Wittgenstein and Qualia

Draft of November 6, 2006
Wittgenstein and Qualia
Ned Block
NYU
Abstract
(Wittgenstein, 1968) endorsed one kind of inverted spectrum hypothesis
and rejected another. This paper argues that the kind of inverted spectrum
hypothesis that Wittgenstein endorsed (the “innocuous” inverted spectrum
hypothesis) is the thin end of the wedge that precludes a Wittgensteinian critique
of the kind of inverted spectrum hypothesis he rejected (the “dangerous” kind).
The task of this paper is to explicate the difference between the innocuous and
dangerous scenarios, to give arguments in favor of the coherence of the
dangerous scenario, and to show that the standard arguments to the effect that
the dangerous scenario is impossible are no good. I will also agree with what I
take to be the Wittgensteinian position, that the kind of inverted spectrum
hypothesis Wittgenstein rejected lets qualia in the door, where qualia are (for the
purposes of this paper) qualities of experiential states whose phenomenal
character cannot be expressed in a certain way (to be described) in natural
language.1 Further, I will argue, if there are qualia in this sense, Wittgenstein’s
perspective on the mind is wrongheaded. In other words, I will be saying that
there is a recognizably Wittgensteinian position that is incoherent, undermining
Wittgenstein’s deepest views about the mind.
Two Inverted Spectrum Hypotheses
“Wittgenstein’s Notes for Lectures on “Private Experience” and “Sense
Data”” (Wittgenstein, 1968) were written in English, apparently between 1934
and 1936 (Rhees, 1968). Wittgenstein is notoriously difficult to interpret, even to
the extent that scholars cannot agree whether claims that are clearly formulated
in his writings are being asserted or denied. I will put forward an interpretation of
Wittgenstein’s view of the inverted spectrum hypothesis and of the nature of
sensory experience, not as a proposal that meets the standards of Wittgenstein
scholarship, but rather as a suggestion in a recognizably Wittgensteinian
framework that is worthy of discussion (and refutation) on its own merits. When I
attribute a view to Wittgenstein, you may wish to understand that as
“Wittgenstein-according-to-one-non-expert”.2 The seeming endorsement of the
1
Alex Byrne (Byrne, 2006) usefully discusses a variety of uses of the term ‘qualia’, and
Tim Crane (Crane, 2000) has an useful discussion of the term’s history. According to
Crane, C.S. Peirce used ‘qualia’ in something like its modern sense, but Byrne holds that
C.I. Lewis introduced the term into contemporary philosophy.
2
An ancestor of this paper was written for a conference on the work of Paul Horwich in
Pec May 15-16, 2006. Horwich kindly gave me a chapter on the inverted spectrum from
a draft of a book on Wittgenstein to respond to. In this draft, Horwich also accepts one
kind of inverted spectrum and rejects another.
Wittgenstein and Qualia
possibility or at least coherence of one kind of spectrum inversion (the kind I am
calling “innocuous”) begins in this passage (p. 284):
“Consider this case: someone says ‘I can't understand it, I see everything
red blue today and vice versa.’ We answer ‘it must look queer!’ He says it
does and, e.g., goes on to say how cold the glowing coal looks and how
warm the clear (blue) sky. I think we should under these or similar
circumstances be inclined to say that he saw red what we saw blue. And
again we should say that we know that he means by the words ‘blue’ and
‘red’ what we do as he has always used them as we do.”
In his introductory notes Rush Rhees (Rhees, 1968, p. 274) notes “All that is
printed here is a collection of rough notes or memoranda which Wittgenstein
made for his own use. He would never have published them—he would not even
have had them typed—without revising and rearranging them. Certainly he
would have revised the language.” I will be arguing that the kind of inverted
spectrum hypothesis that Wittgenstein endorses (as possible or at least
coherent) in this passage commits him to something he would not agree with, so
a defender of Wittgenstein might respond that after proper consideration,
Wittgenstein would not have endorsed it. However, I think what Wittgenstein
describes here is obviously coherent. Further, subsequent technological
developments have shown it is possible, even technically feasible now. Colors
are easily reversed in digital television. I myself have appeared, inverted in color,
in an interview on a German television station. And it is feasible right now for
virtual reality goggles to make use of such technology in producing that inversion
experience in a subject. I think most vision scientists would agree that the same
could in principle be accomplished by circuits embedded between the eye and
the brain, although no one knows how to do this now. (I will describe this science
fiction scenario as having “wires crossed” in the visual system.)
Later in these notes (316), Wittgenstein confirms the endorsement and
introduces the version of the inverted spectrum hypothesis that he rejects (what I
will call the “dangerous” type):
We said that there were cases in which we should say that the person
sees green what I see red. Now the question suggests itself: if this can be
so at all, why should it not be always the case? It seems, if once we have
admitted that it can happen under peculiar circumstances, that it may
always happen. But then it is clear that the very idea of seeing red loses
its use if we can never know if the other does not see something utterly
different. So what are we to do: Are we to say that this can only happen
in a limited number of cases? This is a very serious situation.—We
introduced the expression that A sees something else than B and we
mustn’t forget that this had use only under the circumstances under which
we introduced it.3
3
One difference I will not comment on beyond this note is that the first passage discusses
red/blue inversion whereas the second discusses red/green inversion. I will treat the first
passage as if it concerned red and green.
2
Wittgenstein and Qualia
Note the words “We said that there were cases…” suggesting that those
cases are coherently describable and perhaps possible. What is the difference
between the innocuous and dangerous cases? In the innocuous case, colors are
inverted but certain properties of them—warm and cool—are not, and this
happens suddenly to someone whose color experience has been normal before.
So it is detectable because of two significant changes: (1) Things the subject
knows to be red such as fires look green and things the subject knows to be
green such as grass look red and (2) red things suddenly seem cool colored and
green things seems warm colored. The dangerous scenario is widespread and is
not behaviorally detectable in either of these ways.
(Note that I am assuming color realism, i.e. that things often have
objective colors, for example that fire hydrants really are red and grass really is
green. This position contrasts with error theories that say that nothing really is
colored but rather color involves an erroneous projection of mental properties
onto the world. (See two joint papers by Paul Boghossian and David Velleman
(Boghossian & Velleman, 1989, 1991).) I will not be able to discuss error
theories further in this paper.)
There is an immediate problem in figuring out what is supposed to be
happening in the dangerous scenario. The most straightforward way for the
innocuous case to become widespread would be if the odd thing that happened
to the subject of the innocuous inverted spectrum scenario simply happened
repeatedly, the result being many inverted people who knew that red things
looked green and green things looked red and also saw red things as coolcolored and blue things as warm-colored, making the widespread inversion
behaviorally detectable. But that is not what Wittgenstein is imagining in the
dangerous scenario, since he takes the dangerous scenario to be one that is
behaviorally undetectable. He says about the dangerous scenario that “we can
never know if the other does not see something utterly different.” Perhaps what
he has in mind is that the odd thing that happened to the subject in the innocuous
scenario happened at birth. Then one might suppose that inversion is
widespread but not detectable because the inverted people have not
experienced a change and they have learned to use color words and the word
‘cool’ in conjunction with the color experiences they get on seeing, e.g. grass. So
one might take the dangerous scenario to be one in which warm and cool are
supposed to be inverted along with the colors themselves (Block, 1990; Levine,
1991).
Intrasubjective vs Intersubjective
Philosophers have made much of the difference between an
intrasubjective spectrum inversion--in which a person at one time is said to be
inverted with respect to the same person at another time--and an intersubjective
spectrum inversion in which one person is said to be inverted with respect to
another. The key dialectical difference can be explained as follows. Suppose we
have a pair of identical twins at birth. One of them has the wires crossed in his
visual system whereas the other does not. The twins are raised normally and
acquire color terminology in the normal way. On the point of view that I favor, we
3
Wittgenstein and Qualia
may suppose that it is possible that the way red things (which they agree are red)
look to one twin is the same as the way green things (which they agree are
green) look to the other. Now a vulnerability in this line of thought stems from the
following objection (Block, 1990, 1994; Harman, 1990).
Notice that it is not possible that the brain state that one twin has when he
sees things they both call "red" is exactly the same as the brain state that the
other twin has when he sees things they both call "green". At least, the total brain
states can't be the same, since the first causes the subject to say "Its red", and to
classify what he is seeing as the same color as blood and fire hydrants, whereas
the second causes the other twin to say "Its green", and to classify what he is
seeing with grass and Granny Smith apples. Suppose that the color-relevant
brain state that one twin has when he sees red things and that the other has
when he sees green things is X-oscillations in area V4, whereas the colorrelevant brain reaction in the first twin to green and to the second twin to red is Y
oscillations in area V4. The objector can say that phenomenal properties should
not be thought to be based in brain states that are quite so "localized" as Xoscillations in V-4 or Y-oscillations in V-4, but rather color experience should be
seen as based in more holistic brain states that include the brain bases of
reporting and classification behavior. (I am assuming agreement on some form or
other of physicalism or physicalistic functionalism.) Thus the objector whom I am
thinking of will want to say that one twin’s holistic brain state that includes Xoscillations and the other twin’s holistic brain state that includes Y-oscillations are
just alternative realizations of the same experiential state: that experiential
state has a disjunctive realization. So the fact that red things cause Xoscillations in one twin but Y-oscillations in the other doesn't show that their
experiences are inverted.
The defender of the claim that inverted spectra are possible (and
coherently describable) can point out that when something looks the way red
things look to one twin, he has X-oscillations, whereas when something looks the
way green things look to him, he has Y-oscillations, and conversely for the other
twin, and so the difference in the phenomenal aspect of color experience
corresponds to a local brain state difference. But the objector can parry by
pointing out that the experiential difference has only been demonstrated intrasubjectively, keeping the larger brain state that specifies the roles of Xoscillations in classifying things constant. The X to Y difference in each subject
is a color experience difference, but this gives us no knowledge of cross-person
comparisons. The objector can insist on typing brain states for inter-personal
comparisons holistically. And most friends of the inverted spectrum are in a poor
position to insist on typing experiential states locally rather than holistically, given
that they normally emphasize the "explanatory gap", the fact that there is nothing
known about the brain that can adequately explain the facts of experience. So
the friend of the inverted spectrum is in no position to insist on local physiological
individuation of qualia. At this stage, the debate seems a standoff, and that is
4
Wittgenstein and Qualia
where the intrasubjective inverted spectrum comes into the picture.4
The intrasubjective inverted spectrum scenario can be seen as a way for
the defender of the inverted spectrum to avoid this response, for if the change
happens in the life of an individual person, we have introspective and behavioral
evidence of an inversion and do not have to ask the question of whether brain
states should be thought of in a localistic or holistic manner (Block, 1990, 1994;
Shoemaker, 1982).
I agree with this superiority of the intrasubjective inverted spectrum
scenario, and will be pursing it in detail later. But for now, my point is that the
intra vs inter difference is not directly involved in the difference between the
innocuous and the dangerous inverted spectrum. Note that it is clear from the
first passage that I quoted that for the kind of inverted spectrum that Wittgenstein
is willing to accept, he accepts both intrasubjective and intersubjective versions
of the innocuous scenario. The individual he describes sees everything
differently today than yesterday—that’s the intrasubjective version. And he also
endorses the intersubjective version in saying “I think we should under these or
similar circumstances be inclined to say that he saw red what we saw blue.” So
it does not appear to be the intrasubjective/intersubjective difference that is at
issue as between the innocuous and dangerous scenarios.
Qualia
So what is the difference between the “innocuous” inverted spectrum
case, the one Wittgenstein regards as coherent and maybe possible, and the
“dangerous” case, the one that he rejects? One difference is that the innocuous
case occurs suddenly and “under peculiar circumstances:” the subject agrees
that fire and the sky now “look queer”; whereas in the dangerous case “it may
always happen”, even not under peculiar circumstances. So they differ in how
peculiar and how widespread they are. Another difference is that the innocuous
case is behaviorally detectable, whereas it seems that what Wittgenstein is
imagining in the dangerous case is not behaviorally detectable.
I do not think that the underlying issue that concerns Wittgenstein is
primarily any of these things. Rather, I think the underlying issue is the existence
of qualia, features of experience that are not expressible in ordinary language in
a canonical way. By a canonical way I mean that, for example, one cannot
express what red things look like colorwise in normal circumstances using terms
like “looks red”, since things that look the way red things look to me (in normal
circumstances) may look to you the way green things look to me. And I think the
crucial difference between the innocuous and dangerous cases that is relevant to
qualia, is whether normal people (or of one normal person at different times) in
normal circumstances can be said to be color inverted or shifted with respect to
4
It should also be said that although some kinds of visual experience are localized in the
brain, notably experience as of a face or as of motion, the same cannot be said for color.
Color appears to be represented at a number of occipital sites, mainly in V4 and V8.
5
Wittgenstein and Qualia
one another5. In the innocuous case, the color inverted person is abnormal. But
if “one section of mankind has one sensation of red and another section another,”
(Wittgenstein, 1958, Section 272, p. 95) then it would seem that normal people
can have inverted color experience with respect to one another, and that is the
source of the “very serious situation”. Here is why.
It is incoherent to suppose that there are normal people in normal viewing
circumstances for whom red things look green. To see this, suppose you are
one of those putatively normal people for whom red things always look green. I
say to you that whereas red things look red to me, red things look green to you.
Suppose that you and I agree that there is a color-perception difference between
us, and that the difference can be described by the locution “Things we agree are
red look to you the way things we agree are green look to me.” Still, you can
object to the idea that red things look green to you. You can reply
Who says that red things look green to me? Why are you the one
for whom red things look red? I would rather say that red things
look red to me and green to you!
If we are both normal, you have as good a case as I do. Since we are both
normal, the right response is that neither you nor I have a superior claim to be
the one for whom red things look red and green things look green. The situation
is relevantly symmetrical. What we should say instead is that the way red things
look to me is the same as the way green things look to you, and in allowing that
there are ways things look that cannot be expressed in terms of “looking red” or
“looking green” or any phrase of the form “looking F” where F is a color name, we
step into the realm of qualia. (Given what I mean by ‘qualia’, this is definitional.)
I am making heavy use of the concept of normality, but without saying
what it means. I think we all know well enough what it means, but of course that
stance leaves my position open to alleged ambiguities or obscurities that affect
the argument.
Using the familiar “what is like” terminology (Farrell, 1950; Nagel, 1974),
the point is that if we acknowledge the innocuous inverted spectrum scenario, we
can say that green is what it is like for the abnormal person to see red, whereas
for the normal person, red is what it is like to see red. So the innocuous scenario
does not require us to suppose that there are color experiences that cannot be
canonically expressed in natural language. However, if we allow the coherence
of the dangerous scenario, in which normal perceivers are inverted with respect
to one another, we cannot say of either of them that green is what it is like to see
red. If we acknowledge the coherence of an inverted spectrum in this sense, we
have to agree that no color name expresses what it is like for either one of the
inverted people to see red. I am not saying that what it is like for them to see red
cannot be referred to in English. For example, we can refer to it by saying “What
it is like for that person to see red”. What we cannot find is a color name F, such
that what it is like for one of these people to see red can be expressed in the
5
Shifted spectra could be used to make the same points as made here in terms of inverted
spectra, but I will stick to inverted spectra in this paper. See (Block, 1999).
6
Wittgenstein and Qualia
form “looking F”, and in that sense we can say that the experiential property is an
ineffable quale.
Sydney Shoemaker (Shoemaker, 1982) distinguishes between intentional
and qualitative similarity. If something looks red to you and me, then our
experiences are thereby intentionally similar, but if your spectrum is inverted
relative to mine, our experiences are thereby qualitatively dissimilar. If
something looks red both to you and to me, our experiences thereby have similar
intentional content, but if our spectra are inverted with respect to one another,
they also have dissimilar qualitative contents. Looking red is an intentional
content of color experience, not a qualitative content. For both members of the
inverted pair, red things look red and green things look green. I will argue below
that the upshot is that color language in application both to the outer and the
inner is keyed to intentional contents of experience (Block, 1990). The problem
for Wittgenstein, then, is that the dangerous inverted spectrum hypothesis
mandates qualities of experience that go beyond its intentional properties.
Those who have a “use theory of meaning” that focuses on external rather
than internal uses—and Wittgenstein is certainly the source of and a plausible
candidate for the attribution of this perspective—will find the description I am
giving unacceptable. When Wittgenstein says “But then it is clear that the very
idea of seeing red loses its use if we can never know if the other does not see
something utterly different,” I doubt that he is talking about the use of ‘red’ to
refer to the color of fire hydrants: that use is unaffected by the scenario. The use
that is affected is the use in describing what it is like to see red in Shoemaker’s
qualitative sense.
As (Byrne, 2006) notes, Gottlob Frege (Frege, 1884/1953) held a similar
view:
The word ‘white’ ordinarily makes us think of a certain sensation, which is,
of course, entirely subjective; but even in ordinary everyday speech, it
often bears, I think, an objective sense. When we call snow white, we
mean to refer to an objective quality which we recognize, in ordinary
daylight, by a certain sensation. If the snow is being seen in a coloured
light, we take that into account in our judgement and say, for instance, ‘It
appears red at present, but it is white.’ Even a colour-blind man can speak
of red and green, in spite of the fact that he does not distinguish between
these colors in his sensations; he recognizes the distinction by the fact
that others make it, or perhaps by making a physical experiment. Often,
therefore, a colour word does not signify our subjective sensation, which
we cannot know to agree with anyone else's (for obviously calling things
by the same name does not guarantee as much), but rather an objective
quality. (§26)
And, although it is a bit of a stretch, one can see Wittgenstein’s beetle in the box
passage as endorsing a similar view6.
6
“Suppose everyone had a box with something in it: we call it a ‘beetle’. No one can
look into anyone else's box, and everyone says he knows what a beetle is only by looking
7
Wittgenstein and Qualia
Opponents of there being an intelligible inverted spectrum hypothesis tend
to make the inverted spectrum hypothesis look bad by putting it in terms like
“Red things look green to one person and red to the other” in a scenario where
the two are clearly normal, instead of “things that the two people agree are red
look to one the way things they agree are green look to the other”.
For example, Gilbert Harman (Harman, 1990, p. 47) argues as follows”
Things that look red to Alice look green to Fred, things that
look blue to Alice look orange to Fred, and so on…Now, in the
normal case of perception, there can be no distinction between how
things look and how they are believed to be, since how things look
is given by the content of one’s perceptual representation and in
the normal case one’s perceptual representation is used as one’s
belief about the environment. The hypothesis of the inverted
spectrum objection is that the strawberry looks different in color to
Alice and to Fred. Since everything is supposed to be functioning
in them in the normal way, it follows that they must have different
beliefs about the color of the strawberry. If they had the same
beliefs while having perceptual representations that differed in
content, then at least one of them would have a perceptual
representation that was not functioning as his or her belief about
the color of the strawberry, which is to say that it would not be
functioning in what we are assuming is the normal way.
A further consequence of the inverted spectrum hypothesis
is that, since in the normal case Alice and Fred express their beliefs
about the color of the strawberries and grass by saying “it is red”
and “it is green,” they must mean something different by their color
words. By “red” Fred means the way ripe strawberries look to him.
Since that is the way grass looks to Alice, what Fred means by
“red” is what she means by “green.””
Harman goes on: “It is important to see that these really are
consequences of the inverted spectrum hypothesis.” But these are
consequences of an inverted spectrum hypothesis that no proponent should
accept, one that combines the assumption of normality of Alice and Fred with the
claim that their experiential contents can be wholly expressed in terms of phrases
like “looking red”. Those of us who accept the existence of qualia can say that
Alice and Fred have the same beliefs about the color of the ripe strawberry and
how it looks, namely that it is red and looks red—because those beliefs function
so as to ascribe the same intentional content, which is what is relevant to
perceptual belief. And those of us who accept qualia can say that “the way ripe
at his beetle. — Here it would be quite possible for everyone to have something different
in his box. One might even imagine such a thing constantly changing. — But suppose the
word ‘beetle’ had a use in these people's language? — If so it would not be used as the
name of a thing. The thing in the box has no place in the language game at all; not even
as a something: for the box might even be empty. No, one can ‘divide through’ by the
thing in the box; it cancels out, whatever it is”. (Wittgenstein, 1958, §293).
8
Wittgenstein and Qualia
strawberries look to him” is ambiguous, having a qualitative and an intentional
sense. So Harman’s critique depends on another critique of the
intentional/qualitative distinction and does not stand on its own.
Color Terms Express Intentional not Qualitative Contents
What exactly is the argument that color terminology (‘red’, ‘green’, etc.) is
keyed to intentional contents rather than qualitative contents? In daily life, we do
not distinguish between intentional and qualitative content, just as we do not
normally distinguish between weight and mass or between rest mass and
relativistic mass. Shouldn’t we suppose that there is some sort of indeterminacy,
as with ‘mass’, where ordinary uses of the term ‘mass’ partially denote both, as
argued plausibly by Hartry Field (Field, 1973)? In (Block, 1990), I argued that if
spectrum inversion is actually rife, it would make sense to think of our tacit
semantic policy as one of using color terms as applied to experience to denote
intentional contents of experience, since when we say of someone that the fire
hydrant looks red to him, we often know what the intentional content is, but not
what the qualitative content is. So how can we be understood as attributing a
qualitative content? However, we do not know whether spectrum inversion is
actually rife (cf. (Byrne, 2006), §3.8). Is there an argument from our lack of
knowledge whether or not inversion is rife to the same conclusion?
Yes, the same argument applies. When we say of someone that the fire
hydrant looks red to him, we can know what the intentional content is—say if we
know that conditions are normal and the perceiver’s visual system is normal
(excluding crossed wires), but not what the qualitative content is. If phrases like
“looking red” were intended to apply to qualitative contents, we would have a
vulnerability to widespread error that we do not take ourselves to have. For
example, we assume that we all stop at stop-lights in part because they look red
to us, and no one would take that assumption to be overturned by finding out that
inverted spectra are rife. But if actual inverted spectra undermined such
attributions then our uncertainty about inverted spectra would spread to ordinary
attributions. We would be uncertain whether people stop at stop lights in part
because they look red. But we are not uncertain about that fact, which suggest
that our semantic policy is one that links those attributions to intentional contents
rather than qualitative contents.
I have some agreement with Frege when he says “Often, therefore, a
colour word does not signify our subjective sensation, which we cannot know to
agree with anyone else's.” I would take the “cannot know” to mean in practice
rather than in principle, since I think perceptual neuroscience is making great
strides in that direction. The Fregean point is very easily available to language
users. I became aware of it as a child. And my seven-year old daughter
commented on first hearing the inverted spectrum hypothesis that it explained
why some people didn’t have purple as their favorite color.
Shifted Spectra
There is another argument that favors the same semantic policy, an
argument not in terms of inverted spectra but in terms of shifted spectra.
9
Wittgenstein and Qualia
The first premise is the fact that there is enormous variation in color vision
in the normal population. In a recent article in the Journal of the Optical Society
of America, (Malkoc, Kay, & Webster, 2005, p. 2155) comment: “Subjects with
normal color vision have been previously shown to vary widely in the stimuli they
select for the unique hues and in the focal stimuli they select for basic color terms
Thus a yellow that appears distinctly reddish to one observer might appear
strongly greenish to another.” 7 When they speak of “the stimuli they select for
the unique hues” they mean the stimuli that are supposed to be the best
examples of red, green, blue and yellow as opposed to colors that seem to
subjects to be mixes such as orange, purple, blue-green and yellow-green.
These are mixes in the sense that, for example, orange seems both reddish and
yellowish. In particular, they comment (p. 2156) that “As in previous studies, the
range of variation in the hue settings is pronounced, to the extent that the range
of focal choices for neighboring color terms often overlap. Thus some subject
chose as their best example of orange a stimulus that other subjects selected as
the best example of red, while others selected for orange a stimulus that some
individuals chose for yellow.” (I hope you are surprised!) Using eight color
categories, there was 80% agreement among normal perceivers in normal
circumstances on only two of thirty-four samples. The eight categories were not
fine shades but coarse colors.8
A particularly useful experimental paradigm for assessing individual
differences uses the anomaloscope (devised in the 19th Century by Lord
Rayleigh), in which subjects are asked to make two halves of a screen match in
color, where one half is lit by a mixture of red and green light and the other half is
lit by yellow or orange light. The subjects are asked to adjust the intensities of the
red and green lights on one side so as to match the yellow or orange light on the
other. The result is that when a normal person is satisfied that they match,
another normal person generally sees the lights as different and will adjust the
red/green mixture differently to match. (Neitz, Neitz, & Jacobs, 1993) note that
“People who differ in middle wavelength sensitivity (M) or long wavelength
sensitivity (L) cone pigments disagree in the proportion of the mixture primaries
required” (p. 117). That is, whereas one subject may see the two sides as the
same in color, another subject may see them as different--e.g. one redder than
the other, and that difference is based in differences in the physiology of the
retina.
But how do we know that those who match differently are all normal?
When red and green lights are adjusted to match orange, women tend to
see the men's matches as too green or too red (Neitz & Neitz, 1998).
And a major difference between men and women that at least in part
7
I learned about this paper from : (Cohen, Hardin, & McLaughlin, 2006). The
same points are made on the basis of older evidence in (Block, 1999; Hardin,
1988).
8
There was no more disagreement about the binary hues (purple, blue-green, yellowgreen and orange) than about the unique hues (red, blue, green and yellow).
10
Wittgenstein and Qualia
explains these results is a difference in the underlying physiology of vision.
Let me explain
There are three kinds of cone in the retina that respond to long, medium
and short wave light. The designations "long", "medium" and "short" refer to the
peak sensitivities. For example, the long cones respond most strongly to long
wavelengths but they also respond to medium wavelengths. Two normal
perceivers chosen at random will differ half the time in peak cone sensitivity by 12 nm (nanometers) or more. This is a considerable difference, given that the long
wave and middle wave cones only differ in peak sensitivities by about 25 nm.
Further, there are a number of specific genetic divisions in peak sensitivities in
the population. The most dramatic of these is a 51.5%/48.5% split in the
population of two types of long wave cones that differ by 5-7 nm, roughly 24% of
the difference between the peak sensitivities of long and middle wave cones.
((Neitz & Neitz, 1998)9) This characteristic is sex-linked. The distribution just
mentioned is for men. Women have smaller numbers in the two extreme
categories and a much larger number in between. As a result, the match on the
Rayleigh test (described earlier) “most frequently made by female subjects
occurs where no male matches” ((Neitz & Jacobs, 1986), p. 625).
(Neitz & Neitz, 1998) explain the result as follows. Genes for long and
medium wave pigments are on the X chromosome. Men have a single X
chromosome which is roughly equally likely to be either of the two forms, and
hence they show a matching distribution with two spikes corresponding to the
peak sensitivities of the two kinds of cones. Women have two X chromosomes.
In roughly half the cases, they have the same allele in both chromosomes—in
the other half the alleles are different. That is, a quarter of the cases are XAXA, a
quarter XBXB, and a half are XAXB. In the XAXB case, one gene de-activates the
other. But that happens independently in each cone in the retina, the result being
that the average cell in these women is intermediate between the extreme
values, and so these women have long wave absorption peaks roughly in
between the two groups of men.
Further, variation in peak sensitivities of cones is just one kind of color
vision variation. In addition, the shape of the sensitivity curves vary.
These differences are due to differences in macular pigmentation, which
vary with “both age and degree of skin pigmentation” (Neitz & Jacobs,
1986, p. 624). Hence races that differ in skin pigmentation will differ in
macular pigmentation. There is also considerable variation in amount of
light absorption by pre-retinal structures. And this factor also varies with
age.
I emphasize gender, race and age to stifle the reaction that one group
should be regarded as normal and the others as defective.
The upshot is that normal people vary in their color classifications, and this
difference is at least in part due to variation in phenomenology which is in turn
due to variation in the underlying physiology of color vision.
9
(Neitz et al., 1993) report a figure of 62%/38%.
11
Wittgenstein and Qualia
That is premise 1 of my argument. Premise 2 is that despite variation in
human visual apparatus from person to person, the borders for our application of
color terminology to objects must to some extent be constrained by needs of
communicating and cooperating with other people. Imagine, contrary to fact, that
we were all given rigorous training in the application of accepted color
terminology to the point where variation in application of color terms among us
was minor. Things we almost all categorize as ‘borderline between red and
purple’ will still look quite different to us, just as, in the inverted spectrum
scenario, a fire hydrant that the members of the inverted pair both categorize as
‘red’ looks different them. My premise 2 is that to some extent, that is the way
things are. That is, color terminology is to some extent subject to social and
linguistic constraint.
How do we know that? Normal people see colors as having the same
similarity relations. In the Farnsworth-Munsell 100 hue test, subjects are asked
to arrange 100 chips in a circle with one chip fixed as the starting chip. Normal
subjects make nearly the same arrangement (Hilbert & Kalderon, 2000). One of
the things social and linguistic influence does is to impose a set of categories on
that similarity space. Before the introduction of oranges into England, the color
we call ‘orange’ was considered a shade of ‘red’. (The OED’s first listed use of
‘orange’ as the name of a color is from 1512.) Cultural and linguistic influences
no doubt affect the borders of our color categorizations. And of course culture
also affects our shade categories. I am sure that those who were school age in
the USA after 1949 will agree pretty much on the forty shades of Crayolas
introduced then and in use (with more added and a few subtracted in 1958, 1972,
1990, 1993 and 1998) since then, such as burnt sienna and turquoise blue-despite variation in the way those colors look to us. And that is a shifted
spectrum.
Why is this relevant to the issue of whether color terms express intentional
contents? Given that language is public and used in communicating and
cooperating, and given that except in special experimental setups, we have no
way of tracking individual variation in qualitative content of color perception, the
only rational policy, and the policy I think that we have tacitly followed, is to tie
color terminology to the colors things actually have. We say something looks red
if it is red and everything else we know of is normal. Although there are many
indeterminacies in the actual colors of things, there is also widespread
agreement in part due to commercial enterprises (e.g. the naming of crayon and
paint colors). The color-relevant intentional content of a color experience is the
color our experience represents something to have—that is the content that color
language of the form “looks red” tracks, not qualia.
Behavioral Indistinguishability
One upshot of the line of argument that I have been pursuing is that
(contrary to everything I have read on the inverted spectrum) behavioral
indistinguishability is not relevant to the use of the inverted spectrum hypothesis
in arguing for the existence of qualia. If half of mankind is inverted with respect
to the other half, the inversion shows the existence of qualia even if the inverted
12
Wittgenstein and Qualia
spectrum is behaviorally detectable—unless one half is thereby seen to be
abnormal or unless the behavioral difference undermines the claim of inversion.
To see this, note that the little dialectic I rehearsed earlier about who has
authority over the word ‘red’ would apply to the two normal groups as well as to
two normal individuals. Neither group could pretend to be “the” group for which
red things look red. If it is alleged that for members of the other group, red things
look green, the other group can complain that the situation is relevantly
symmetrical. The key to qualia is whether ‘looking red’ expresses an intentional
content rather than a qualitative content, and for that conclusion, behavioral
indistinguishability does not matter except to the extent that it can reveal that one
group is not normal, or that the groups are not inverted with respect to one
another.
Virtually every argument that inverted spectra are impossible that I have
read or heard appeals to one or another sort of asymmetry in color space.10 (I
am talking about arguments that spectrum inversion is impossible not that it
makes no sense.) For example, Bernard Harrison (Harrison, 1973) argues that
there are more colors between red and green going one way around the color
circle (via blue) than going the other way (via yellow). Another example: dark
yellow is brown whereas dark blue is still blue, so hue is not symmetrical with
regard to brightness. And desaturated red is pink but desaturated green is
green, so hue is not symmetrical with regard to saturation.11 These are not just
differences in categorization, but also differences in similarity. A light and dark
blue look much more similar to one another than do light yellow and dark yellow,
i.e. brown (Hilbert & Kalderon, 2000). Just how significant such asymmetries are
is hotly debated. Stephen Palmer (Palmer, 1999) discusses a far larger number
of ways in which the color space can be mapped onto itself than had appeared in
the prior literature, some of which may avoid such problems. And as
(Shoemaker, 1982) notes, we could always move to an example in which the
color inverted pair see in black and white, imagining black and white inversion
instead.12 That would be enough to disprove functionalism about that person’s
experience.13 In addition, as Shoemaker notes, it is not clear why minor
asymmetries ought to be significant given that we could imagine a slight variant
10
One argument against inverted spectra that is not based on an appeal to asymmetries is
given in (Byrne & Hilbert, 1997). However, that argument is directed towards the
conjunction of the possibility of an inverted s
pectrum together with the thesis that there is a failure of match between the phenomenal
character of a sensation and its intentional content. It is not an argument that purports to
show that inverted spectra are impossible.
11
See (Byrne, 2006; Hilbert & Kalderon, 2000) for more details. A slightly more
accurate description of the point about red and green is that desaturated near-red is pink
but a desaturated near green is near-green.
12
See (Nordby, 1996) for a personal account of achromatic vision and (Byrne, 2006) for
an objection.
13
The first use of an inverted spectrum argument against functionalism was (Block &
Fodor, 1972).
13
Wittgenstein and Qualia
of human color vision in which those asymmetries are absent but color
experience is much the same as ours. But this argument is unsound according
to Daniel Dennett (Dennett, 1991). Dennett notes that red is advancing, warm
and exciting, whereas green is receding, cool and calming. He thinks that such
asymmetries in function are intrinsic to color experience. So Shoemaker’s
imagined race for whom these “minor” asymmetries are ironed out, could not
have color experience much like ours.
But this whole set of controversies can be sidestepped—at least if the
existence of qualia is the issue. For as I have noted, the argument for qualia
depends not whether the inversion is behaviorally undetectable but whether the
groups that are inverted with respect to one another are both normal. All
arguments that an inverted spectrum is impossible that I have seen focus on the
claim that there could not be a behaviorally undetectable inverted spectrum. I
know of no argument that anyone has given that there could not be a widespread
detectable inverted spectrum in which the groups that are inverted with respect to
one another are all normal. But if such an inverted spectrum is possible, then
there are qualia.
In the rest of this paper, I will discuss an argument that an inverted
spectrum is possible. I will focus on an intrasubjective inverted spectrum
because of the advantage I mentioned earlier. I will be considering an inversion
in which any differences among the inverted pair are compatible with rough
functional equivalence. (Earlier versions are in (Block, 1990, 1994).)
The Intrasubjective Inverted Spectrum Argument
I will argue that the innocuous inversion scenario that does not directly
involve qualia leads to the dangerous inversion scenario that does involve qualia.
Let us start with a scenario described in a way that is not (NOT!) acceptable to a
Wittgensteinian. See Figure 1.
14
Wittgenstein and Qualia
Figure 1
Inverted spectrum scenario described in a way that would not be acceptable to
Wittgenstein or any Wittgensteinian. The captions express the way things look to the
subject at each stage in terms of ways of looking, R and G. R is the way the red tomato
looks at Stage 1. G is the way the red tomato looks at Stages 2, 3 and 4.
We start at Stage 1 with you at your 18th birthday, at which time you are a
normal perceiver. You call red things ‘red’ and red things look red to you. You
agree to undergo an experimental color inversion in which a chip is inserted in
your optic nerve which changes signals from red things into the signals that
would have been produced by green things, and so on for other colors.14 I have
been calling this crossing the wires in the visual system. If we were relaxed
about what counted as a visual “input”, we could imagine instead virtual reality
goggles that do this trick without any surgery.
The operation is a success: at Stage 2, you are disposed to call red things
‘green’ and they look green to you, and so on: you see colored things as having
the complementaries of their actual colors. The experimenters publish their
14
This is an oversimplification. There are many different types of “inversion” because
there are many different ways to map the color solid onto itself. See (Palmer, 1999)
15
Wittgenstein and Qualia
results and you get on with your life. At first, you have to fight your tendency to
call red things ‘green’ and your tendency to stop at green lights, to suppose that
red tomatoes are unripe and so forth, but after some years have passed, you
naturally and spontaneously react to colors in the normal way. You unthinkingly
call stop-lights ‘red’ and stop at them. How do we know that the colors have not
reinverted so that the way things look is the same at Stage 1? You remember
very clearly what things used to look like before the operation, and if asked, you
say that everything used to look the complementary color of the way it looks now.
At first, your status as the subject of the inverted spectrum experiment makes
you a celebrity, but after years pass, the fickle public forgets, and you yourself
rarely think about the way things used to look to you. At Stage 3, you are 50,
and you haven’t thought about your operation in many years. Someone says to
you “Aren’t you the person who underwent the experimental spectrum
inversion?” You say “Oh yes, I haven’t thought about that in many years. But I
do clearly recall the look of the ripe tomatoes at my 18th birthday party. They
looked to me then the way grass looks to me now, colorwise.” If asked to paint a
picture of the way things look to you today, you paint the grass green and the sky
blue. If asked to paint a picture of the way things looked to you before the
operation thirty-two years earlier, you paint the grass red and the sky yellow
(Taylor, 1966). Another 10 years pass, during which time no one asks you about
the operation and during which time you don’t think of the days before 18 or any
of the times someone has questioned you about the inversion. At age 60, you
develop amnesia for the period up to age 50. Now, and this is Stage 4, you have
no memories for the period before your operation, nor do you have any
memories for episodes of remembering that period. You are in relevant respects
functionally identical to your 18 year old self, that is, your experience of red
functions in your mental economy at age 60 the way your experience of red
functioned at age 18. However, the way red things look to you at Stage 4 is the
way green things looked to you at Stage 1. I think it makes sense to suppose
that the inversion is not behaviorally detectable, but as I mentioned earlier, that
assumption is not necessary for the argument. If this is to be an argument
against functionalism, Stage 1 and Stage 4 have to be functionally equivalent.
But functional equivalence is compatible with subtle differences that could reveal
that there has been an inversion. However, if the purpose of the argument is to
demonstrate qualia—and that is my main purpose here—the differences could be
more substantial.
At Stage 1, there is a way that red things look, R. At Stage 2, there is a
way that red things look, G, and that G reaction to red things persists through
Stages 3 and 4. Despite the ‘R’s and ‘Gs’ there is, as I will explain, no argument
for qualia yet. However, if it is OK to quantify over ways things look and give
names to them, then we have an argument against functionalism. (I will argue
soon that even if it is not OK to quantify over ways things look, we have an
argument against functionalism, but it will help to see the argument in a starker
form.) Functionalism is the view that what it is like to have an experience is a
matter of its causal relations to sensory inputs, behavioral outputs and other
mental states. The way that red things look does not supervene on functional
16
Wittgenstein and Qualia
organization, since your functional organization is relevantly the same at Stages
1 and 4 but the way red things look differs at the two times, since R differs from
G. Conversely, the experience of a ripe tomato at age 18 and the experience of
an unripe tomato at age 60 are phenomenally the same, colorwise, but you are in
significantly different functional states. That is, ripe tomatoes at age 18 look the
way unripe tomatoes look to you at age 60, but you are in significantly different
functional states. So functionalism fails in both directions, even without qualia.
Of course, I am not saying that Wittgenstein was a functionalist. His
metaphilosophy makes any such attribution suspect. But to the extent that there
is a metaphysical view of the mind that Wittgenstein’s views in philosophy of
mind point to, it is functionalism (if not behaviorism).
Note that the latter argument—that ripe tomatoes at age 18 look the same
to you as unripe tomatoes at age 60 despite the very different functional states
does not depend on any sort of behavioral undetectability of the inversion. For
all the argument requires is sufficient functional difference. The former
argument, the one that compares the way red things look at age 18 with the way
red things look at age 60 is more vulnerable to issues of undetectability. To use
Wittgenstein’s example, if you see red and yellow things as warm and green and
blue things as cool at age 18, with many years of experiencing fire and water,
you also see red and yellow things as warm and green and blue things as cool at
age 60 despite the fact that at age 60 you see red things as green and green
things as red. Whether the tendency to see red things as warm is
environmentally acquired or whether it is innate in the visual system is, I believe,
simply not known at this time. Whether a person who saw red as cool would be
having a different color experience is also not known.
Now it is not hard to see where a Wittgensteinian might object to the line
of argument summarized in Figure 1. The subject at Stage 3 says “I do recall the
look of the ripe tomatoes at my 18th birthday party. They looked to me then the
way grass looks to me now, colorwise.” But what are these ways and why should
we countenance them? And if these suspicious ways are coming in at Stage 3,
presumably they must have been involved at the outset, in the argument at Stage
1. The Wittgensteinian may say that in effect the argument diagrammed in Figure
1 amounts to this, eliminating all but the essentials:
Stage 1: There is a way red things look to the subject, R.
Stage 2: The way red things look to the subject changes to the different G
Stage 3: The way red things look to the subject remains G
Stage 4: The way red things look to the subject remains G, however the
subject is relevantly functionally identical to Stage 1
So relevant functional identity is compatible with difference in the way red
things look.
The problem, according to the Wittgensteinian objection, is the postulation
and naming of the ways things look, even at Stage 1.
However, and this is the crux, there is no need to see the admittedly
provocative talk of R, the way red things look, colorwise, as amounting to
anything more than the innocuous use of color words that Wittgenstein allows.
17
Wittgenstein and Qualia
Consider Stage 2: red things look green. If our subject, the person who has
undergone something “peculiar” is asked how he sees red things, he says he
sees them as green. (Wittgenstein says he says “I see everything red blue today
and vice versa.”) But now if asked how the way he sees things now contrasts
with the way he saw things yesterday, it is perfectly natural for him to say. “I see
everything red as green today, but yesterday I saw everything red as red.” So
we are justified in saying that at Stage 1, he sees red things as red. I know
Wittgensteinians are likely to object to this point, but I don’t see a good ground
for it. If in the peculiar situation, our subject is justified in saying red things
suddenly look green, he is also justified in saying that yesterday red things
looked red. Imagine the subject talking to his doctor where it is better to err on
the side of over-explicitness. The doctor is wondering whether there might have
been something wrong with the patient even at Stage 1. To be absolutely clear,
the patient says that red things look green today but yesterday red things looked
red. And that justifies the descriptions given below in Figure 2 of Stage 1 and
Stage 2.
It would be weird for someone to claim, in non-peculiar circumstances,
that red things look red. What use could it have? But we now live in a postGricean world in which pointlessness cannot be conflated with nonsensicality.
And further, as just mentioned, it would not be pointless for our subject at Stage
2 to note that at Stage 1, red things looked red to him.
18
Wittgenstein and Qualia
Figure 2
The same scenario described in Figure 1, but without explicit mention of the ways things
look or of ‘R’ and ‘G’. The focus is on the colors that things look to have.
And once we have gone this far, it is difficulty to see how to avoid the
descriptions I have given of Stage 3 and 4. If red things look green at Stage 2,
then they still look green at Stages 3 and 4. Here is a summary of the argument.
Stage 1: Red things look red (to the subject).
Stage 2: Red things look green
Stage 3. Red things still look green. The subject says that ripe tomatoes
used to look green, apparently disagreeing with my claim about his
experience. But the apparent disagreement stems from the fact that his
color terminology has changed, and this comes out when he talks about
19
Wittgenstein and Qualia
the colors things look to have now rather than the colors they actually
have now. He has adapted in that he now says the stop light looks red,
even though we can see from other things he says that it looks as green to
him as it did right after the operation. He does not say it looks green
however, because he has adapted his color terminology so as to talk the
way others do. But it is not possible for him to completely match the way
others talk. At Stage 2 he got the colors of things wrong while correctly
reporting how they looked. At Stage 3, his color language has inverted so
that he matches other people on the colors of things, but the effect is to
mislead on how tomatoes actually look to him. He could have kept his
color terminology the way it was, in which case he would have been
following the normal rules for describing how colors look, or he could
invert his color terminology so as to describe things as having colors in
conformity with the way others describe them. He could not do both
because the way things look, colorwise, has changed.
Stage 4. Red things still look green, but the subject is relevantly the same
functionally as Stage 1, so functionalism is false
(White, 1986) has suggested as against Stage 4 that once the amnesia
hits, the subject’s experience instantaneously reinverts. This may be the best
way to save functionalism, but it is hard to see what else can be said in favor of
it.
Note that the force of the argument presented earlier is preserved but
without any quantification over ways or use of ‘R’ or ‘G’.
The Argument for Qualia
I have given a way of conceiving of spectrum inversion, but that certainly
does not show that an inverted spectrum is possible. It is a platitude in
philosophy that one cannot easily move from conceivability to possibility
(Chalmers, 2002; Gendler & Hawthorne, 2002). However, since all the serious
arguments I know of against the possibility of an inverted spectrum depend on
the inversion being behaviorally undetectable, I do not think the transition from
conceivability to possibility is so problematic here.
Our subject at Stage 4 is inverted with respect to Stage 1, but that is still
no argument for qualia, because our subject at Stage 4 is abnormal. After all, he
has undergone an operation in which the wires in his visual system were
“crossed”. And what could be more abnormal than that. But wait! “Crossed” is a
metaphor. The subject at Stage 4 is inverted with respect to Stage 1, but why
suppose it is Stage 4 that is abnormal rather than Stage 1? If we were to
discover that all the rest of us were born with the physiological formation of the
visual system that obtains in Stage 4 and that it is Stage 1 that is “peculiar”, we
would regard Stage 1 as “crossed”. If we were to find out that our subject at
Stage 4 was born with an unusual visual system, we would naturally take the
view that the operation “uncrossed” the wires in his visual system, making him
normal again. The metaphor of crossing and uncrossing is itself keyed to
normality: it is the normal visual system that is “uncrossed”.
20
Wittgenstein and Qualia
But what would we say if half the population has a visual system like
Stage 1 and half like Stage 4? Then we would say that both are normal, that
there are two different varieties of normal visual systems. If the intrasubjective
inversion scenario described above is possible, then so is the scenario just
mentioned in which half the population has a visual system like Stage 1 and half
like Stage 4. What it is like for one sub-population to see red things is what it is
like for the other to see green things, and so on. But neither group can claim that
they, uniquely see red things as red and green things as green, while the other
group sees red things as green and green things as red, for both groups are
normal. So color experience cannot be expressed using the “looks F” form,
where F is a color. So we have an argument for qualia.
Daniel Taylor and Sydney Shoemaker (Shoemaker, 1975; Shoemaker,
1982; Shoemaker, 1996; Taylor, 1966) have argued from the possibility of
intrasubjective inversion to the possibility of intersubjective inversion. They
argue that if Fred undergoes intrasubjective spectrum inversion, then the color
experience of others must be radically different from Fred’s either before his
inversion or after it (or both). I am also arguing from intra to inter, but the intra
and inter are not important, in themselves, for my case. My focus, rather, is on
the normality of both members of the inverted pair. (Another difference: I am
starting with the comparison of “crossed” and “uncrossed” wires in visual systems
and using that to argue for different experiences.)
Wittgensteinians could avoid this conclusion by rejecting the assumption
that what it is like to see something depends on whether wires are “crossed” in
the visual system. That assumption is involved in the reasoning about what to
say if half the population is like Stage 1 and half like Stage 4. That is,
Wittgensteinians could assume that crossed wires have no effect on what it is
like to see colors so long as the crossing is congenital. This view is natural for
representationists like Alex Byrne, Fred Dretske, Bill Lycan, Michael Tye, and
“vehicle externalists” like Alva Noë and Susan Hurley.15 But these are heavy
duty philosophical theories which Wittgensteinians should reject.
Another line of argument to the same effect is the functionalist view
mentioned earlier that the different brain realizations that supposedly determine
the different color experiences are localistically individuated whereas they should
be holistically –i.e. functionally—individuated. However, once the functionalist
concedes the possibility of an intrasubjective inverted spectrum, functionalism is
refuted, so using functionalism against an intersubjective inverted spectrum
would be unconvincing. If the functionalist is to be effective in blocking the
argument from intra to inter, the functionalist will have to do better on refuting the
intra case itself. Further, since crossing wires in the visual system has a color
inverting effect on an adult, it is just a piece of scientifically informed common
sense that it has that effect on a child. This piece of scientifically informed
common sense should not be denied by anyone who rejects philosophical
theorizing and insists on leaving everything as it is.
15
(Byrne, 2001; Dretske, 1995; Hurley, 1998; Lycan, 1996; Noë, 2004; Tye, 1995, 2000)
21
Wittgenstein and Qualia
Another option for the Wittgensteinian is to adopt what (Shoemaker, 1982)
calls the “Frege-Schlick” viewpoint, according to which intrasubjective
comparisons of the qualities of experience make sense, but intersubjective
comparisons do not. Again, this is a heavy-duty theoretical point of view. In
ordinary life, we do make such comparisons. For example, it is a fact known to
many of us that if you have just eaten grapefruit and I have not, some wines will
taste quite different to you than to me. And further, Wittgenstein himself accepts
the intersubjective version of the innocuous inverted spectrum scenario.
It may be said on Wittgenstein’s behalf that in comparing the taste of the
wine between you and me, we are talking about the intentional content of the
taste, not its qualitative content. But that defense would amount to turning
Wittgenstein into a representationist, not something that one who respects
Wittgenstein’s metaphilosophical views should welcome.
Memory
(Dennett, 1988; Dennett, 1991)has emphasized the unreliability of
memory in arguing against inversion scenarios. Figure 3 shows a schematized
Stage 1 at the top (that is the topmost of the 4 schematic depictions). Seeing a
ripe tomato causes the subject to experience red which causes him to say
“Red!”, indicating a normally functioning experience as of red. The second
portion shows the subject at Stage 2 where seeing the red tomato causes an
experience as of green, which causes the subject to say “Green!” The bottom
half of the diagram indicates two different versions of the situation at Stages 3
and 4. According to the possibility that I have been emphasizing (“phenomenal
realism”), at Stages 3 and 4, red things look green. The subject says “Red!”
because there has been a terminological inversion that cancels out the color
perception inversion. But memory skeptics such as Dennett wonder how we can
rule out the case schematized in the last line in which the compensating
inversion occurs earlier than I imagined and the subject misreports his
experience, remembering falsely what red used to look like.
22
Wittgenstein and Qualia
Figure 3
The phenomenal realist contrasted with the memory skeptic. See text.
Dennett’s argument is that the two possibilities represented in the two
scenarios at the bottom half of Figure 3 cannot be distinguished so the distinction
makes no sense. However, he gives no reason to doubt that they can be
distinguished empirically, as indicated in Figure 4.
23
Wittgenstein and Qualia
Figure 4
The bottom two hypotheses from Figure 3 elaborated
Figure 4 presents an implementation of the difference indicated in the
bottom two scenarios of Figure 3. In the top case in Figure 4, the possibility that I
have been emphasizing is elaborated. The subject at Stage 3 sees a ripe
tomato. It looks green. The subject contrasts its looking green with the way ripe
tomatoes used to look. The machinery of this comparison involves a comparator
device which compares the neural implementation of the experience with the
neural implementation of the memory of previous tomato experiences. Of
course, this story is unflinchingly physicalistic, but no more so I think than the
Tuesday New York Times. A Wittgensteinian who wishes to deny this level of
physicalism would have to adopt a revisionary theoretical perspective, something
that a Wittgensteinian should not do.
One version of a Dennettian memory skeptical scenario is indicated at the
bottom of Figure 4. The idea is that the memory is itself inverted, as is the
experience, so when the subject compares them, the subject says the current
experience is not what he remembers. The report is the same but the internal
reality is not. Dennett wonders how these two hypotheses could possibly be
distinguished. Since they cannot be distinguished, he thinks there is no real
24
Wittgenstein and Qualia
difference. My response is that there is no reason to doubt that the normal
methods of science could distinguish them. For example, perhaps it could be
shown that the memory representation does not change in the relevant period.
Dennett might object that one can tell whether the memory representation in the
brain changes only if one can distinguish the memory representation from the
physical basis of the quale itself, whereas his point is that these cannot be
distinguished. But the point of the detail illustrated in Figure 4 is that once one
makes the hypothesis concrete, the claim that the normal methods of science
cannot possibly resolve such issues begins to look like mere skepticism.
My overall argument is this: The kind of color inversion that Wittgenstein
accepts seems to allow a case such as that depicted at Stage 2 in Figure 2 in
which an abnormal person has inverted color experience. But once that
description of Stage 2 is accepted, it is hard to see what is wrong with the
description I gave in Figure 2 of Stage 1. And once those descriptions are
accepted, it is hard to see why we cannot proceed to Stages 3 and 4. And once
we accept that Stage 1 is inverted with respect to Stage 4, it is hard to see why it
could not be that half the population could resemble Stage 1 whereas the other
half resemble Stage 4. And that yields an argument for qualia that does not
depend on behaviorally indistinguishable inversion. But qualia are deeply
incompatible with the Wittgensteinian point of view. The upshot is that
Wittgenstein is right to say “This is a very serious situation.”
References
Block, N. (1990). Inverted Earth. In J. Tomberlin (Ed.), Philosophical
Perspectives 4. Atascadero, CA: Ridgeview.
Block, N. (1994). Qualia. In S. Guttenplan (Ed.), A Companion to Philosophy of
Mind. Oxford: Oxford University Press.
Block, N. (1999). Sexism, Racism, Ageism and the Nature of Consciousness.
Philosophical Topics, 26(1 & 2).
Block, N., & Fodor, J. A. (1972). What psychological states are not. Philosophical
Review, 81, 159-181.
Boghossian, P., & Velleman, J. D. (1989). Color as a secondary quality. Mind,
98, 81-103.
Boghossian, P., & Velleman, J. D. (1991). Physicalist theories of color.
Philosophical Review, 100, 67-106.
Byrne, A. (2001). Intentionalism Defended. The Philosophical Review, 110, 199240.
Byrne, A. (2006). Inverted Qualia. In E. N. Zalta (Ed.), Stanford Encyclopedia of
Philosophy (Vol. Winter 2006 Edition). Stanford, CA: The Metaphysics Research
Lab of the Center for the Study of Language and Information
Byrne, A., & Hilbert, D. R. (1997). Colors and reflectances. In A. Byrne & D. R.
Hilbert (Eds.), Readings on Color, Volume 1: The Philosophy of Color: MIT
Press.
25
Wittgenstein and Qualia
Chalmers, D. (2002). Does Conceivability Entail Possibility. In T. Gendler & J.
Hawthorne (Eds.), Conceivability and Possibility. Oxford: Oxford University
Press.
Cohen, J., Hardin, C. L., & McLaughlin, B. (2006). True colours. Analysis, 66(4),
335-340.
Crane, T. (2000). The Origins of Qualia. In T. Crane & S. Patterson (Eds.),
History of the Mind-Body Problem. London: Routledge.
Dennett, D. (1988). Quining Qualia. In A. Marcel & E. Bisiach (Eds.),
Consciousness in Contemporary Science (pp. 381-414). Oxford: Oxford
University Press.
Dennett, D. (1991). Consciousness Explained. Boston: Little Brown.
Dretske, F. (1995). Naturalizing the Mind. Cambridge MA: MIT Press.
Farrell, B. (1950). Experience. Mind, 59, 170-198.
Field, H. (1973). Theory Change and the Indeterminacy of Reference. Journal of
Philosophy, 70(14), 462-481.
Frege, G. (1884/1953). The Foundations of Arithmetic. Oxford: Blackwell.
Gendler, T., & Hawthorne, J. (2002). Conceivability and Possibility. Oxford:
Oxford University Press.
Hardin, C. L. (1988). Color for Philosophers: Hackett.
Harman, G. (1990). The Intrinsic Quality of Experience. Philosophical
Perspectives, 4, 31-52.
Harrison, B. (1973). Form and Content. Oxford: Blackwell.
Hilbert, D. R., & Kalderon, M. E. (2000). Color and the Inverted Spectrum. In S.
Davis (Ed.), Color Perception: Philosophical, Psychological, Artistic and
Computational Perspectives. Oxford: Oxford University Press.
Hurley, S. (1998). Consciousness in Action. Cambridge, MA: Harvard University
Press.
Levine, J. (1991). Cool Red. Philosophical Psychology, 4, 27-40.
Lycan, W. (1996). Consciousness and Experience. Cambridge, MA: MIT Press.
Malkoc, G., Kay, P., & Webster, M. A. (2005). Variations in normal color vision.
IV. Binary hues and hue scaling. Journal of the Optical Society of America A,
22(10), 2154-2168.
Nagel, T. (1974). What Is it Like to Be a Bat? The Philosophical Review,
LXXXIII(4), 435-450.
Neitz, J., & Jacobs, G. (1986). Polymorphisms of long-wavelength cones in
normal human color vision. Nature, 323, 623-625.
Neitz, J., Neitz, M., & Jacobs, G. (1993). More than three different cone pigments
among people with normal color vision. Vision REsearch, 33(1), 117-122.
Neitz, M., & Neitz, J. (1998). Molecular genetics and the biological basis of color
vision. In W. G. Backhaus, R. Kliegle & J. S. Werner (Eds.), Color Vision:
Perspectives from Different Disciplines. Berlin: DeGreuter.
Noë, A. (2004). Action in Perception. Cambridge MA: MIT Press.
Nordby, K. (1996). Vision in a complete achromat: A personal account, from
http://consc.net/misc/achromat.html
Palmer, S. (1999). Color, Consciousness and the Isomorphism Constraint.
Behavioral and Brain Sciences, 22, 923-943.
26
Wittgenstein and Qualia
Rhees, R. (1968). Notes for Lectures on "Private Experience" and "Sense Data":
Note on the Text. The Philosophical Review, 77(3), 271-275.
Shoemaker, S. (1975). Functionalism and Qualia. Philosophical Studies, 27, 291315.
Shoemaker, S. (1982). The Inverted Spectrum. The Journal of Philosophy, 79,
357-381.
Shoemaker, S. (1996). Intrasubjective/intersubjective. In S. Shoemaker (Ed.),
The first-person perspective and other essays. Cambridge: Cambridge University
Press.
Taylor, D. (1966). The Incommunicability of Content. Mind, 75, 527-541.
Tye, M. (1995). Ten Problems of Consciousness. Cambridge MA: MIT Press.
Tye, M. (2000). Consciousness, Color, and Content: MIT Press.
White, S. (1986). Curse of the Qualia. Synthese, 68, 333-368.
Wittgenstein, L. (1958). Philosophical Investigations, Second Edition. Oxford:
Blackwell.
Wittgenstein, L. (1968). Notes for Lectures on "Private Experience" and "Sense
Data". The Philosophical Review, 77(3), 275-320.
27