualities enter.

Two crucial claims emerge from Tye’s account:

1. If we stipulate that something is a visual phenomenal quality or quale only if it is directly accessible quality of an experience, then there are no visual phenomenal qualities or qualia.

2. There are qualities of which the subject of visual experiences are directly aware via introspection. They are qualities of external surfaces (and volumes and films) if they are qualities of anything. By being aware of these qualities, we are aware of phenomenal character.

There seems to me to be a number of problems, which I shall discuss in the following sections:

1.2 Two Senses of ‘Awareness of’
1.3. Phenomenal Character and Introspective Awareness
1.4 The Block-Searle Objection
1.5 Transparency of Pain
1.6 An Alternative Account
1.7 Projection

1.2 Two Senses of ‘Awareness of’

In developing his account of phenomenal character, Tye relies upon an account of introspective awareness of phenomenal character. (I shall examine that account in the following section. Here I want to highlight a potential problem.) Introspection of phenomenal character is said to be a reliable automatic process that takes one from being in one state to being in another, from an experience to a conceptual state. It is a process that takes, as input, awareness of external qualities and yields, as output, awareness that a state is present with a certain phenomenal character.

Unfortunately, in explaining transparency, there seems to be equivocation on “awareness of”. In seeing the facing surfaces, the perceiver is said to be “immediately and directly aware of a whole host of qualities as qualities of the surfaces” [p. 46]. Here the awareness consists in having the experiences which are nonconceptual. On the other hand, it is also stated that ‘via introspection you are directly aware of a range of qualities that you experience as being qualities of surfaces at varying distances away and orientations”. Here the awareness would seem to be conceptual. Introspection is said to be a process that results in your being in a conceptual state: a state of awareness-that.

Now, it may well be that Tye is using the expression “awareness of a quality” ambiguously, once in a nonconceptual sense, once in a conceptual sense, and everything is in order. But it would be better if he didn’t. It would be better if he made it clear which sense is being used. This is especially crucial if, as when talking about introspection, he uses the phrase “directly aware of a quality”.

I suspect that the ambiguity is not benign. Tye’s argument depends on our accepting as a fact a certain phenomenon: the transparency of experience. In asking us to focus our attention on the scene before one’s eyes, one in which you see various objects by seeing the facing surfaces, Tye asks us to take as obvious, the fact that one is immediately aware of a whole host of qualities. But which sense of “aware of” is appealed to here?

I am inclined to think that there is a third sense of “awareness of”, one that combines elements of the other two. Intuitively, I am presented with an instance of a quality (or complex of qualities: redness, roundness, hard-edgedness, . . . ,) and I am aware of it as, say, being present and before me, and perhaps as an instance of red. In accepting the phenomenon of transparency, I suggest, we are adopting some such sense. I may well be wrong in my account of the intuitive sense, but there seems to be grave possibility of error here in Tye’s “intuitively appealing” account.

1.3 Phenomenal Character and Introspective Awareness

In explaining what introspective awareness of phenomenal character amounts to, Tye emphasises that this awareness consists of awareness-that: awareness that an experience with a certain phenomenal character is present (p. 52). Introspection of phenomenal character is said to be a reliable automatic process that takes one from being in one state to being in another, from an experience to a conceptual state. It is a process that takes, as input, awareness of external qualities and yields, as output awareness that a state is present with a certain phenomenal character.

On this account, there are two distinct states: the first is an experience or a feeling or pain, which is an awareness of some objective quality, either a property of an external object (in the case of a visual experience) or a disturbance in one’s body, in the case of feelings and pains; the second is a conceptual state: an awareness-that. The first state is a nonconceptual state that is apt to give rise to the conceptual state. If one introspects one’s experiences or feelings, one forms a conceptual state: a state of awareness that an experience with a certain phenomenal character is present.

What this second kind of awareness amounts to is that I am aware that there is present something (external) that has certain public objective qualities. These qualities are qualities specific to the representational content of the experiences. It is claimed that I am not aware of any of the intrinsic qualities of this experience itself, I am only aware of the content of the experience, i.e., of the qualities that the experience represents certain objects as having.

One possibility that Tye does not consider seriously enough is that the perceiver might be aware both of intrinsic qualities and content. It may be that the intrinsic qualities contribute to the content. It could be, for example, as Hume and others have suggested, that we project (or ‘project’) some of the intrinsic qualities onto the object. Tye does reject the notion of projection, but for reasons I shall discuss later, I think that the notion is defensible.

I have a more serious objection. I think that there is an unresolved tension in Tye’s account ( a tension shared with the similar account given by Dretske). In what sense is the nonconceptual state the “awareness of an external quality” an awareness at all, an experience at all?  It is a state that carries nonconceptual content about the external quality and it is the last stage in a process that results (other things being equal) in a conceptual state, an awareness-that there is something red and moving, say in front of one. But that doesn’t make it an experience or awareness. Furthermore, to say that it has a certain intrinsic quality that is causally related in an appropriate way with the external quality. (What exactly the relationship is may be a matter of dispute and it is not easy to specify it in detail: witness the writings on the subject by Fodor.) Is that all it means to say that the state constitutes “awareness of the qualities”: that it carries this content?  That surely is not awareness in anything like a normal sense. Nor is it remotely like an experience. Why call it awareness or experience then? (I think that John Heil makes a similar criticism of Dretske in his essay in the Dretske and His Critics volume.)

Phenomenal concepts are held to be conceptually irreducible concepts that function in the right sort of way. It is part of their characteristic functional role, qua phenomenal concepts, that they enable us to discriminate phenomenal qualities and states, “directly on the basis of introspection”. Thus far, so good. But Tye’s theory is that the phenomenal qualities that we discriminate are qualities specifiable in the content of the experience, and are not intrinsic qualities of the experience. But if this is so, I find it difficult to understand how there is a difference between what results when I introspect my perceptual experience of a brown shoe on the floor as opposed to what results when I introspect my belief that there is a brown shoe. In either case, all I can discriminate are the qualities contained in the content of the relevant mental state, and on Tye’s theory, the content is the same. I know that he believes, as we all do, that the states are different, but once we explain phenomenal character in terms of content, then I cannot see how his theory has the resources to explain the difference. Nor will it help to be told that the content of the perceptual experience is nonconceptual and the content for the belief is conceptual. For it seems to me that when one introspects the perceptual experience, and discriminates the relevant qualities, one becomes aware of the qualities by conceptualising them.

One may be tempted to say that when one has the nonconceptual experience one is aware of a quality, that is, that one is aware of an instance of that quality. But this is not what Tye’s theory entails. One is not aware of an instance of the quality, i.e. of an actual instance of that particular quality. One is aware of an experience which has a certain content, that is one is aware that something has that quality. The quality, for Tye, is something abstract.

It may be replied that we can distinguish between two applications of the phenomenal concept. One is a recognitional application: one recognises a current instance of an experience of red as such an instance. the other is a non-recognitional application. But from the point of view of the subject, it is hard to see how the subject can distinguish them. It is hard to see how Tye’s theory can explain the capacity that subjects have to distinguish them in the way that they do.

1.4 The Block-Searle Objection:

Tye considers an objection raised by Block and by Searle to the representationalist approach to the phenomenal character of pain. The objection is that the possibility of seeing one’s damaged leg, say, while one is feeling pain in the leg shows that the phenomenology of the pain experience cannot be captured by its representational content since the content of the perceptual experience as one views the damaged leg is the same as, or very similar to, the content of the pain experience. The perceptual experience nonconceptually represents features such as color, shape, orientation of surface, presence of an edge, etc. It does not nonconceptually represent tissue damage (or any other comparable quality or cluster of qualities). The leg, says Tye will not look damaged in the phenomenal sense of the term ‘looks’.

But how can Tye maintain this? He argues in the case of looks red, that given that what it is to be red is to be disposed to reflect certain percentages of light, then if a spot looks red, then the spot looks disposed to reflect such and such percentages of the light. If that is so, why should it not be that the damaged leg look damaged in the phenomenal sense. In any case, it seems to me that the phenomenal sense applies to a range of features or cluster of features besides the ones he cites: there are rusty looks, wooden looks, jarrah looks, metallic looks, and damaged-tissue looks. That is, there is a characteristic experience-type that things which are rusty, wooden, jarrah, metallic, and damaged tissue cause in optimal conditions.

1.5 Transparency of Pain

Tye’s Representationalist thesis has considerable plausibility concerning the transparency of, say, visual experience, though I think as Moore suggested, it is possible to give a different interpretation of the transparency. However, I cannot for the life of me see how the experience of pain is transparent. I acknowledge that our experiences of pain have representational content and that they represent a bodily region, say a leg or head or tooth, as being a certain way, and that they can be used to represent a bodily disturbance, but how could it be thought that the experience is transparent? The natural account of pain is surely that we are aware of the bodily disturbance through being aware of the intrinsic quality of the pain. The experience of pain has content all right: it represents the intrinsic quality as going on in the body. The ache is experienced in the head, behind the eye, say. Of course, we do not believe that the feeling is there, but the phenomenology of pain is that we experience the pain as if it were there. I feel the pain in the foot just as I feel movement and pressure in the foot. Indeed the feeling of pressure in the foot can quickly become a feeling of pressure and pain in the same region in the foot.

1.6 An Alternative Account:

The example of pain and the possibility of providing an alternative explanation of transparency for visual experience should lead us to take seriously that alternative. That is to say, we should consider the sense-datum alternative (in one sense of ‘sense-datum’).

Tye considers the sense-datum hypothesis very briefly. The theory is “unacceptable, however, for a whole host of familiar reasons”. In my experience most of the familiar reasons usually given are either bad ones or not relevant to the most carefully formulated versions. An example are the reasons given by John Austin in Sense and Sensibilia, where he almost exclusively concentrates on A.J.Ayer’s version and then often not fairly. (Austin is cited with approval in Putnam’s Dewey Lectures as providing the best arguments against sense-datum theories.) I trust that Tye’s “familiar reasons” are not these. In my opinion, the arguments presented by Howard Robinson in his Perception are very persuasive, if not compelling. They are certainly strong enough to deserve consideration. I do not think that they fail for any of the usually cited “familiar reasons”.

It is interesting that at a later stage, (p. 48) Tye asks “what then is visual character?”. One possible hypothesis, he writes, is that it is a quality of the surface experienced. That hypothesis, he claims, is intelligible only if it is assumed that the surface is an immaterial one of the sort the sense-datum theorists posited. (By ‘intelligible’, I assume he means ‘plausible or ‘reasonable’.) He does not consider further the claim, going on instead to suggest that the best hypothesis (no doubt for avoiding that consequence) is that visual phenomenal character is representational content of a certain sort. I think that he (and we) should consider the alternative more seriously. I would have thought that, intuitively, it is highly plausible that there is such a thing as a visual phenomenal character which is a quality of the surface experienced, and that for the innocent perceiver, it is a quality of an actual physical existing surface. It is captured in the thought that most people have that there is a range of objects (or their surfaces) that have a sensuous phenomenal character: a Perth (or Californian) sunset, a field of poppies, the Pacific Ocean, a film by Visconti. Likewise for other forms of  perception: the taste of a ripe peach or of a fine red wine, the sound of a piece of music.

If Tye wishes to reject (as he does) this ordinary intuitive belief, that this character is a quality of an actual existing surface, then it would seem that if it is a quality of anything it is a quality of an immaterial surface. Tye rejects that too, but it is hard to see how the Representationalist account that he espouses can handle the sensuous character of the objects described above.

This point is similar to the point raised in the objection levelled by Block and Searle against the representationalist account of pain. It is particularly significant in handling color experience. It is non-controversial that we experience color as part of the physical world, as on the surface of physical bodies, as spread through volumes and films. One of the central features of these colors is their sensuousness. It is the feature which Tye described in an earlier paper. [Tye (1992), ‘Visual Qualia and Visual Content’ in Crane Tim (ed.) (1992), The Contents of Experience, Cambridge: CUP, pp. 158-77.]  There he considers a hypothetical appeal to first person phenomenology:

Standing on the beach in Santa Barbara a couple of summers ago on a bright sunny day, I found myself transfixed by the intense blue of the Pacific Ocean. Was I not here delighting in the phenomenal aspects of my visual experience? And if I was, doesn’t this show that there are visual qualia? [p.160]

He is not convinced. It seems to him

that what I found so pleasing in the above instance, what I was focussing on as it were, were a certain shade and intensity of the color blue. I experienced blue as a property of the ocean not as a property of my experience. My experience itself certainly wasn’t blue. Rather it was an experience that represented the ocean as blue. What I was really delighting in then were specific aspects of the content of my experience. It was the content, not anything else, that was immediately accessible to my consciousness and that has aspects I found so pleasing. [p. 160]

But how does the representational account, in his version, handle the sensuousness of the experience? On the face of it, it is part of the phenomenal character of the experience. In the example given, the phenomenal character is part of the representational content, of the experience. But I cannot see how Tye’s thesis can handle this situation, given how content is construed. It may be that the content can be specified in terms of the property redness, but we have to think of what redness amounts to, on Tye’s theory: it is a reflectance profile. We might point out that this property is not the right kind of property: for one thing it is a dispositional property and if I am enjoying anything it not the dispositional property but the exercise of the disposition. But put that point aside for the moment. There is a more general problem.

It is not clear to me how one is supposed to be aware of this property when I delight in the sensuous character of the colored objects. I am in a certain state, one that carries nonconceptual content about color. What it is for this state to carry nonconceptual content is for it to be of a certain type, one that is causally related (in the right way) with red-type states. This means that on a specific occasion when I delight in the colored experience, the property red may not be instantiated. And even if it is instantiated, how exactly can I be experiencing it?

I am not aware of the intrinsic qualities of the experience, but how can it help that I am aware of the content which is nonconceptual.  To be part of the content, a property must be causally related in the right kind of way to my experience, but given Tye’s account, to be aware of the property is simply for me to be in a certain state, which stands in a complex relation to instances of the property. It does not count as any normal kind of awareness. Nor does it help to bring in reference to the conceptual state that the nonconceptual state leads to. This second state is a state of awareness-that, but consider what I am aware of: I am aware that I am having an experience which represents an object as being a certain way. But it is hard to see how this kind of awareness is anything I could rejoice in. Why should it make any difference to me that I become aware that the sunset is red, any more that I learn from reading it in the paper that the sunset at 5 p.m. is red. Of course they have different causes, but given that in each case the conceptual state is an awareness-that, there does not seem to be any significant difference between the two.

The point might be put this way. The nonconceptual content of an experience is informational content. It must be more that information, but it is at least that. What makes it representational is related to the causal history of the state. The quality represented is one that has had a role in the person’s acquiring that kind of state, and the use that the perceiver can make of that information. But what can that quality have to do with my current experience. It seems to me that what is required is either a current instance of the quality is there for me to be aware of, or alternatively the perceiver is disposed to act towards the object bearing the quality in a way that is sensitive to its presence.

1.7 Projection

In arguing that the phenomenal character of an experience is given by the content of the experience,  Tye does not consider that the perceiver might be aware both of intrinsic qualities and content. It may be that the intrinsic qualities contribute to the content. It could be, for example, as Hume and others have suggested, that we project (or ‘project’) some of the intrinsic qualities onto the object. Near the end (p. 165), Tye says that projectivism, upon reflection, seems incomprehensible. He claims that he does not understand how subjective qualities can be projected, and especially onto physical objects. The answer to this is that the projection is not literal, but metaphorical: it is a case of ‘projection’. One example is that a picture of a man in a red coat can be such that the red in the picture is used to represent the red of the man’s coat. That is to say, the red in the picture exemplifies the red of the coat represented. Likewise two actors kissing on stage can exemplify two characters kissing. In the case of pain, the pain one feels in one’s leg is a subjective feeling that one ‘projects’ onto the leg. It is not a real projection. One has a body image which represents the body. The pain is projected on to, is located on, that part of the body image that represents the leg. Likewise with colors. There is a subjective quality which one ‘projects’ onto an external object, say to the moon, to represent it as yellow. One does not actually projects the quality onto the physical moon: one projects it onto that part of one’s subjective visual field that represents the moon. The basis of the metaphor is that the perceiver automatically and naturally takes  the representation of the leg, to be identical to ‘that which I reach for when trying to ease the pain’, i.e. the part of the leg that is the source of the pain. Likewise, s/he takes the moon-representation to be identical to ‘that which I point at when indicating the moon’.

Indeed, the notion of projection seems perfectly fitted to Tye’s Representationalism, or to a suitably modified version of it. The intrinsic quality of the subjective experience need not be thought of as projected onto a physical object, or onto the subject’s leg: it is ‘projected’ onto the representational content of the experience. That is to say, it contributes to, or is a part of, the content, in the way that a property of a photograph may contribute to the content of the photograph.

2. Color

2.1 Tye’s position on Color

In the chapter on color, Tye begins by describing a view that has a distinguished history, and has been recently defended by a number of different theorists: Cosmides and Tooby, Boghossian and Velleman, and L. Hardin. The view is that science has rebutted the common-sense view of color. This commonsense view is that colors are objective, perceiver-independent qualities of physical objects.

Tye argues that the scientific facts allow for another proposal consistent with the commonsense view: colors are objective, microstructural properties, e.g., reflectances. It is important that such a proposal be defended since, as Russell pointed out, and Tye agrees, if things in the world lack color, then the only way to avoid the conclusion that there is a deep and pervasive error in our color experiences is to claim that things we experience are not outside the mind at all.

Tye’s treatment in this chapter is largely confined to treating some of the arguments developed by some of the theorists cited above. I do not propose to discuss these arguments apart from the one by Hardin because I think that there are other powerful reasons for thinking the commonsense view false. Tye’s discussion is relevant, however, for I think that in this discussion he makes a number of questionable assumptions.

There seem to be two major elements in his characterisation of the commonsense conception:

1. The colors we see objects and surfaces to have are observer-independent properties of those objects and surfaces. We think of colors as inhering in surfaces and in volumes and films. We take it for granted that objects typically retain their colors when not seen, thereby helping us to re-identify them,

2. Another important fact about color which is manifest to us in our everyday life is color constancy. Objects do not typically appear to change their colors during the day as the sunlight changes.

There are, Tye thinks, three objective accounts consistent with this conception, but there are other reasons for ruling out two of them. The most plausible account, he holds, is Reductive Physicalism: the thesis that colors are physical properties whose natures are discoverable by empirical investigation. The version of this broad thesis that he seems to favour is some form of Anthopocentric Objectivism. (Colors are identified with some objective properties which in themselves are of little interest except in so far as they have a distinctive effect on those organisms, especially humans, built to respond to them.) In the case of surface colors, reflectances are the natural candidates.

My major criticism of Tye’s approach is that the conception of the commonsense view of colors that he is working with, is under-described. It is clear that the two elements Tye attributes to the common-sense conception are not enough, they do not enable us to distinguish the concept of color from many other concepts. There are, that is to say, important elements other than the two he describes. Once we take these elements into consideration, it will become apparent that we will have trouble retaining the commonsense conception. Although he does consider some of them in his discussion I do not think that he does enough to justify dismissing them. I think that some of these elements were in the background of the thinking of many of the traditional philosophers and scientists, and if they were not, they should have been. They should be present in our thinking.

2.2  Reductive Physicalism

In order to defend reductive Physicalism about colors we need to give an accurate characterisation of the common-sense conception. In order to do that we need to say something about how the ordinary concept of color operates, e.g., about the practices, (discursive and non-discursive) in which the concept is embedded.

We should note that in many scientific microstructural reductions, what happens, in effect, is that the scientists, or the philosophers who do the reducing on their behalf, redefine the prior concept and having replaced it with a new concept, then identify the property picked out with the new concept with a new microstructural property. Of course it is not arbitrary that this is done, but given that it is done, it is clear that it is being assumed that the prior concept is inadequate and needs to be replaced. Tye can hardly apply this approach to reductive Physicalism for colors, since it involves rejecting the commonsense conception of colors. He wants to argue that reductive Physicalism is consistent with that conception. (As Searle has pointed out, the reconstruction of the prior concept often omits elements that have to do with subjective elements. To omit these in the case of color hands the reductionist a pyrrhic victory.)

If physical reduction for colors is to be consistent with the common-sense conception, then it needs to respect the way the concept of color operates. That is, it must acknowledge that the concept operates within a certain set of conceptual practices, central to which are the naming practices, learning and teaching practices. If color terms are taken as functioning as names for certain properties, then there are different ways in which they might do so. There are three primary candidates we have to choose from: causal reference-fixing models, descriptivist models and models which combine descriptive and causal elements. These theories offer different models for understanding what kind of properties color terms name, and how they name them. Putting the point in relation to concepts, we can say that the color concepts expressed with terms such as “brown”, “olive”, “turquoise”, “crimson”, etc., conceptualise colors as certain kinds or properties. The different models offer different ways for the conceptualisation to work.

It seems to me that since there are different models that might apply, the reductionist must defend their choice. Tye does not do so but rather presupposes one of the models. More to the point, I think that there is sufficient reason to think that he has made the wrong choice.


There are several problems. In the first place Tye assumes a certain view about how color terms operate, e.g. in terms of a certain way of construing the sense-reference distinction. That is controversial. It seems to me that our ordinary color terms function in such a way that they are understood to be ostensively defined. One teaches and learns the terms through ostensive situations and through dealing with examples of colors.

If we accept this view, we need to say something about how the ostensive definition should be understood, or at least filled out. that is to say, we need to characterise the ostensive definition I such a way as to be faithful to the teaching and learning practices. It seems to me that the more faithful way is to characterise red, say, as “this manifest feature exemplified here (and in blood, and sunsets, and John Bull coats, and Essendon football jumpers, and so on”, and not as “red is the hidden feature responsible for the way that blood, sunsets, . . . appear”.

I do not see the problem with saying that our ordinary practices assume that color is a manifest, rather than a hidden, feature. If color is not a manifest property, what is? And surely the fact that we theorists function very well with a notion of hidden essences (i.e., with properties whose natures are non-manifest) shows that we have a viable concept of being manifest. Given the centrality of the ostensive teaching and learning practices, we can identify other features essential to the characterisation of the commonsense conception: colors are intrinsic, non-relational and non-dispositional features. Finally, as was argued in section 1.6 above, there is the sensuous character of the objective color properties that must be acknowledged.

There is a second important problem for Tye’s account. The microstructural properties associated with the various colours are very different for surfaces than they are for lights, and as they are for volumes or for films, or for spots of light on the wall, or for objects such peacock feathers where diffraction effects operate. For the ordinary concept, it does not matter much whether we are talking about houses, or bird feathers or sunsets, or skies or glasses of wine, we use the terms blue, red, burgundy, claret etc to apply to all of these. That we do so is compatible with our recognising that there are certain important differences in the ways these colors appear. There is no question that sky blue is similar (enough) to a blue-bird’s coat or a Cambridge rugby jumper. This point is important. In teaching situations we do not hesitate to use examples for all of these kinds of objects, to teach ‘blue’ for example with respect to the sky, the ocean, the blue in a peacock’s tail, the blue on a flag, and so on. If we mix lights and cast them on a screen, there is little hesitation in identifying them as yellow, blue, orange, and so on.

As I pointed out earlier, it is clear that the two elements Tye attributes to the common-sense conception are not enough, that  they do not enable us to distinguish the concept of color from many other concepts. Once we have filled out the characterisation of the commonsense conception in more detail, we can see that there are crucial features that Tye has ignored. There are two of particular importance. The first is that colors are treated as properties whose natures are manifest. The second is that volumes, films and surfaces share the same property. The first element rules out the possibility that according to the commonsense conception the natures of colors could be hidden. The second rules out the possibility that the nature of surface color could be different for that of volume color and film color etc. This fact precludes (identifying or reducing) the property picked out by the commonsense concept with any microstructural property. The objectivist it seems to me is forced to reject the common-sense conception.

There are other features that are also important: that colors are intrinsic, non-relational, non-dispositional and sensuous properties. Even without them, but especially with them, the commonsense concept is in trouble.

2.4 Unitary/Binary Structure of Colors

Tye considers and rejects Hardin’s objection to physicalism with respect to colors: some colors are unitary, in particular red, yellow, blue, black and white. Others are binary, for example orange, pink, purple and blue. Tye argues that, contrary to what Hardin contends, this provides no problem for physicalism. To handle it, however, Tye treats orange, for example, not simply in terms of having a certain reflectance profile, i.e., as being disposed to reflect different proportions of light from the range of wavelengths associated with standard illumination. Rather it is reflectance which has a certain power defined in terms of the effect it has on the perceiver’s opponent-processing channels (suitably specified, as in the best available theory). Tye may well be right that in this way a theorist can explain the unitary/binary structure of color appearance. Hardin’s point surely was directed against certain objectivist conceptions of color: those that construe colors as objective, perceiver-independent properties of objects, that is, as either intrinsic properties of surfaces (volumes, etc.,) or as dispositional properties such as reflectances. These properties could not explain the underlying unitary/binary structure. It might be possible to reconstruct a new physicalist property that was no longer perceiver-independent, i.e, a property defined in terms of its effect of the perceiver’s opponent-processing channels. Perhaps Hardin’s argument fails against that view, (without further argument), but that view was not the original target. More importantly for Tye’s purposes, however, such a concept would be different from the commonsense one.

Moreover, if we recall Tye’s characterisation of the transparency of experience, it is hard to see how this concept of color fits his description of the transparency of experience. There he argued that in seeing the facing surfaces of objects before you “you are immediately and directly aware of a whole host of qualities. You experience these qualities as qualities of the surface. You do not experience any of these qualities as qualities of your experiences.” [p.46] If we can be so confident of what we experience here, that the qualities are not qualities of our experiences, for example, is it all plausible that these qualities that we claim to experience are the reflectance properties of the type that Tye requires. I suggest not.

2.5. Conclusion

I do not think that Tye has done enough to justify his claim that his version of Representationalism can explain our experience of color, and in particular the phenomenal character of color. It seems to me that his account of the commonsense conception of colour is underdescribed. Once we fill out that conception in an adequate manner, I claim, the difficulties for Tye’s thesis clearly emerge. Although the problem of consciousness seemed to be something of a side-issue in the development of naturalistic accounts of mind in the mid to late twentieth century, real progress has been made recently. We now possess, thanks to Michael Tye and other philosophers, an impressive variety of sophisticated theories of consciousness. To my mind, the most significant of these are the representational theories of consciousness.

top | back to symposium index | Tye's reply