Common Sense about Proper Sensibles and the Senses



COMMON SENSE ABOUT QUALITIES

AND SENSES

Peter W. Ross

Department of Philosophy

CSU, Pomona

3801 W. Temple Ave.

Pomona, CA 91768

Internet: pwross@csupomona.edu

COMMON SENSE ABOUT QUALITIES

AND SENSES

What are the limits on scientific understanding of perceptual states? One common way of addressing this question is to consider the problem of the explanatory gap. Setting aside the issue of the metaphysics of perceptual states, the problem is whether there can be an explanation of the qualitative properties of perceptual states in scientific terms. Some theorists, for example, Ned Block (1995, pp. 380-382) claim that we should characterize the qualitative properties of perceptual states as qualia, that is, as mental properties of perceptual states which are what it is like to be conscious of qualitative properties. These theorists claim that qualia at least resist explanation in scientific terms, and allow that the problem of the explanatory gap might always be a serious epistemological problem. Consequently, qualia mark at least a potential limit on scientific understanding of perceptual states.

Much ingenuity has been devoted to arguments attempting to show that perceptual states do or don't have qualia. Recently, attempts along these lines take up the question of how we distinguish sensory modalities.[i] For example, Block (1996, p. 28) has suggested that distinguishing sensory modalities requires that perceptual states have qualia. Thus the issue of distinguishing the senses might offer the basis for a new argument for qualia.

However, at another extreme, Brian Keeley (2002) argues that our common-sense way of distinguishing the senses in terms of qualitative properties is misguided, and offers a scientific eliminativism about common-sense sensory modalities which avoids appeal to qualitative properties altogether. This eliminativism about the senses could make way for a non-qualitative characterization of perceptual states which avoids the potential limit imposed by qualia.[ii] Considering these opposing claims, we can use the issue of how we distinguish the senses as a vantage point from which to consider what limits, if any, circumscribe scientific understanding of perceptual states.

I'll argue, contrary to Keeley, that common sense is correct that qualitative properties are necessary for distinguishing senses. But, contrary to Block, I'll argue that our common-sense distinction doesn't show that perceptual states have qualia. Thus, a non-qualitative characterization of perceptual states isn't needed to avoid the potential limit on scientific understanding imposed by qualia. For, as far as the issue of distinguishing senses indicates, we need not characterize these qualitative properties as being qualia.

In section I, I'll describe two general strategies for distinguishing senses: a common-sense strategy which relies on qualitative properties and an eliminativist one which denies that qualitative properties are necessary for distinguishing senses. Keeley's (2002) attempt to provide individually necessary and jointly sufficient conditions stated in non-qualitative terms provides an example of the eliminativist option which requires serious consideration. In section II, however, I'll show Keeley fails to demonstrate that qualitative properties are dispensible in distinguishing senses. In section III, I'll offer a minimalist account for distinguishing senses. According to this account, qualitative properties are necessary for distinguishing senses, but the prospects are poor for building on this single necessary condition to establish independently necessary and jointly sufficient conditions. However, I'll point out that we can address the issue of the limits on a scientific understanding of perceptual states while avoiding the problem of providing full-fledged individuation conditions for senses. I'll then show in section IV that we need not characterize the qualitative properties that are necessary for distinguishing senses as being qualia. Consequently, the necessity of qualitative properties in distinguishing senses does not found a new argument for a potential limit on scientific understanding of perceptual states.

I. Strategies for distinguishing senses

I'll first set the stage by describing what I take to be the default view about distinguishing senses; I'll call this the core common-sense view.

Core common-sense view: if a sense is distinct, it produces perceptual states characterizable by a distinctive qualitative property (such as color for vision).[iii]

This view doesn't attempt to give a full account of the distinction among the senses. In its provision of a single necessary condition, the core common-sense view says nothing about whether there are additional necessary conditions and it says nothing at all about sufficient conditions for distinguishing senses. In section III, I'll argue that this single necessary condition is in fact the only plausible necessary or sufficient condition.

Furthermore, the core common-sense view abstracts from the metaphysics of qualitative properties, holding that common sense distinguishes among senses by way of distinctive qualitative properties, whatever the metaphysics of these properties might be (that is, whether they are, for example, properties of physical objects or properties of mental states). This abstraction from the metaphysics of qualitative properties reflects the fact that common sense doesn't offer a claim about this metaphysics.

For example, common sense doesn't claim simply that colors are mind-independent properties of objects. While it might be a bit of common sense that the red of a fire engine is a mind-independent property of the fire engine (although even this doesn't offer a claim about metaphysics since it's not clear what common sense makes of the concept of mind-independent property), it's not clear at all what common sense has to say about the red of an afterimage of the sun. Michael Tye's contention (2000, p. 147) that common sense claims that colors are mind-independent properties is plausible largely because he focuses on the colors of ordinary physical objects. However, if we consider a broader range of colors--including afterimage colors, for example--common sense offers no decision. Thus common sense offers no claim about the metaphysics of colors tout court (which is not to say there aren't theoretical claims about the metaphysics of colors tout court, or that there isn't a theoretical justification for the claim that colors tout court are in fact mind-independent properties of objects).

Thus, the core common-sense view is incomplete in various ways. Nevertheless, it is useful in that it serves to clarify the crucial difference between a common-sense strategy for distinguishing senses which relies on qualitative properties, and an eliminativist strategy which claims that qualitative properties are not necessary. Indeed, the core common-sense view is precisely what eliminativism rejects.

Keeley (2002) proposes eliminativism about common-sense sensory modalities in the spirit of Paul Churchland's eliminativism about propositional attitudes. I take this spirit to be, basically, that a common-sense classificatory scheme for some aspect of mentality should be replaced by a significantly different scientific classificatory scheme. Thus, Churchland, allowing that there's mental representation, rejects the common-sense division of mental representations into sentence-like structures in favor of a scientific classificatory scheme. Similarly, Keeley's target is a common-sense classificatory scheme. Allowing that there are sensory modalities, Keeley rejects the common-sense distinction among senses in terms of qualitative properties and offers a significantly different scientific classificatory scheme.[iv]

Keeley (2002, pp. 6-8) claims that we must make a move to eliminativism in order to specify a scientifically useful concept of sensory modality which neuroethologists and other perceptual scientists can apply to non-human animals. To this end, he proposes four criteria--which are intended to shun qualitative properties--as being independently necessary and jointly sufficient for distinguishing modalities:

(1) distinctive external physical conditions, particularly distinctive stimuli--for example, electromagnetic radiation for vision and pressure waves for hearing (a criterion which Keeley labels physics);

(2) distinctive sense organs, including their connection with the brain (labeled neurobiology);

(3) distinctive abilities "to discriminate behaviorally between stimuli" (labeled behavior); and

(4) distinctive evolutionary or developmental importance of a putative modality (labeled dedication).

To further support his eliminativism, Keeley also denies that the following common-sense criteria are necessary:

(5) distinctive qualitative properties of objects (for example, color for vision, flavor for taste, and so on), the proper sensible criterion,[v] and

(6) distinctive introspectible mental qualitative properties of perceptual states, the distinctive quale criterion.[vi]

I'll argue, however, that Keeley's attempt to avoid an appeal to qualitative properties fails. I'll show that Keeley's behavior criterion makes tacit appeal to qualitative properties and that an appeal to qualitative properties need not be at odds with science in any event, despite Keeley's claim that a scientifically useful concept of modality must be shorn of qualitative specification. To the contrary, Keeley underestimates how effectively science, in particular psychophysics, can shore up the core common-sense view.

II. A defense of the indispensibility of qualitative properties

A. The use of MDS in distinguishing senses

Indeed, psychophysics is the first place that a scientifically informed supporter of the common-sense distinction will look. As Austen Clark (1993, Ch. 5) points out, psychophysicists employ the techniques of multidimensional scaling (MDS) for scientific investigation of the sensory modalities of non-human animals well as human beings.

MDS is a set of statistical techniques which can generate a spatial representation of the relative qualitative similarities among a range of qualitative properties such as colors, where relative qualitative similarities are gauged in non-human animals by techniques such as stimulus generalization tests. (Stimulus generalization refers to a psychophysical effect in which the strength of a conditioned response grades off as the stimulus increasingly differs from the original conditioned stimulus.) And in particular, given data for the relative qualitative similarities among a range of properties, MDS generates dimensions along which qualitative properties are ordered. Thus, using MDS we can map out what is called a psychological quality space for that range of qualitative properties.[vii]

The psychological quality space that's most commonly discussed is the psychological color space with dimensions of hue, saturation, and lightness. This quality space orders colors with respect to relative qualitative similarity along these three dimensions; for example, it represents that: orange is more similar in hue to red than it is to green; that fire engine red and pink are determinates of red which differ with respect to saturation; and that pink and maroon are determinates of red which differ with respect to lightness.

MDS allows us to use behavioral data from human beings or non-human animals to address whether a candidate modality perceives qualitative properties at all, and, if so, whether it has a distinctive qualitative property. In particular, MDS provides the following characterization of qualitative properties: a determinable property D is a qualitative property of perception (thus setting aside properties of states such as pain) if and only if its determinates D1, D2,…Dn can be ordered into a psychological quality space (that is, an N-dimensional order in which the ordering is determined by relative similarity--which can be tested in terms of stimulus generalization) (Clark, 1993, p. 79, pp. 91-94, pp. 117-119). Furthermore, MDS can help us tell whether a candidate modality alone perceives a qualitative property by investigating whether the qualitative property has a distinctive psychological quality space.

Again color provides a useful example. MDS can help us find out whether an animal uses a candidate modality to perceive a range of colors, and whether the modality is distinctive in this respect. First, MDS can help us determine whether an animal perceives a qualitative property with a psychological space akin to the human color space (where by "akin" I mean to include psychological spaces with greater or fewer dimensions than the human color space, but which nevertheless overlap it--such as a one-dimensional grey-scale space).

Then, if this psychophysical investigation indicates that the candidate modality perceives a qualitative property with a kindred psychological space, and this indication is supported by findings from other sciences such as physiology and genetics, we can consider whether the animal perceives a range of colors. (For the relevance of evidence from physiology and genetics, see C. L. Hardin's 1993, pp. 145-154; Evan Thompson's 1995, Ch. 4; and Peter Bradley and Michael Tye's 2001, pp. 471-475. These authors also stress the challenge of attributing color vision to other species, considering our incomplete evidence in these sciences. Even so, they all claim that such attributions are empirically testable given more evidence.)

If this further investigation finds that the animal perceives a range of colors, MDS--along with simpler psychophysical techniques--can help us find out whether the candidate modality alone perceives color.

Furthermore, MDS makes precise what is meant by "a range of colors" so that we can use this concept to support the claim that color is necessary to distinguish vision from other modalities. A range of colors is a range of properties such that: a property is a member of the range if and only if there is a qualitative resemblance route through a color space from it to any other property in the range; a qualitative resemblance route is a span of qualitative properties which is continuous in that points along the span which are sufficiently close are indiscriminable, even though points farther apart are discriminable. So, for example, there is a qualitative resemblance route from red (or magenta) to any other color (determinable or determinate), but there is no such route from red to C-sharp.

The first clue to difficulties with Keeley's eliminativism is provided by his misleading label for criterion (5), the so-called behavior criterion. He discusses this criterion in terms of the use of MDS by psychophysicists. But in this case, this criterion involves not simply behavior but instead the dispositional relation between stimuli, sensory response, and behavior which is the focus of psychophysics. (In fact, considering his labels for the first two criteria, Keeley should have called the third psychophysics.)

By labeling this criterion "behavior," though, Keeley misleadingly suggests that qualitative properties have not entered the picture. But they have, since psychophysics in general, and the use of MDS to generate psychological quality spaces in particular, does not merely study behavior, but rather uses behavioral evidence to study qualitatively characterized sensory responses to stimuli.

1. A methodological worry

Keeley's motivation for subsuming psychophysics under behavior is methodological.[viii] He's interested in a concept of sensory modality which applies to non-human animals. Since it's a tricky question what qualitative properties other animals experience, it might seem methodologically best to simply appeal to behavior.

This methodological worry isn't compelling, however. The best explanation of other people's qualitative similarity classifications is that they have sensory modalities which can be characterized in terms of qualitative properties (either qualitative properties of physical objects, or of mental states, or of both). It's not clear why this consideration shouldn't apply to at least many kinds of animals as well (where relative qualitative similarity classifications are gauged by such methods as stimulus generalization tests)--and particularly if the inference is also supported by findings from other sciences such as physiology and genetics.

The deepest problem with Keeley's eliminativist proposal is that by subsuming psychophysics under behavior, he artificially separates the behavior criterion from the proper sensible and distinctive quale criteria, that is, from the criteria that involve qualitative properties. For if psychophysics can be used to support the claim that qualitative properties are necessary for distinguishing senses, then psychophysics shows that it's false that a scientifically useful concept of sensory modality must be eliminativist.

2. A mismatch between psychophysics and common sense?

In response, Keeley argues that, as it turns out, psychophysics can't be used to support the necessity of qualitative properties in distinguishing senses. Keeley contends that MDS's differentiation of modalities, by differentiating too many, would badly mismatch our common-sense differentiation:

…not only can you not get from red to C-sharp…you also cannot get from red to 'moving left to right across the visual field' (or however motion sensations ought to be described), nor from C-sharp to 'darn that's as loud as a 747 engine from ten feet away' (or however auditory intensity sensations ought to be described). Lacking the appropriate matching relationships, there is no reason to class the "color" and "motion" submodalities of vision as both being submodalities of vision.…Such an account fails to provide the resources for grouping together as "the same modality" sensory qualities that we would intuitively group together. (2002, p. 16; emphasis in original).

Keeley's aim is to show that if psychophysics meshes badly with common sense, there's little that psychophysics can do to support our common-sense distinction. Thus, a scientific concept of sense replaces rather than supports the common-sense concept. And, according to Keeley, since the lack of a qualitative resemblance route cannot plausibly be necessary and sufficient for distinguishing sensory modalities, psychophysics does in fact mesh badly with common sense.

Yet, while Keeley is right that the lack of a qualitative resemblance route cannot plausibly be necessary and sufficient for distinguishing senses, this isn't how we should try to mesh psychophysics and common sense. The core common-sense view is that a qualitative property distinctive of a modality is merely necessary for distinguishing modalities. Accordingly, so long as psychophysics can be used to characterize such distinctive qualitative properties, science supports rather than replaces common sense.

Keeley's first example--that we cannot get from red to 'moving left to right across the visual field'--overlooks the fact that we need not think that every distinction that MDS allows us to draw is sufficient for differentiating modalities. Instead, a qualitative property distinctive of a modality is necessary. In this case, the relevance of MDS is in characterizing qualitative properties such as color which are distinctive to a modality, as opposed to spatial properties which are not.[ix]

The second example--we cannot get from C-sharp to 'darn that's as loud as a 747 engine from ten feet away'--does not take seriously how MDS maps a quality space for sound. The dimensions of this space include pitch and loudness; as well, since different pitch-loudness pairs can combine into chords, auditory quality space can include a number of pitch-loudness pairs up to a maximum complexity beyond which the addition of pitch-loudness pairs produces no discriminable difference (Clark, 1993, pp. 134-138). So, the quality space for sound in fact would provide a qualitative similarity route from (some particular loudness) of C-sharp to the pitch-loudness pairs characterizing some particular complex sound produced by a 747 engine.

Against Clark's general strategy of using psychophysics to distinguish modalities, Keeley claims it "…attributes the wrong modalities to the wrong organisms.…[H]umans are easily capable of discriminating fully charged nine-volt batteries from 'dead' ones, simply by sticking them to the tongue" (2002, p. 16). Here, Keeley's claim is that if the ability to discriminate electrical potency were necessary and sufficient for having an electrical modality, then the mistaken conclusion drawn from psychophysical data would be that "…humans would have an electrical modality!" (2002, p. 16).

But, again, Keeley doesn't consider that the relevance of psychophysics is with regard to characterizing a qualitative property which is distinctive of a modality. It might be that we identify battery potency with our tongues on the basis of a qualitative property (or a combination of qualitative properties) distinctive of familiar senses (such as touch, or a combination of taste and touch); if so, the conclusion that we have an electrical modality would not be plausible. If, however, MDS were to indicate that we identify battery potency on the basis of a qualitative property not perceived by familiar senses and which thus has its own psychological quality space, then the claim that we have an electrical modality would at least become worthy of further investigation. While the indication of a distinctive qualitative property might not provide sufficient evidence for attributing an electrical modality (for it could be an additional distinctive qualitative property perceived by touch, for example) nevertheless this indication is quite plausibly necessary evidence.

Thus, when considered in the right way, the psychophysical evidence does support common sense, despite Keeley's claims to the contrary. He claims that qualitative properties of objects play no role in a useful scientific concept of sensory modality. At the same time, he claims that the behavior criterion is necessary for distinguishing senses. My argument has been that we can strike down the artificial separation of the behavior criterion from the criteria that involve qualitative properties, and use MDS to support a necessary role for distinctive qualitative properties.

It is true that scientists and philosophers are hesitant to attribute conscious qualitative states to other species, but this is typically in recognition of the inadequacies of our current grasp of consciousness. It may seem that these inadequacies enforce a de facto eliminativism. However, a de facto eliminativism doesn't provide reason enough for Keeley's in-principle eliminativism. And in section IV, I'll argue that there's no reason to accept even de facto eliminativism.

However, Keeley follows up these considerations regarding psychophysics with independent arguments against the necessity of each of the proper sensible and distinctive quale criteria.

B. Proper sensibles

The argument against the necessity of the proper sensible criterion is borrowed from Paul Grice's "Some Remarks about the Senses" (1962) and as Keeley deploys it, the argument amounts to the point that proper sensibles help distinguish sensory modalities only because qualia distinguish proper sensibles. Consequently, Keeley claims, "…the cogency of [the proper sensible criterion] rests on the foundation provided by [the distinctive quale criterion]" (2002, p. 22).

This reliance on Grice's argument is odd, however, because Keeley does not even consider what is by now an obvious response, namely that provided by intentionalism about the qualitative aspect of sensory experience.[x] According to intentionalism, this qualitative aspect is characterized in terms of its representation of physical qualitative properties of objects. Thus, the proponent of intentionalism claims, contrary to Grice, that the introspectible qualitative properties of the distinctive quale criterion just are represented physical properties of objects of the sort referred to in the proper sensible criterion.[xi] In this case, the proper sensible criterion provides the foundation, and we can dispense with the distinctive quale criterion.

C. Modalities without qualities?

In section IV, I'll show that Grice's argument for qualia can be countered.[xii] However, in his final attempt to finish off the need for qualitative properties in distinguishing senses, Keeley presents an additional objection, which, although it targets the necessity of the distinctive quale criterion, could easily be modified to target the necessity of the proper sensible criterion as well; indeed, its target is best thought of as the core common-sense view.

There could be, Keeley contends, sensory modalities which produce states that can't be characterized in terms of qualitative properties at all,[xiii] taking as an example the possibility that human beings have a vomeronasal system which perceives pheromones. The upshot is that qualitative properties are not necessary for distinguishing sensory modalities, and the common-sense concept of sensory modality goes by the wayside.

This argument, more than Keeley's other arguments, vividly captures the difference between the common-sense and eliminativist strategies. Since Keeley's necessary and sufficient conditions don't seem to eliminate the human senses of, for example, vision and hearing (and don't seem to be intended to do so), his eliminativism might appear to be an idle view--an eliminativism about common-sense modalities without elimination. Yet with the argument for modalities without qualities, it becomes clear that the main result of rejecting qualitative properties would be to expand the number of modalities, even while recognizing at least some familiar ones in somewhat unfamiliar terms. In addition, it might seem that this argument, as straightforward as it is, avoids tendentious assumptions about the proper use of psychophysics in supporting common sense.

Nevertheless, I'll show that this appearance is misleading. Take the possibility of a human vomeronasal system. If human beings have such a system, then either it produces states that can be ordered into a psychological quality space or it doesn't. It's common ground that for something to be a sensory modality, it must produce perceptual states. Furthermore, it's uncontroversial that perceptual states are a species of mental state. But if states of the vomeronasal system can't be ordered into a psychological quality space, Keeley's contention is vulnerable to the objection that they are not mental states at all but rather non-mental physiological states, and consequently the vomeronasal system isn't a sensory modality. Keeley notes "…the vomeronasal sense seems akin to an olfactory version of blindsight" (2002, p. 25). (Blindsight patients have a scotoma--and so a lack of consciousness--in a portion of the visual field, but nevertheless can make guesses about what's contained in that portion of the visual field with an accuracy above chance.) Although the comparison with blindsight suggests that these states are mental, this comparison is misleading. Though it is uncontroversial that visual states are mental, the worry remains that, if states of the vomeronasal system have no quality space, they are not mental but rather are merely physiological.

However, if there is a psychological quality space for pheromones (involving a qualitative resemblance route along at least one dimension), then the vomeronasal system produces qualitative states.

The point of contention is the first horn of this dilemma, since the interesting question is whether a human vomeronasal system could be a modality without qualities. Considering Keeley's view, an obvious objection to the first horn of this dilemma is that it rests on question begging assumptions. In particular it assumes that perceptual states are a species of mental state, and that what distinguishes perceptual states from non-mental states are qualitative properties.

Again, it is uncontroversial that perceptual states are mental. There are robotic sensing systems, but to the degree that robotic systems perceive, they have mental faculties. Indeed, such robotic systems have a form of non-biological mentality. Moreover, Keeley doesn't dispute this assumption.

The controversial assumption is that what distinguishes perceptual states from non-mental states are qualitative properties. But, presumably, there is some way of distinguishing perceptual states from non-mental physiological states.

Drawing the distinction between perceptual states and non-mental states in terms of qualitative properties is defensible on the following grounds. First, if the states of a physiological system can be ordered into a psychological quality space, they are mental. David Hilbert (1992, p. 357) notes that some invertebrates' abilities to discriminate behaviorally among different color stimuli amount to color tropisms, and do not indicate that these creatures have perceptual states. Hilbert offers the following example: consider an invertebrate with separate response mechanisms which are sensitive to wavelength, so that due to one mechanism it would move toward red light (predominantly long wavelength light) but due to the other it would move away from blue light (predominantly short wavelength light). The creature would discriminate behaviorally among different color stimuli, but this wouldn't show that it has perceptual states of (chromatic) color. Rather, in order to perceive color, "[t]he information acquired through the different photo-receptor mechanisms must interact somehow" (p. 357). Hilbert continues that one indicator of having met this additional necessary condition "…would be the ability to classify stimuli as more or less similar on a dimension different from brightness" (p. 357).

Furthermore, if we describe the biological function of (chromatic) color vision in ecological terms--for example, to promote object detection against a background (see J. D. Mollon, 1989 for a description of this sort)--then the ability to classify stimuli in terms of relative qualitative similarity is plausibly sufficient. For considered from the standpoint of ecology, the function of color vision is fundamentally to identify objects (or creatures) and their states (of ripeness or emotion) (see Thompson, 1995, chs. 4-5, and Gary Hatfield, 1992 for arguments for the ecological approach to the function of color vision). From this standpoint, this function is so directly served by the ability to make qualitative similarity classifications--for example, the ability to easily differentiate contrasting pairs, for instance, red and green, and yellow and blue--that this ability is plausibly sufficient for having (chromatic) color vision. So it's plausible that a state's being characterizable in terms of qualitative properties is sufficient for its being mental; no non-mental states are characterizable in terms of qualitative properties.

Second, if the states of a physiological system are perceptual, they can be ordered into a psychological quality space. Perceptual states are typically characterized in terms of their causal roles as well as in terms of qualitative properties. So, for example, perceptual states are characterized in terms of leading to the production of perceptual beliefs, and other beliefs and other intentional states. Yet, if a state not characterizable in terms of qualitative properties leads to the production of, for example, perceptual beliefs, we are hesitant to call it a perceptual state. And this is reasonable given that it would seem to amount to a power of divination (for a similar claim, see Grice 1962, p. 248).

However, blindsight patients, who are not conscious of qualitative properties within blind regions of their visual fields, give no credence to their guesses about what's contained within these regions. Yet we do attribute a form of sight to these patients with respect to their unconscious regions, despite the lack of credence. But it's plausible that they have sight in these regions only because they can accurately discriminate, for example, shapes, patterns, and locations, which (in vision) require discrimination of (chromatic or non-chromatic) color (Holt, 2002, pp. 20). In other words, if patients were not able to accurately discriminate the qualitative properties of stimuli, and thus were not able to accurately discriminate shapes and so on, there would be no question of attributing sight at all. So a state's being characterizable in terms of qualitative properties is necessary for its being perceptual; no perceptual states are not characterizable in terms of qualitative properties. Taking this point together with the point that no non-mental states are characterizable in terms of qualitative properties, qualitative properties distinguish perceptual states and non-mental states.

Furthermore, in the context of considering our vomeronasal states, Keeley merely asserts that they could be perceptual states not characterizable by qualitative properties. But for this assertion to be plausible, he must address how we distinguish perceptual states from non-mental physiological states in a way that counts vomeronasal states as perceptual, and offer an alternative which is at least as defensible.

Taking a broader perspective, Keeley's contention that a scientifically useful concept of modality avoids qualitative properties has the effect of drawing too sharp of a distinction between common-sense and scientific concepts. These concepts might diverge at points, and where they diverge it can be that common sense goes wrong in ways that science can identify and correct. For example, perhaps there is a human vomeronasal modality. If it is a common-sense assumption that there isn't one, then psychophysics can provide reasons to think that this assumption is false.[xiv] Nevertheless, science supports the common sense view that qualitative properties are necessary for distinguishing senses. So there are some points of agreement as well, and the distinction between scientific and common-sense concepts of sensory modality is much blurrier than Keeley suggests. Agreement with Keeley's optimism about having a scientifically useful concept of sensory modality needn't bring eliminativism.

III. A minimalist account

My positive account of sensory modalities ventures only as far as the core common-sense view, that is: if a sense is distinct, it produces perceptual states characterizable by a distinctive qualitative property. I'll argue that the core common-sense view is correct, but that the prospects are poor for building on it to establish independently necessary and jointly sufficient conditions for distinguishing senses. Yet all is not lost; I'll point out that a full account of individuation conditions is unnecessary for addressing the problem of the limits on a scientific understanding of perceptual states.

My response to Keeley's modalities without qualities argument provides some support to think that the core common-sense view is correct. For according to that response, senses must produce states characterizable in qualitative terms (to be distinguishable from non-mental physiological capacities). But the core common-sense view claims more than this: it claims that distinct senses must produce states characterizable by distinctive qualitative properties.

Necessary conditions for distinguishing senses are conditions where it's not possible for distinct senses to violate the condition. For example: if having distinct physical stimuli is a necessary condition, then it's not possible for distinct senses to violate that condition (and have the same stimulus). Another example: the core common-sense view claims that it's not possible for distinct senses to perceive the same ranges of qualitative properties (that is, the same qualitative determinables).

The core common-sense view presents a plausible necessary condition. Consider a modification of a thought experiment from Grice's (1962).[xv] Martians land on Earth. Oddly, they have two sets of what look like human eyes. These sets of eyes are identical in various respects: they are sensitive to the same physical stimulus, and perceive the same set of properties of objects (and, in fact, they function just like human eyes in these ways as well). Let's assume that the sensory capacities conveyed by these sets of eyes have the same evolutionary or developmental importance as well. It seems that the two sets of eyes are organs of the same sense.

Now let's consider another scenario where we hold constant that two Martian sense organs perceive the same set of qualitative properties, but allow that these sense organs differ, as do the physical stimuli that these organs sense and the evolutionary importance of the sensory capacities of these different organs. Does it makes sense to distinguish these senses on the basis of a combination of distinctive physical stimulus, distinctive sensory organ, and distinctive evolutionary importance, even while they perceive the same ranges of qualitative properties? Or is perceiving the same ranges of qualitative properties sufficient for being the same sense (the contrapositive of the claim that perceiving different ranges of qualitative properties is necessary for distinguishing senses)?

Richard Gray (in press) offers an example relevant to this question: imagine a creature that perceives radiant heat by way of two infra-red sensitive pits located between its nose and eyes, and kinetic heat by contact with its body (excluding the infra-red sensitive pits). Furthermore, its capacities to sense radiant and kinetic heat (different kinds of energy from the standpoint of physics and so different physical stimuli) have different evolutionary importance (since the pits have evolved to detect prey). Nevertheless, if psychophysicists were to show that the creature's infra-red detecting states could be ordered in a way that best fit the psychological quality space for temperture, it would be plausible that in the cases of both radiant and kinetic heat the creature uses differently evolved organs of a single sense, touch, to perceive temperature.

However, if, instead, psychophysicists were to show that the creature's infra-red detecting states could be ordered in a way that best fit the psychological quality space for (perhaps the black-white dimension of) color, it would be plausible that the creature uses differently evolved organs of a single sense, vision, to perceive different parts of the electromagnetic spectrum. Either way, what's sufficient for being the same sense is the range of qualitative properties perceived--temperature or color. Thus it's plausible that perceiving the same ranges of qualitative properties is sufficient for being the same sense; consequently, it's plausible that perceiving different ranges of qualitative properties is necessary for distinguishing senses.

The creature that Gray describes is in fact the pit viper. As Gray points out, neuroethologists, distinguish the pit viper's pits and eyes as organs of different senses, and he conjectures that this is largely because these organs are used independently and have different evolutionary importance. However, the hypothesized psychophysical evidence is not available. My claim is that if the hypothesized psychophysical evidence we to become available, neuroethologists would, or at least would have good reason to, reclassify the pit viper's senses.

Moreover, this claim is bourne out by considerations that show that classifications of sense organs are derivative of qualitative properties that they are used to sense. In fact, while the core common-sense view presents a plausible necessary condition, there is no other; as it turns out, proposals of additional independently necessary conditions are unpromising.

Take physical stimulus. It seems we should characterize physical stimulus in terms of physics. But in that case, sound waves are a kind of pressure wave that is on the same spectrum as vibrations that can be felt. But then if physical stimulus is independently necessary, hearing and touch are not distinct senses because they have the same physical stimulus. But, I take it, that's not plausible.

A way of getting around this problem is to characterize physical stimuli in terms of qualitative properties--for example, to distinguish a physical stimulus as sound waves because it is in a portion of a spectrum of pressure waves that can be heard. But then qualitative properties are doing all the work, and adding the stimulus condition as independently necessary is superfluous.

The addition of either sense organ or evolutionary importance as an independently necessary condition is similarly vulnerable to the charge of being derivative of qualitative properties. For example, if the core common-sense view is right, we can characterize sense organs in terms of the distinctive qualitative properties of the states they produce. But then distinguishing sensory organs according to modality becomes derivative of qualitative properties. For example, if antennae are used to sense texture, they are tactual organs; if they are used to sense sounds, they are auditory organs. Again, distinctive qualitative properties are doing all the work, and adding the sense organ condition is superfluous.

So it seems that it's at least difficult to get a proposal of multiple independently necessary conditions off the ground.[xvi]

And sufficient conditions are even more obviously difficult to come by. Distinctive qualitative properties aren't sufficient for distinguishing senses: touch is associated with multiple distinctive qualitative properties, for example, temperature, texture, and hardness. But is this sufficient to claim that touch should be divided into multiple modalities? We can stipulate an answer, of course, but it's difficult to know how to do better.

I'm not claiming to have decisively put to rest the idea of formulating necessary and sufficient conditions for distinguishing senses. But the prospects are poor (and, I suspect will get poorer with the bringing of messy pit-viper-type cases to bear on the issue). Thus it's best to avoid--as the core common-sense view does--the problem of offering independently necessary and jointly sufficient conditions.

Furthermore, even while avoiding the problem of full-fledged individuation conditions we can do substantial work by focusing on the issue of the necessity of qualitative properties in distinguishing senses. After all, the basic questions relevant to the problem of the limits on a scientific understanding of perceptual states are these: are qualitative properties necessary for distinguishing senses? If so, are qualia necessary?

So far, I've addressed the first of these two questions: I've argued that qualitative properties are necessary for distinguishing senses, and so Keeley's eliminativism fails. I'll now argue that qualia aren't necessary.

IV. The dispensability of qualia

A. Qualitative properties and consciousness

Because my aim so far has been a defense of the core common-sense view, and this view abstracts from the metaphysics of qualitative properties, the discussion has been neutral on the question of whether qualitative properties are physical qualitative properties of physical objects or mental qualitative properties of mental states. However, I'll now argue that, as far as the issue of distinguishing the senses shows, the qualitative properties necessary for distinguishing senses need not be qualia, that is, they need not be mental qualitative properties that are characterized in a certain way: as what it's like to be conscious of qualitative properties. I'll offer an alternative way of characterizing mental qualitative properties which avoids marking a potential limit on scientific understanding of perceptual states.

The standard way to characterize qualia is as what it's like to be conscious of colors or smells or sounds (or any other qualitative property). This characterization has it that these mental qualitative properties are necessarily conscious--there must be something it's like to have them, where there being something that it's like requires consciousness.

According to the alternative I'm suggesting,[xvii] qualitative properties can be characterized in terms of psychological quality spaces. In this case, a non-human animal has a determinable mental qualitative property just in case MDS tells us that it makes qualitative similarity classifications (as indicated by stimulus generalization tests) which can be ordered into a psychological quality space.

This way of characterizing mental qualitative properties avoids the issue of consciousness. For the mental qualitative properties which we infer on the basis of behavioral evidence are neural properties which need not be conscious. In the right neurophysiological context (so, for example, barring damage resulting in blindsight) these mental qualitative properties are conscious. But with regard to blindsight, we can infer that blindsight patients have mental qualitative properties which process information contained in the blind spots of their visual fields on the basis of behavioral evidence--and in this case, these qualitative properties are not conscious.

An unconscious qualitative property may sound like an oxymoron. After all, proponents of qualia define these mental qualitative properties as being necessarily conscious. However, in providing an alternative, the proposal is to characterize mental qualitative properties in terms of psychological quality spaces rather than in terms of consciousness. As David Rosenthal (1991 and 2005) argues, the idea the mental qualitative properties are necessarily conscious is unmotivated. Even common sense allows that mental qualitative properties need not be conscious (for example, common sense allows that a headache that dips out of consciousness doesn't thereby cease to exist).

Moreover, this alternative offers important benefits. It allows us to set aside the question of conscious perception while distinguishing senses in non-human animals. Nevertheless, it retains the common-sense characterization of perceptual states in terms of qualitative properties. Consequently, unlike eliminativism, it doesn't leave us at sea with respect to the issue of how to distinguish perceptual states from non-mental physiological states.

In section II, I noted that, due to the inadequacies of our current grasp of consciousness, scientists and philosophers are hesitant to attribute conscious qualitative states to other species. This hesitancy may seem to amount to a de facto eliminativism. However, if we can distinguish between consciousness and qualitativeness as independent properties of perceptual states, there's no reason to accept eliminativism de facto or otherwise. For by drawing this distinction, we allow that perceptual states can be characterized in terms of qualitative properties--on the basis of behavioral evidence which is best explained by perceptual states which can be ordered into a psychological quality space--independently of consciousness.

B. Grice's visiting Martians

However, there has been recent interest, largely the result of a reconsideration of Grice's "Some Remarks," in whether the problem of distinguishing senses shows that perceptual states have qualia--that is, mental qualitative properties which are what it's like to be conscious of qualitative properties.

Grice argues in "Some Remarks" that qualia are necessary for distinguishing senses.[xviii] I'll counter that the qualitative properties needed to distinguish senses are--so far as Grice shows--physical qualitative properties of physical objects and mental qualitative properties that are not necessarily conscious. Unlike qualia, neither of these properties mark a potential limit on scientific understanding of perceptual states, since when not conscious neither property presents a special difficulty for scientific understanding.

There are, of course, other arguments for the existence of qualia. However, Grice's argument would provide a new and interesting way to make this case. My main objective here is to show that Grice's argument doesn't succeed.

(Although Grice doesn't use the term 'qualia', instead favoring the term 'introspectible character' throughout "Some Remarks", and the using the terms 'phenomenal character', and 'experiential flavor or quality of experience' in his "Retrospective Epilogue," he describes how perceiving with the two sets of eyes are different for the Martians in terms of considering "…whether [perceiving] something to be round [with one set of eyes] was like [perceiving] it to be round [with the other set], or whether when something [appeared] blue to them [with one set] this was like or unlike its [appearing] blue to them [with the other set]" (1962, p. 261). Following the current convention, I'll refer to such mental qualitative properties that are 'what it's like' to be conscious of color as qualia.)

Grice argues for the necessity of qualia on the basis of the Martian thought experiment: Martians have two sets of eyes which are sensitive to the same physical properties, and perceive the same ranges of properties (colors and shapes for example). Due to these similarities, it seems that the two sets of eyes are organs of the same sense. But, Grice continues, if we ask a Martian whether perceiving blue with one set of eyes is like perceiving blue with the other set, the Martian says "There is all the difference in the world" between these perceptual experiences, pointing out a qualitative difference.

Grice claims that we would conclude that the two sets of eyes are of different senses. Since physical stimulus, sense organ, and perceived properties of objects aren't sufficient for distinguishing these different senses, then, Grice argues, another condition--in particular, the distinctive qualia condition--is necessary. In "Some Remarks," Grice's main focus is the relationship between perceived properties of objects and qualia, and in particular to consider whether qualia are dispensable with regard to distinguishing senses; the point of the Martian thought experiment is to demonstrate that they aren't.[xix]

I'll show, however, that, Grice's argument is unsuccessful. Perhaps the Martians distinguish their two sets of eyes because each set perceives a proper sensible (a distinctive qualitative property of physical objects). If so, we can account for "all the difference in the world" in terms of proper sensibles, and qualia aren't needed to distinguish senses. Grice seems to rule this possibility out in the set up for the thought experiment--because supposedly the two sets of eyes perceive the same ranges of properties of objects. But without an independent reason for thinking that the two sets of eyes perceived the same ranges of properties, this claim is question begging.

So Grice needs to provide an independent reason for thinking that we must account for qualitative differences in terms of qualia. But he suggests an independent reason in his comments about the generic resemblance among perceptual states of the same sense.

C. Generic resemblance

Grice claims:

…the way to describe our visual experiences is in terms of how things look to us, and such a description obviously involves the employment of property-words. But in addition to the specific differences between visual experiences, signalized by the various property-words employed, there is a generic resemblance signalized by the use of the word 'look', which differentiates visual from non-visual sense-experience. This resemblance can be noticed and labeled, but perhaps not further described (1962, p. 267).

Thus visual experiences share a generic resemblance; auditory experiences share a generic resemblance; and so on. Grice intends for this point about generic resemblance to support his case for the claim that qualia are necessary for distinguishing senses. I'll offer an interpretation of what Grice has in mind, and state why this consideration doesn't indicate that qualia are necessary.

Visual experiences are distinctive in that they involve color. Consequently a generic resemblance among visual experiences derives from a generic resemblance among colors; broadly speaking, a generic resemblance among the perceptual states of a modality derives from a generic resemblance among determinable (or generic) qualitative properties that are distinctive of that modality. What Grice might have in mind is that each determinable qualitative property distinctive of a sense is a dispositional relation between a range of physical properties of physical objects and a determinable quale.

This proposal is a common version of dispositionalism about qualitative properties, according to which a color, for example, is a disposition of a certain range of physical properties of physical objects to produce color qualia. According to this kind of dispositionalism, we don't attribute qualia to physical objects. Nevertheless, we are aware of them in perception in an indirect way since they determine the qualitative character of properties (such as dispositional colors) which we do attribute to physical objects. Thus, they can make "all the difference in the world" in the perception of the properties of physical objects. Furthermore, qualia are, as Grice says, introspectible. In the case of the visiting Martians, the idea would be that, in addition to a blue quale which the two sets of eyes have in common, there is also a quale distinctive of each set. Supporting this interpretation, Grice suggests that he favors this kind of dispositionalism in his "Retrospective Epilogue," (p. 343).

And perhaps the idea that different senses are distinguished by different determinable qualia is what Block has in mind when he remarks (1996, p. 28) that considerations about distinctions among the senses provide a basis for accepting qualia.

Yet this line of reasoning for qualia is unconvincing. Psychophysics' use of MDS again provides a useful way of looking at the issue. If we can understand a generic resemblance among perceptual states in terms of a determinable qualitative property, we can, in turn, understand this determinable qualitative property in terms of a psychological quality space. Viewed this way, generic resemblance among visual states amounts to our recognition that there is a qualitative resemblance route through the psychological color space from any color determinate to any other color, but not to a sound. Furthermore, understood this way, generic resemblance "can be noticed and labeled, but perhaps not further described" in the sense that our recognition that we can place a determinate color in the psychological color space is a non-inferential recognition that we can't redescribe in other terms (at least not in other non-scientific terms).

But, if this interpretation of Grice is along the right lines, generic resemblance provides no reason to think that perceptual states have qualia. We do think of psychological color spaces in terms of qualitative properties of which we are conscious. But this by itself doesn't demonstrate that mental qualitative properties must be conscious. Instead, it could be that a better way of characterizing mental qualitative properties is as properties are inferred as the best explanation of behavioral evidence, whether they are conscious or not. So an understanding of generic resemblance in terms of a psychological quality space offers an opportunity to avoid qualia, rather than forcing us to accept them.

To decide whether perceptual states have qualia, we should set aside the issue of how we distinguish sensory modalities, and directly address the question of whether mental qualitative properties are necessarily conscious. For, even if we assume that the core common-sense view is correct, support for qualia requires the additional claim that mental qualitative properties are necessarily conscious. Thus, while a consideration of how we distinguish the senses tells against eliminativism, it leaves us where we started with respect to the question of qualia. Instead, the issue of qualia requires that we directly address the problem of the metaphysics and epistemology of qualitative properties such as color. But this is a discussion, as Grice (1962, p. 267) puts it, "for which at the moment I have neither time nor heart."[xx]

NOTES

REFERENCES

Author's papers (2001), (manuscript)

Block, Ned (1995). "On a Confusion about a Function of

Consciousness," Behavioral and Brain Processes, Vol. 18, No. 2 ( 1995): 227-287. Reprinted in The Nature of Consciousness: Philosophical Debates, eds. Ned Block, Owen Flanagan, and Güven Güzeldere. Cambridge, Mass.: The MIT Press, 1997, 375-415. Page numbers refer to reprint.

Block, Ned (1996). "Mental Paint and Mental Latex," in

Philosophical Issues, 7: Perception, ed. Enrique Villanueva. Atascadero, Cal.: Ridgeview Publishing Company, 19-49.

Bradley, Peter and Michael Tye (2001). "Of Colors,

Kestrals, Caterpillars, and Leaves," Journal of Philosophy, Vol. 98, No. 9 (September 2001): 469-487.

Clark, Austen (1993). Sensory Qualities. Oxford:

Clarendon Press.

Cytowic, Richard E. (1995). "Synaesthesia: phenomenology

and neuropsychology," Psyche, Vol. 2. Reprinted in Synaesthesia: Classic and Contemporary Papers, eds. Simon Baron-Cohen and John E. Harrison. Oxford: Blackwell Publishers Ltd., 1997, 17-39. Page numbers refer to reprint.

Gray, Richard (in press). "On the Concept of a Sense,"

Synthese.

Grice, H. P. (1962). "Some Remarks about the Senses," in

Analytical Philosophy, First Series, ed. R. J. Butler. Oxford: Blackwell, pp. 133-153. Reprinted in his Studies in the Way of Words, Cambridge, Mass.: Harvard University Press, pp. 248-268. Page numbers refer to reprint.

Grice, H. P. (1989a). "Retrospective Epilogue," in his

(1989b), 339-385.

Grice, H. P. (1989b). Studies in the Way of Words.

Cambridge, Mass.: Harvard University Press.

Hardin, C. L. (1993). Color for Philosophers: Unweaving

the Rainbow, Expanded edition. Indianapolis: Hackett.

Hatfield, Gary (1992). "Color Perception and Neural Encoding:

Does Metameric Matching Entail a Loss of Information?" in Proceedings of the 1992 Biennial Meeting of the Philosophy of Science Association, Vol. 1, Contributed Papers, eds. David Hull, Micky Forbes, and Kathleen Okruhlik, 492-504.

Hilbert, David R. (1992). "What Is Color Vision?" Philosophical

Studies, Vol. 68, No. 3 (December 1992): 351-370.

Holt, Jason (2002). Blindsight and the Nature of Consciousness.

Peterborough, Ontario: Broadview Press.

Keeley, Brian (2002). "Making Sense of the Senses:

Individuating Modalities in Humans and Other Animals," Journal of Philosophy, Vol. 99, No. 1 (January 2002): 5-28.

Kind, Amy (2003). "What's So Transparent About

Transparency?" Philosophical Studies, Vol. 115, No. ?, (?? 2003): 225-244.

McIver Lopes, Dominic M. (2000). "What Is It Like to See

with Your Ears? The Representational Theory of Mind," Philosophy and Phenomenological Research, Vol. 60, No. 2 (March 2000): 439-453.

Mollon, J. D. (1989). "'Tho' She Kneel'd in That Place Where

They Grew…': The Uses and Origins of Primate Colour Vision," Journal of Experimental Biology, Vol. 146 (September 1989): 21-38.

Noë, Alva (2002). "On What We See," Pacific Philosophical

Quarterly, Vol. 83, No. 1 (March 2002): 57-80.

Nudds, Matthew (2004). "The Significance of the Senses,"

Proceedings of the Aristotelian Society, Vol. 104, Part 1: 31-51.

Sorabji, Richard (1971). "Aristotle on Demarcating the Five

Senses," The Philosophical Review, Vol. 80, No. 1 (January 1971): 55-79.

Rosenthal, David M. (1991). "The Independence of

Consciousness and Sensory Quality," in Philosophical Issues, 1: Consciousness, ed. Enrique Villanueva. Atascadero, Cal.: Ridgeview Publishing Company, 15-36.

Rosenthal, David M. (2005). "Sensory Qualities,

Consciousness, and Perception," in his Consciousness and Mind. Oxford: Clarendon Press.

Thompson, Evan (1995). Colour Vision: A Study in Cognitive

Science and the Philosophy of Perception. London: Routledge.

Tye, Michael (2000). Consciousness, Color, and Content.

Cambridge, Mass.: The MIT Press.

-----------------------

[i] In the last few years there has been growing interest in this issue; recent papers largely or exclusively devoted to it include Dominic M. McIver Lopes (2000), (author's paper [2001]), Brian Keeley (2002), Alva Noë (2002), Matthew Nudds (2004), and Richard Gray (in press).

[ii] Keeley (2002, p. 27) allows that perceptual states might have qualia. Thus Keeley's view isn't presented as an attempt to avoid qualia as such. Instead it presented an attempt to avoid qualia in the provision of individuation conditions for senses, which, according to Keeley, is necessary for the concept of sense to be scientifically useful.

[iii] It might seem that synaesthesia shows that there are no such distinctive qualitative properties. Synaesthesia is a rare condition where a property distinctive of one modality supposedly is perceived through another modality; for example, in one manifestation, colors supposedly are heard. If colors really can be heard (and sounds can be perceived by a modality other than hearing, and so on), then it seems that the core common-sense view is false.

It's difficult to tell, though, if synaesthesia really does indicate that color is not distinctive of vision, or if, instead, certain colors are merely associated with certain sounds. A suggestion in favor of mere association is that particular color-sound pairings are idiosyncratic. Richard E. Cytowic points out, "Scriabin and Rimsky-Korsakov, for example, disagreed on the color of given notes and musical keys" (1997, p. 26). Admittedly, this point is not decisive. But, as well, our current understanding of synaesthesia is too murky to decisively reject the core common-sense view.

[iv] Although Churchland assumes that common sense models mental representation on linguistic representation, this assumption is controversial. By contrast, Keeley's assumption that common sense divides sensory modalities according to qualitative properties is not controversial.

[v] Keeley calls this the proper object criterion; I'll use the label "proper sensibles" to underscore that they are properties not objects. As Keeley notes, the idea of distinguishing modalities on the basis of distinctive qualitative properties of objects goes back to Aristotle. For historical background, see Richard Sorabji (1971).

[vi] Keeley calls this the sensation criterion.

[vii] For an extremely clear and detailed description of MDS, see Austen Clark (1993) pp. 76-101 and pp. 210-221.

[viii] Keeley's (2002) suggests this, and he's confirmed this in conversation.

[ix] Of course, that spatial properties are not distinctive is determined by the fact that they can be perceived by more than one modality, each of which perceives a distinctive qualitative property. Since qualitative properties can be characterized in terms of psychological quality spaces, but this characterization doesn't require reference to any particular sense, there is no danger of circularity.

[x] This reliance is also odd because Grice's aim is just the opposite of Keeley's--Grice wants to show that qualia are indispensable in distinguishing modalities. But it turns out that the discussion of Grice is superfluous since, as I'll discuss next, Keeley offers another argument, and this further argument can be used to reject the necessity of either the distinctive quale criterion or the proper sensible criterion.

[xi] A possible reason that Keeley doesn't consider intentionalism is that he might consider physical properties in general as being physical stimuli; also he might consider qualitative properties in general to be qualia. If so, we can see why Grice's idea that physical properties are qualitative only in virtue of (perhaps dispositional relations with) qualia would be appealing. Furthermore, we can see why the issue of physical properties being qualitative independently of qualia--the intentionalist view--would never come up.

[xii] See [author's paper] for an intentionalist response to Grice's argument. In section IV, my response appeals to mental qualitative properties, and so is not intentionalist; however, I do not characterize mental qualitative properties as qualia, and so still side with intentionalism against qualia.

[xiii] In this argument Keeley's target is qualia, which, following the typical characterization, he understands to be necessarily conscious. I'll take up the question of whether mental qualitative properties are necessarily conscious in section IV. In the meanwhile I'll respond to Keeley's argument in a way that doesn't assume that mental qualitative properties are necessarily conscious.

[xiv] So science could correct common sense with respect to what Nudds calls the counting question, "that is, why do we have five senses?" (2004, p. 31). In the case of a vomeronasal system, science could help to show that the underlying assumption of the counting question--that human beings have five senses--is false.

[xv] I've modified the set up of the thought experiment: Grice's thought experiment assumes a difference in senses for which a difference in qualitative properties is necessary. By contrast, I'll argue for the contrapositive: a sameness in senses for which sameness in qualitative properties is sufficient.

More importantly, I'll modify how the thought experiment is used: in arguing for the core common-sense view, I'm making no assumption as to whether qualitative properties are properties of physical objects or of mental states. By contrast, Grice uses the thought experiment to argue for the necessity of qualia in distinguishing senses.

[xvi] In [author's paper] my idea was to offer disjunctively necessary conditions (as I'll suggest that Grice's conditions can be interpreted as being). But I don't think that this will work either. Consider the proposal: if senses are distinct, then they have distinctive physical stimuli or distinctive sense organs or distinctive evolutionary history or distinctive qualitative properties. This proposal allows that senses are distinct even if they don't differ with respect to qualitative properties. Now I think that the core common-sense view is correct in claiming that a distinctive qualitative property is necessary for a distinct sense (not that a distinctive qualitative property is just a disjunct in disjunctively necessary conditions).

[xvii] This alternative is described and defended at length by David Rosenthal (2005).

[xviii] The fact that Grice (1962, p. 259) stresses the diaphanousness--what is currently called the transparency--of experience is consistent with his arguing for qualia. For an incisive discussion of the complexities of the relationship between transparency and qualia, see Amy Kind (2003).

[xix] In fact, Grice doesn't intend to propose independently necessary and jointly sufficient conditions for distinguishing senses. At best, Grice might suggest that distinct physical stimuli, sense organs, perceived properties, and qualia are disjunctively necessary for distinct sensory modality. In that case, if a sense x-ing and a sense y-ing don't satisfy any of the disjuncts, then x-ing and y-ing are the same sense, and the Martian thought experiment is intended to show that the distinctive qualia condition is needed as an additional disjunct.

[xx] But see (author's paper [manuscript]) for a development and defense of a proposal about the nature of color, in particular, a version of physicalism which avoids qualia. I owe many thanks to Brian Keeley for helpful comments on a draft of this paper. I also thank my colleagues in the Cal Poly philosophy department for their helpful responses; the audience of a CUNY Graduate School Cognitive Science colloquium, in particular David Rosenthal, Pete Mandik, Doug Meehan, Josh Weisberg, and Tony Dardis; and the Claremont Colleges Work in Progress group, in particular Dion Scott-Kakures, Amy Kind, Peter Graham, and Carrie Figdor. I also owe very many thanks to Richard Gray, with whom I've had many helpful discussions about the issues in this paper.

-----------------------

2

39

41

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download