Qualia

Realism

qualia cartoon

  

Back to Home Annotated Bibliography


 

"Qualia" (singular, "quale") is a term introduced by C. I. Lewis (1929, p. 121) to stand for "recognizable qualitative characters of the given". Lewis’s examples were red, blue, round, and loud. Although the predicates for these qualia are also used to denote properties of physical objects, Lewis was explicit that properties of physical objects are not qualia: qualia are properties only of the given. To give a helpful, although not perfectly unambiguous example, the roundness of a coin is not a quale, but if we look at the coin straight on, so that it looks round to us, then the roundness that characterizes our visual experience will be a quale. A frequent and convenient extension of Lewis’s usage allows the phrase "a quale" to refer to an instance of a qualitative character of the given. Thus, while red and round are distinct qualia (distinct qualities of possible experiences) they may both be exemplified in a (single) red, round quale (i.e., there can be an experience of the qualitative kind that one might well have when looking at a red poker chip from straight on).

Clarifying qualia evidently goes hand in hand with motivating and clarifying the distinction between properties of ordinary physical objects and properties of experiences. We shall begin with some very traditional approaches to this key distinction, and then set out some more recent arguments for qualia. While this article will focus on the positive and will not offer a survey of objections to qualia with qualia realists’ replies, it will be necessary to discuss the main motivation for views that oppose qualia realism.

The Argument from Illusion

Hepatitis patients sometimes have a bitter taste when they eat honey. Honey from the same pot eaten by others tastes just as sweet as usual, and honey eaten from the same pot by the same person before or after the bout of hepatitis tastes sweet. For these reasons, it is not at all plausible to say that the honey is bitter. But the bitterness was not merely thought of, or imagined: there was a palpable taste and it had a definite, bitter character. If the bitterness cannot be attributed to the honey, it must be attributed to something else – and the natural candidate in this case is the taste, that is, the experience that was caused by eating honey while ill with hepatitis.

Shoppers have learned to be wary of buying clothes in stores with fluorescent lighting, for it is perfectly possible to buy something that looks blue in the store, only to find that it looks green when one gets it outside. Since no one has washed, dyed, burned, chemically treated, or otherwise altered the composition of the fabric or what adheres to the fabric, it is not plausible to say that the color of the clothing has changed. But there is a color difference of some kind, and it is not merely a matter of thinking or imagining a different color. The difference is plausibly attributed to a difference in color experience, caused by the change in lighting conditions.

Cases of the kinds just instanced can be multiplied indefinitely. They focus our attention on the need to attribute qualities to something other than the ordinary physical objects that are affecting our sense organs and thus causing our experiences. Experiences themselves have seemed to many to be the natural candidates for the required attributees of such qualities. When they are thus thought of, individual experiences are qualia (individuals) in the qualia realists’ sense, and the properties we have mentioned are qualia (in the sense of properties of experiences).

The argument from illusion can be used to introduce a further claim that is central to qualia realism. If we make a mistake about the color of some clothing, we are, it seems, led to it by having a certain quale. We may say that such a quale is misleading, meaning that it helps induce us to form a false belief. But we do not have any false belief about which quale characterizes our experience – for example, it is precisely because we do not make a mistake about the blue character of our experience that we do make a mistake about the color of the clothing. To generalize, qualia are not thought to be subject to mistaken belief in the way that ordinary physical objects are. (Just how strong this protection from error should be taken to be is a controversial matter. See the section below on Qualia and Knowledge.)

Illusions provide a convenient way of focusing our attention on qualia, as opposed to the properties of physical objects. But it must not be thought that we have qualia only when we are in some way deceived. In a straightforward case of a person with normal vision viewing a piece of green clothing outdoors on a sunny day, there is still a difference between the molecular-structure properties of the dye, and the way the clothing looks. Only the latter is a quale. In general, qualia can be regarded as the ways things appear, whether or not there is anything problematic about the physical objects that appear, or the circumstances of their appearing.

This last feature of qualia makes it easy to distinguish them not only from properties of the ordinary physical objects that impinge upon our sense organs, but also from the properties of the sense organs themselves and from properties of the neural activations resulting from stimulation of our sense organs. Qualia realism holds that certain neural activations cause our experiences to have qualia such as green, sweet, peppery, cold, pain, itchiness, nausea or sexual pleasure. On the one hand, no one would suggest that these properties are literally had by sets of neural activation events. And, on the other hand, nothing appears to us in ordinary experience as having the kind of complexity that neural events actually have, according to our best neuroscience. Thus, qualia realists easily distinguish between qualia (= qualitative properties of our experiences) and properties of neural activations.

The Argument from Science

Consider what science tells us about our seeing of the colored surfaces of ordinary physical objects. The surfaces themselves are collections of atoms, bound into molecules that, among other properties, have a resonant frequency. Atoms and molecules are colorless. The latter absorb light at their resonant frequencies, and reflect light of other wavelengths. Wavelength is – evidently – length, and it is correlated with energy level. Light waves of different energy levels interact differentially with various chemicals in the cone cells of our retinas. The products of these interactions cause neural activations, which consist of progressive sequences of ion transfers across neural cell membranes. At the ends of the axons of activated cells, neurotransmitters are released, and these affect the probability of similar activations in adjacent neurons. After this kind of transmittal of activity has happened in a good many neurons (perhaps with recurrent connectivity), we experience the lovely orange and blue of the bird of paradise blossom that blooms a few feet in front of us.

This story is abbreviated in many respects, but it is not false. The point of rehearsing it is to emphasize the fact that it is almost entirely a story of energies, distances, frequencies and the like. There is not one word about color in it, until the very last step. A natural reaction to this account would be to say that color turns out to occur only in experiences and nowhere else, all the other items in the story being only causes of conditions that, in the end, cause qualia.

A likely response to this description is the view that of course there are ordinary physical objects that are colored -- these are just objects that can normally cause color experiences of a certain kind. Thus, an object that would normally cause a green experience (in people with normal color vision) can perfectly sensibly be said to be a green physical object. Apart from a technical reservation, this essentially Lockean (1690) response may be acceptable. However, experiencing green does not itself reflect light or thereby cause green experiences, so the main point that we must distinguish qualia from properties of physical objects (that may go by the same name) is reinforced. (The technical point is that, in general, physical surfaces of very different microphysical structure can cause indistinguishable qualia. So, there may be no property that explains why, e.g., green physical objects should be able to be grouped by color, except for the fact that they happen to cause the same qualia when they stimulate normal human sensory equipment.)

The argument from science evidently generalizes to enable us to distinguish taste, smell, and sound qualia from the chemical or vibrational properties of objects that cause those qualia. The argument further generalizes to enable us to distinguish the qualities of bodily sensations from their causes. Thus, for example, tissue damage produces progressive sequences of ion transfers across the membranes of several kinds of neurons. After these sequences have caused other sequences of the same general kind in other neurons, a pain is experienced. That is, pain is plausibly attributable to the experience itself, with the activities in neurons functioning as causes that bring the pain about.

Afterimages

In a magic shop one can sometimes find a novelty consisting of an American flag printed in the complementaries of its usual colors. The instructions advise one to stare fixedly at the oddly printed flag for a half minute, and then look at a white wall. The result is that one "sees" the flag in its normal colors.

The causes of afterimages are exceedingly complex, and we shall not go into them except to say that afterimages, like ordinary color perceptions, are caused by conditions of neurons in the visual system. For our purposes, the interest of afterimages lies not in the details of their causes, but in the fact that one can have an experience just like seeing red stripes, when nothing in the eye or in the physical world beyond the eye is red. Thus, afterimages give us another case in which we seem forced to recognize a definite character that can only be attributed to the experience or, in short, we are forced to recognize a quale. Similar considerations yield taste and tactile qualia, as when an artificial sweetener leaves an aftertaste, or removal of a tight hat leaves the feeling of contact where the hatband was.

The Unbearability of Qualia

Reflection on the way we have introduced qualia will make it clear that the arguments for them can function also as arguments for dualism, that is, as arguments for the view that there is something in the world besides our bodies, the parts of our bodies, and the ordinary physical objects that lie outside us. For this reason, acceptance of qualia has often been seen as an intolerable offense to the dominant late-twentieth century ideology of materialism. Thus, an enormous literature exists, the point of which is to explain away the threat that qualia present to materialism. A lesser, but growing body of literature has arisen in which qualia are defended against these attacks.

Since other articles in this Field Guide address various materialist arguments against qualia realism, the focus in the present article is upon positive arguments for qualia. There is, however, one piece of dialectic that may be usefully identified here. Some materialists take comfort in the view that qualia realists (sometimes known as "qualophiles") cannot prove that dualism is true. The argument for this denial, however, sometimes boils down to the claim that qualia realists cannot prove that materialism is false. It is evident that materialism is most secure against attack by qualia realists when the content of "materialism" is very small. We thus find some materialists whose firm attachment to materialism is proportional to the emptiness of the materialism they assert. Experience, they hold, is certainly material; but as to how that could be true, or how we might find out whether it is true, or even what in the future will be held to be essential for a thing to be "material", are matters on which we are now ignorant. It has even been held (by McGinn, 1991) that we are constitutionally incapable of coming to understand how qualia could be material.

To empty materialisms, qualia realists may reply that, although no refutation is possible, none is needed. Empty materialism may indeed not be contradictory, in which case a counterdemonstration is not possible. A view that stands only by studied reticence, however, is a form of skepticism, and although it may correctly identify the frailty of our knowledge, it gives no specific reason for doubt about dualism. In parallel with theorizing in science, it may be more reasonable to hold a risky, but contentful, qualia realism than an impregnable, but empty, materialism.

Many recent qualia realists, however, have not been content with this kind of response, and have sought to provide reasons for discontent with materialist views. Unlike the arguments already reviewed, these newer arguments have been framed to explicitly engage a philosophical milieu in which materialism is a default assumption. The sections below turn to arguments of this kind.

The Knowledge Argument

Frank Jackson (1982; 1986) offered what is now a famous argument for qualia. His character, Mary, is a brilliant scientist who knows everything there is to know about the physical world, including, of course, light waves, the molecular structure of surfaces, and brain states. She has, however, been confined to a room in which there are only black and white objects, including a black and white TV set. She has thus learned all about (chromatic) color experiences without ever having had one. In particular, she knows all about the structure of the skins of ripe Delicious apples, all about light wave reflectances, and all about what goes on in John’s brain when he sees a ripe apple in daylight. Now, she is to be allowed out of her special room, where she will, for the first time, see a ripe, red apple in daylight. The question is, whether she will come to know anything about John that she did not know before her release?

It seems overwhelmingly plausible that she will. Although she knew everything about John’s neural changes when he looked at ripe apples, she did not know what a ripe apple looked like to John, or what it was like for something to look red to him. After she has had her own red experience, she does know what John’s experience upon looking at ripe apples is like. Thus, there is some fact – the fact that John’s experience had this definite character – that she did not know before her release, but that she does know afterward. But she knew all the physical facts before her release. Thus, there is a fact that is not a physical fact. And since this fact is the fact that John’s experience has a certain quale, such facts about qualia are not physical facts.

We will refine this argument below. For the present, let us note that its conclusion has been mightily resisted. Among the most cited attempts to avoid Jackson’s conclusion is that of Lewis (1983, 1990) and Nemirow (1980, 1990), who hold that Mary acquires some new abilities but no knowledge of new facts. Thus, after her release, Mary is able to imagine John’s state or remember what it was without calculation of his neural state; but the state she knows is the same one that she once could only reach by calculation. In contrast, qualophiles do not deny that Mary acquires some new abilities, but they doubt that these can adequately be described without reference to her knowledge of new facts. For example, post-release Mary can indeed imagine John’s state, but what she can then imagine is how things look to John, namely red; and whereas before her release she could have known that "red" is the right word for how things look to John, she would not have known what she now knows, namely, what looking red consists in.

Jackson has recently disavowed the success of the knowledge argument, citing the following maxim: "do not have opinions that outrun what is required by the best theory of these opinions’ causal origins" (Jackson, 1998, p. 77). Since Jackson also assumes that Mary’s coming to know what it is like to see red has physical causes, he concludes that what she learns should not outrun "how things are physically". However, from the premise (P) All constituents of Mary’s knowledge have physical causes, it does not follow that (C) All constituents of Mary’s knowledge are physical; i.e., it is possible that Mary’s knowledge of what it is like to see red contains a constituent that is physically caused but not itself physical. It is thus possible that Jackson’s maxim can be accepted while still holding that Mary learns new facts upon her release from the black-and-white room. Whether this possibility can be sustained upon thorough examination is likely to be a subject of lively discussion in the next few years. (see KA entry and annotated bibliography for a fuller discussion)

Distinct Conceivability and Kripke’s Argument

It has seemed to many that qualia are conceivable independently of their neural causes. For example, pain seems to be a distinctive quale, knowable for what it is without reference to its causes. If we were to believe that pain is caused by astral forces, or imbalance of bodily humors, we would believe egregiously false causal theories, but we should not thereby be convicted of self contradiction, or of necessarily misunderstanding what pain is. Similar remarks hold of itches, tingles, and feelings of warmth, and of nonbodily sensation qualia, e.g., red, bitter, or pungent.

Qualia realists can regard this independent conceivability of qualia as supportive of their view, because independent conceivability of qualia is naturally explained if qualia actually are distinct from physical properties. Materialists, however, object that independent conceivability is no argument for real distinction. Many have called upon certain examples from science to support this point. For example, heat is conceivable independently of motion of molecules; yet it is overwhelmingly plausible to hold that heat (of bodies) is nothing but motion of molecules. If motion of molecules is what heat (in bodies) is – if, in short, heat is identical with such motion – then independent conceivability does not show distinctness of properties, and for all we know pain is, in fact, identical with certain physical properties (presumably, properties of certain neural events in our brains).

Kripke (1971), however, has given a powerful argument that undercuts the analogy between pains and heat. In brief, assuming throughout that our present scientific views are correct, heat is necessarily identical with molecular motion. This does not at first appear to be correct, for it seems that we can imagine another possible world in which heat was something else. According to Kripke, this appearance is misleading, and rests on a confusion. What we can imagine is that we might have discovered that what feels hot to us was produced by something other than heat, that is, was produced by something other than molecular motion. And we can imagine that molecular motion might have produced a different kind of feeling in us. Neither of these is imagining that heat is not molecular motion; they are only imaginations that heat (=molecular motion) might not have been what produces heat qualia in us. On this analysis, there is no possibility that heat is not molecular motion, that is, heat is necessarily identical with molecular motion.

We cannot give the same analysis for pain and properties of neural events. This is because, while

      the feeling of heat (heat quale) = the quale produced in us by hot things

it is not true that

the feeling of pain (pain quale) = the quale produced in us by pains.

Instead, the feeling of pain just is a pain. For this reason, the explanation as to how we could become confused, and think that heat is only contingently identical with molecular motion, cannot be used to explain how it could seem to us that pain is only contingently related to neural events in the brain.

Materialists are thus placed in a dilemma, in which they must either (a) retreat and allow the distinctness of qualia from properties of physical events, or (b) insist on the necessary identity of qualia and some physical properties, despite the appearance that their connection can be at best contingent. The cost of (b) is that a necessary identity is asserted without any explanation of how it could appear to be only contingent.

Inversion Arguments

Many children have wondered whether what their playmates see when looking at, e.g., ripe tomatoes, might actually be the same color that they themselves see when they look at summer lawns, and vice versa. What gives this wonderment a special fascination is that it is soon followed by the thought that one will never know. One’s friends have obviously learned to call ripe tomatoes "red" and summer lawns "green"; but it may be that their associations of words to colors are correspondingly inverted with respect to one’s own. In that case, there will be complete agreement in classifications of things according to color, whether one’s friends’ qualia match one’s own or are inverted with respect to one’s own.

If inverted spectra are possible, then there must be qualia that can be inverted with respect to one another. Attempts to dispense with qualia in favor of reports about colors, or other behavioral dispositions, must be abandoned; for the generalization of the possibility of inversion is that all the behavioral dispositions of another person could have been systematically attached to qualia that are inverted with respect to ours. In that case, behavior and dispositions to behavior (i.e., functional characteristics) would be indistinguishable, yet there would be an experiential difference.

Inverted spectrum arguments have stimulated careful investigations (e.g., by Hardin, 1988; Clark, 1993) that have cast doubt upon the empirical possibility of completely inverted spectra. Color space is not symmetrical. For example, there are more distinguishable shades between yellow and red than there are between blue and green. Darkening of most colors leads directly to black, but yellow is an exception: its darkening leads to the distinctive further color of brown. Such asymmetries suggest that undetectable color inversion is unlikely. For example, a person who said that the color of jonquils darkened to the same color as most other things, but that there was something distinctive about darkening of Microsoft’s signature color, might thereby betray a color inversion with respect to other people. Thus, it may be urged, references to color qualia can, after all, be reduced to dispositions to give reports about colors and engage in other forms of behavior that vary with differences among colored things.

One reply to these doubts about the effectiveness of arguing for the indispensibility (or irreducibility) of qualia from the possibility of inverted spectra is given by Shoemaker (e.g., 1996). Qualia must be admitted if inverted spectra are logically possible, even if it turns out that they are not empirically possible. And it seems that even undetectable qualia inversions are logically possible. For, it seems easy to imagine that there could be beings that are very much like us, only simpler. Their color space might be limited to, e.g., light, dark, green, and red. Such a color space might be fully symmetrical, and inversion might then go undetected. We may suppose, however, that the red and green experiences of these beings could be exactly like our own.

A different kind of inversion argument has been advanced by Block (1990), who imagines Inverted Earth – a place where the colors of things are inverted, e.g., grass is red, the sky is yellow, and so on, and where the local language is also inverted with respect to our own. In infancy, your twin was kidnapped from Earth, rendered unconscious, fitted with color inverting devices (too small to be noninvasively detected), and deposited on Inverted Earth. Years later, you undergo the same process. When you are awakened, you look at the sky, and say it is blue. So does your twin. You and your twin share brain states (except for irrelevant differences of memories of events in your respective lives). It thus seems likely that you have the same kinds of qualia. But functionally, you and your twin are inverted; e.g., you will classify the color chip you had in your pocket during the kidnapping differently from your twin, and your twin’s "blue" will mean the same as your word "yellow". (E.g., a paradigm instance of your twin’s "blue" will be the sky, which we are supposing to be yellow, i.e., to be just what you would call "yellow" if you had not had the devices inserted.) What all these suppositions purport to produce is a case of experiential sameness, or sameness of qualia, with different functionality. If this result is accepted, then qualia cannot be held to be reducible to functional predicates, and an adequate account of experience requires more than a description of our verbal and nonverbal responses to various kinds of stimulations.

Zombies

Zombies in Hollywood movies have faces that do not change expression, bodies that move woodenly, and aims from which they are not easily diverted by reasoning. Zombies in philosophy, by contrast, are beings whose behavior is utterly indistinguishable from that of normal humans, but who have no "inner life" at all. There is nothing it is like to be a (philosophical) Zombie. (Hereafter, "Zombie" means philosophical Zombie only.)

Two kinds of Zombies can be distinguished. (1) A-Zombies are somewhat similar to us internally, but not exactly like us internally. Although their behavior is indistinguishable from that of a normal human, the details of their internal mechanisms are just enough different from ours so that the causes of consciousness are missing. Perhaps there are some A-Zombies among us, for it is not clear from their defining description that they are incompatible with the laws of nature of our world (although such incompatibility is not ruled out by their defining description). (2) B-Zombies are exactly similar to us in absolutely every physical respect. But, in the possible world that they occupy, some of the laws of nature are different; and the result is that although their behavior is exactly like ours, the same internal events that produce conscious experiences in our world do not produce any qualia at all in theirs.

In either case, one can argue that Zombies are possible. But their bare possibility requires that there be items that are additional to any behavioral tendencies and even additional to anything in the physical world – items, that is, that can account for the difference between us and Zombies, despite the noted similarities. Moreover, the nature of these items must be such as to account for the difference between presence and absence of an inner life – the difference between its being like something to be moving around in and interacting with the world and its not being like anything at all to be doing exactly the same things. But this is just the role that qualia play; therefore, if Zombies are so much as possible, there are qualia. Naturally, discussions of arguments involving Zombies often turn on disagreements about whether they are so much as possible. It would seem that the claims that (i) one can form a coherent concept of qualia, and (ii) Zombies are logically possible, are equivalent. (see Zombies entry for a fuller discussion)

Representationalism and Determinate Content

Beginning in the mid 1980s, many materialists have espoused some version of representationalism. This view denies that the qualities involved in sensory experiences are actually exemplified by experiences or by any constituent of experiences. Instead, such qualities are merely represented by experiences. It is held that while no brain event could be red, there is no reason why a brain event could not be a representation of a red physical object. Thus, if experiences are held to be essentially representational events, they can be recognized as compatible with materialist assumptions.

Representationalism has naturally called forth arguments for qualia that take the form of denials of the adequacy of regarding experiences as simply representations of qualities. For example, Robinson (1997) has pointed out that we need not be having a red experience ourselves when we think of another’s red experience. In such a case, however, we are representing red (among other things). Thus, an experience of red must contain something more than what is described in classifying it as a representation of red. This ‘something more’ is left out of account by representationalism, but is naturally supplied by a qualia realist view.

Diana Raffman (1995) has pointed out that the limits of our memory prevent us from having concepts that are as determinate as the qualities we can have in our experience. For example, some shade of red – say, red31 – can be distinguished from another shade – say, red 32 – when these are presented together. But we do not have the kind of memory for shades that allows us to distinguish red31 from red32 if these are presented sequentially -- we cannot identify or recognize red31 as such. We can, of course, recognize red31 as red; that is, we can recognize that red31 falls under our representation of generic red. But we have no conceptual representation that is as determinate as our experience can be. Thus, the view that a red experience merely invokes our representational abilities fails to account adequately for the determinateness of the qualitative character of our experiences. Materialists have responded with theories of nonconceptual representational content. It is, however, not clear that materialist theory can provide an interpretation of nonconceptual content that is both adequate to the determinateness of experiences and fully compatible with the avowed commitments of representationalist views. (See the articles mentioned in this section for discussion of this ongoing issue.)

Qualia and Knowledge

Qualia have almost always been assumed to have a specially close relation to knowledge. We may recall, for example, that C. I. Lewis introduced the term "qualia" in connection with the idea of the given. There are, however, several degrees of intimacy that can be distinguished. Roughly speaking, views later on the following list make stronger claims, and are thus more vulnerable to objections, than views listed earlier. (a) Immunity to mistake from false appearance. Qualia do not appear – they are appearances, or the way things (ordinary things such as ties, trains or sewage lagoons) look, sound, smell, etc. Since qualia do not appear, they cannot falsely appear, and this makes them immune to an important kind of error that we can make about ordinary things. (b) Qualia have been taken to be things to which we have privileged access. This means that possessors of qualia are, other things being equal, in a better position than others to know about their own qualia. (c) Qualia have been held to be objects of incorrigible knowledge, that is, objects of beliefs that their possessors could not rationally regard as subject to correction by others (although possibly revisable by the possessors themselves). (d) Beliefs about qualia have been regarded as infallible, i.e., rationally revisable by no one. (e) Beliefs about qualia have been regarded as certain, i.e., not possibly mistaken.

Strong claims about the relation of qualia to knowledge have sometimes led to difficulties about problems such as the speckled hen. A quale one is likely to have when viewing such an object would seem to have parts that are speckles; and if so, there must be some determinate number of speckles. (See Dretske, 1993 for an excellent discussion of the issues raised by this kind of case.) Yet one could surely be wrong that one’s quale has 27 speckles. This fact presents a difficult problem for any view that undergirds the strong connection between knowledge and qualia by, in effect, making the occurrence of qualia itself already a species of believing. If, however, one keeps qualia themselves distinct from knowing them or holding beliefs about them, then one can distinguish kinds of facts about them that are highly knowable (e.g., their color genus) from other kinds of facts that are less knowable (e.g., the number of their speckles, or, in the case of an elliptical afterimage, the ratio of its short axis to its long axis).

Several of the properties just listed are "negative", that is, they attribute lack of openness to various kinds of error on the part of possessors of qualia. There is a corresponding "positive" property that qualia are sometimes held to have; that is, they are sometimes regarded as self-presenting, i.e., as something their possessors must know about. For example, if one is having an orange afterimage, one could hardly be thought of as failing to be aware of the orange character of it. In light of our discussion of the speckled hen, it is clear that views that hold qualia to be self-presenting will be tenable only if some distinction is made between those properties of qualia that are self-presenting, and those properties of qualia that are not.

Three Further Properties

If it is allowed that A’s spectrum is inverted with respect to B’s, then A may still, in one sense, tell B what quale she is having. That is, A can describe her quale using common predicates – e.g., "red" – or by reference to causes, e.g., "what I have when I look at ripe tomatoes". But if the inverted spectrum case has been understood, it will be clear that, in another sense, A cannot convey to B what kind of qualia she has. That is, there is absolutely nothing she could do to satisfy B that her spectrum was or was not inverted with respect to B’s. This property of qualia is often referred to as their ineffability.

Wilfrid Sellars (1963; 1971; 1981) devoted considerable effort to locating the homogeneity of color in a world that physical science represents as made up of discrete neurons and discrete microparticles. Homogeneity is the continuity, or "spread-outness" of color – the property of color expanses that consists in their parts also being color expanses. Homogeneity can be extended to nonvisual properties: foghorn sounds, tastes, smells, pressures, etc. are continuously spread out in time. If, unlike Sellars, one is willing to recognize qualia, it is natural to regard homogeneity as applying to them, i.e., to regard them as having temporal continuity and, in visual and some bodily sensation cases, spatial extensivity.

Finally, let us recall the fact that, as Lewis introduced qualia, qualia were only properties of experience. They are never the same properties as the properties of ordinary things that may have the same name, and that may be causes of qualia. Now, it is not unusual to take experiences to be necessarily conscious. Indeed, in many uses of "experience", "conscious experience" would clearly be a redundant expression. Putting these two points together leads to the conclusion that qualia are necessarily conscious, i.e., that they can be instantiated only in conscious events. This conclusion makes qualia eligible as types of consciousness, and provides a possible answer to the question "What is the difference between conscious events and nonconscious events?" – namely, the answer, that conscious events are those that instantiate qualia.

Natural Dualisms

We have seen how qualia realism diverges from prevailing materialist commitments. Behind materialism, however, is a more general view, namely, a commitment to naturalism. Naturalism is not easily defined, but we can approach the understanding of it in a series of steps. The first step is to recognize as "natural" the world of things recognized by and studied in the sciences. Next, we must realize that science develops and often changes its view both by addition and (perhaps less frequently) by discarding theoretical entities. (Phlogiston and caloric are examples.) Finally, we may say that what is natural is not just what is recognized by the science of today, but rather what can be brought into explanatory unity with that body of future thinking that will be the inheritor, through a continuous line of descent, of today’s sciences.

Naturalism conceived in this way may motivate many materialists, but it does not entail that materialism is true. For the naturalism we have described leaves it open that, from the point of view of a more developed science, present conceptions of material nature will be seen to be inadequate. Naturalism also leaves open the possibility that qualia will be able to be brought into an explanatory unity with a science of the future in a way that will not involve anything plausibly regarded as a reduction of qualia to the physical, on anything like our present conceptions of the "physical". Such views may be called "natural dualisms".

Qualia realists do not claim to know how a future explanatory unity will come about, but there are several ideas that have been offered as pointers to promising developments of this kind. (a) Some thinkers anticipate that consciousness will find an intelligible place in the system of puzzling ideas that occur in interpretations of quantum mechanics. Stapp (1993), for example, focuses on the key role that observation plays in quantum mechanics, and associates observation with consciousness. Penrose (1994) and Hameroff (1994) think that quantum coherence in microtubules will prove essential in understanding consciousness and its contribution to our behavior. Lockwood (1993) sees a parallelism between certain relations among qualia, on the one hand, and certain relations among subsets of quantum-mechanically defined vectors on the other. (b) Differences among qualia can be thought of as differences of information, and the concept of information has rich connections to other concepts in the sciences. It has the advantage of being measurable. Chalmers (1996) has sought a natural dualism that makes information a fundamental property in connecting consciousness with the rest of science. (c) Neutral monism views both consciousness and material nature as ways of regarding a reality that is in itself neither mental nor physical. If the relations of consciousness and material nature to a neutral reality could be adequately made out, their parallel and dependent status might enable them to be embraced by an explanatorily unified theory. Stubenberg (1998) has argued along this line. (d) The property of being a pattern of neural activity has some features that are similar to the homogeneity that we have seen qualia to possess, and it seems that neural events cause qualia in virtue of patterns of neural activity. Causal relations can sometimes be explained, if they can be regarded as consequences of a conservation principle. It is thus conceivable that pattern properties and qualia could be brought into a relation that would be like a conservation principle, and that would be explanatory of their evident causal relations. This line of thought is advanced in Robinson (1999; forthcoming).

Although some of the ideas just summarized have been with us for some time, it should be evident that interest in natural dualism has recently had a renaissance, and that we may expect lively pursuit of developments of theories under this heading.

Annotated Bibliography                                                                                              Top of page