Qualia are commonly taken by scientists and other philosophers to comprise the most obdurate obstacle to scientific explanation of the conscious mind. It is commonly maintained that qualia (in their essential nature) cannot be explained by science even in principle. Against this thesis I will argue that in fact we already have perfectly normal scientific explanations of qualia. To be specific, I will focus on the most difficult case, the case of colour qualia. The point of explanation is to provide understanding. We do not worry anymore about what keeps the moon in the sky or what the principle of life is, because we have scientific explanations of these phenomena. If we treat like cases alike, we should realize that we likewise have a scientific understanding of qualia, despite the sophisticated arguments of the friends of qualia. So, unless we are prepared to resume philosophical worries about the moon and life, we should stop worrying the project of understanding consciousness with quandaries about qualia.
I will first present the explanation, and then review current philosophical ideas on the topic by addressing the most prominent objections to the explanation.
1. The Standard Explanation of Colour Qualia
The explanandum: What is to be explained? Given the thickets of philosophical discussion surrounding qualia, the phrase, “explanation of colour qualia” will mean different things to different people. So, to be more definite, I propose to explain why colours look the way they do.Footnote 1 There are lots of colours, and they look lots of ways. For brevity, let us pick a particular colour: yellow will do as well as any other. So, how does yellow look?
I offer the following description of the way yellow looks, hoping that it is accurate, if not complete (we will consider objections to it in due course):
Y1) It is the brightest of the colours. Blue and green and the rest seem intrinsically darker colours. Blue and green can be bright, as in the case of a bright blue sky, or bright blue light, but they purchase their brightness by sacrificing their hue. As blue gets brighter, it becomes less blue, while an equally bright yellow is very yellow. Accordingly, a dark yellow is less yellow than a bright yellow, while a dark blue can be very blue. Likewise, a paradigmatic yellow is brighter than a paradigmatic blue.
Y2) Yellow has specific, qualitative, colour-relationships. It is most similar to orange, and least similar to blue. Given the intervening colours there is a similarity continuum: orange, red, green, purple, blue.
Y3) Yellow is atomic. It is not an obvious compound in the way in which pink or purple are; the first a blend of red and white, the second a blend of red and blue.
Y4) Yellow has a specific chemistry: it can form compounds with some hues but not others. It combines with red to yield orange, or with green to yield chartreuse. But it cannot combine with blue. One cannot see, or imagine, a yellowish blue. (A juxtaposition of tiny blue dots and tiny yellow dots looks green, and mixing blue paint and yellow paint yields green paint. However, green does not look like a compound of yellow and blue. True compounds, like pink and purple, look like their components: pink looks reddish and white-ish, purple looks bluish and reddish. But green does not look blue, and it does not look yellow. Of course, there are such things as blue-greens, and yellow-greens, but green itself is neither yellow nor blue.)
One could say more, but this is enough for our purposes.
The explanans: To explain Y1 through Y4, we need only draw on the standard, textbook theory of colour, which I will abbreviate here. The human retina contains cone cells, which are sensitive not only to the overall brightness of visible light (as are the rod cells), but to its wavelength as well. There are three types of cone cell, each maximally sensitive to a specific wavelength, located at 440, 550, and 570 nanometres, with the corresponding cone types designated “blue,” “green,” and “red,” (B, G, and R).Footnote 2
The output of the cones is then subjected pairwise to opponent processing, whereby two inputs compete, with only the stronger transmitted as output. So where R and G cells provide input to an opponent process, the output is either some level of R activity, or G activity, but not both — even if the original R and G inputs are of comparable strengths. When the inputs are (nearly) equal there is no output.
Cone output is routed into two separate opponent processes, namely R versus G, and B versus the sum of R and G. The double opponent process constitutes a two-dimensional space, with the R←→G axis orthogonal to (independent of) the B←→(R+G) axis. So cone activity yields two independent tug-of-wars, which we can picture with Cartesian coordinates, a dot on each indicating the outcome. If both contests are tied, both dots are at the origin and no colour is seen, only white, black, or gray. But should either contest have a winner, colour is perceived.
Consider the axes independently. If the dot on the R←→G axis is toward the R end, red is seen; if toward the G end, green. The farther toward either end, the stronger the red or green hue. In order to see pure red or green, the dot on the orthogonal B←→(R+G) axis must be at its midpoint; that is, the (R+G)-level and B-level of activity must be sufficiently well balanced that no signal issues from the opponent processing, in other words, B=(R+G). Likewise, if the dot on the B←→(R+G) axis is toward the B end, blue is seen, and if toward the R+G end, yellow is seen, with hue intensity proportional to its distance from the origin, and purity requiring R=G. When neither dot is at midpoint we get compound hues: yellow and red to get orange, yellow and green for chartreuse, blue and red for purple, and blue and green for the aquamarines.
This theory implies Y1-Y4 (here labelled, qua explananda, as E1-E4, and presented in reverse order) as follows:
E4) Yellow has a specific colour chemistry. This follows from the physiological processes of colour vision. Since yellow results from an opponent process with blue, it cannot blend with blue. Not only do we never see such a compound, we cannot even imagine it, because imagining a colour requires auto-stimulation of neural colour processes. On the other hand, since the yellow-blue opponent process is independent from the red-green process, compounds of yellow with red or green may be seen or imagined. These blends are obvious compounds: orange looks both reddish and yellowish, while chartreuse looks greenish-yellow.
E3) Yellow is atomic. This is because yellow is one of the four basic hues created by human colour processing. It is at one extreme of a single opponent process.
E2) Yellow has specific, qualitative colour relationships. It is least like blue, since it is opposed to blue. It is most similar to orange because orange is a compound of yellow itself (everything being self-similar). The nearly equal similarity of red and green with yellow is readily explained by the orthogonality of the red-green axis with the yellow-blue axis: the independence of the red-green process from the yellow-blue process implies that red and green are equally similar (or dissimilar) to yellow. Many find red more similar to yellow than green is. Red and yellow are said to be “warm” or “approaching,” while blue and green are “cool” or “receding.” This is partly a matter of colour association (fire is red, ice is blue, etc.), thus not a matter of colour as such, and partly one of brightness (to be addressed in E1). The lesser similarity of yellow with purple (than with either red or green) is a direct consequence of the fact that purple is a compound of blue, which is most dissimilar from yellow.
E1) Yellow is the brightest of the colours. This is because it results from an unfair opponent process in which the R and G cones gang up on the B cones. It is not surprising that the sum of R and G activity (yielding yellow) is bright, nor that the B activity (yielding blue) only triumphs when the light is low, when both R and G must be weak. So yellow is paradigmatically bright, and blue paradigmatically dark. A surprising consequence is that the brightness of yellow is, in a sense, an illusion born of the nervous system’s confusion of colour with brightness. There is such a thing as dark yellow, but we normally term this “brown.” As Hardin (Reference Hardin1983, p. 559) points out, we can literally see this for ourselves by looking down a tube of flat-black paper at a piece of brightly lit chocolate, which will then look yellow.Footnote 3 Normally chocolate looks brown because it is dark relative to its surround, but this contrast is eliminated by the tube, revealing the hue-identity of yellow and brown. Perhaps you, like me, have previously noted the affinity of yellow and brown, the way they blend seamlessly into one another in autumn leaves, in Van Gogh’s sunflowers, in the golden tones of antique brass, and so on. If we insist that colour (or hue, properly speaking) is independent of brightness, the apparent affinity is actually an identity, and the brightness of yellow an illusion.Footnote 4 Either way, the brightness of yellow is explained.Footnote 5 That many people find red’s similarity with yellow greater than green’s, noted in E2, may likewise be a matter of brightness.
2. Objections and Replies
Objection 1: Qualia are, in their essential nature, ineffable. So whatever is dealt with, hence described, in the scientific explanation is irrelevant. Knowledge of qualia is given immediately (i.e., without mediation of concepts or theory) in experience: to see yellow is necessary and sufficient for knowledge of the yellow quale. To describe what is thereby known only takes us away from its essentially private, subjective nature. Science deals solely with the domain of describable, publicly accessible, objective phenomena, and so cannot deal with qualia (Nagel Reference Nagel1974; Chalmers Reference Chalmers1996, pp. 22-3).
Reply: Suppose qualia are ineffable. Then, since explanation is a verbal, conceptual, and logical matter, the failure of the explanation is trivial: nothing can explain what cannot be conceived or described. Hence science, in principle, cannot explain qualia. So physical science is incomplete. An easy lemma is that physicalism (the thesis that everything real is physical) is false: what science cannot explain, even in principle, cannot be reckoned among the physical. Major philosophical theses effortlessly fall into place. The only work is showing that qualia are, indeed, ineffable — and real.
Aye, there’s the rub! Demonstrating the joint reality-cum-ineffability of qualia is necessarily a sophisticated task. Whereof we cannot speak, we must remain silent — unless we need to say just what it is, and where, and how we know. This explains the gallons of ink spilt by the friends of qualia on their supposedly ineffable topic. Bringing qualia into the debate demands qualia theory: they are directly known in experience and private (or not, see e.g., Dennett Reference Dennett, Marcel and Bisiach1988), subjective and relative (or not, see e.g., Foss Reference Foss1989, Reference Foss1993), intrinsic (or not, see e.g., Dennett, op. cit., or this article, below), directly known (or not, see e.g., Tye Reference Tye2000 or Gibbons Reference Gibbons2005), transparent (or not, see e.g., Tye Reference Tye2000), in the head (or not, see e.g., Byrne and Tye Reference Byrne and Tye2006), physical (or not, see e.g., Aranyosi Reference Aranyosi2003 or Loar Reference Loar2003). … But now, paradoxically, we find qualia copiously conceptualized and described. Thus the ineffability of qualia must be (carefully) qualified. In certain ways we can and do talk about, hence describe and conceptualize, qualia. Otherwise they would have nothing to tell us, philosophically.
But when we consider what is said about qualia by those who suppose them a barrier to scientific understanding, we see that they (explicitly or implicitly) agree with all that is needed for the explanation at hand. Colour qualia are seen (rather than felt, smelled, etc.);Footnote 6 their visibility is sufficient evidence of their reality; they have innumerably different sorts, distinguished by hue, intensity, and brightness (perhaps among others). In particular, the yellow quale is known, and known to exist, by seeing yellow things, is seen to be different from the red quale, even more different from the blue quale, and so on, as in Y1-Y4.
Sophisticated adherents of the ineffability thesis will still hasten to object that the facts just listed omit the essence of qualia, are merely relational, and so on. But this is to shift from the present objection to different ones that will be addressed below.
Objection 2: The explanation transmogrifies the original explanandum, rich and sensuous qualia, into something more suited to its own devices, namely the skeletal colour-logic derived from the explanans itself. It amounts to a classic bait-and-switch manoeuvre: lured by the promise of an explanation of qualia, we get only an explanation of the scientific ersatz, “qualia.”
Reply: This is a traditional objection when science enters a new field (especially when there are psychological motivations to resist a soulless mechanism). We set out to explain the wanderings of Venus, the Goddess of Love, through the heavens, and in the process she becomes a mere chunk of earth careening through space—all of which makes those with pagan sympathies complain that Venus, the Goddess of Love, has not been explained at all, but left behind as we focused instead on celestial physics. But even when this sort of objection is formally correct, it may have no philosophical merit: as seekers of wisdom, we are better off without the Goddess of Love. In general, understanding a phenomenon may require a reconceptualization of not just the original explanans, but of its resident domain. So, too, with qualia.
This first reply is adequate and fair, but overly dramatic. The explanation of colour is humdrum, from a scientific point of view. No innovative principle is required. Explaining colour is more like explaining rain than it is like Newton’s explanation of planetary motions via his novel and counterintuitive three laws. Colour is a phenomenon long felt to need explanation, and after a history of investigations of varying virtue whereby both data and theories were elaborated, including the data of colour-chemistry (Y4),Footnote 7 a mechanism (to put it provocatively) was discovered that markedly increases our understanding of (and hence explains) the phenomenon, and related phenomena besides.Footnote 8 Sure, the phenomenon was defined and redefined during the investigations leading up to the discovery of the explanation — but that is no objection. How else could we possibly proceed? If we are to surpass the ineffability stance, we must somehow describe what it is we want explained, and this in itself may require some (scientific) investigation.
One boring aspect of the philosophy of mind has been the refusal of those who claim the mind is inexplicable to say what they want explained, rather than just pointing at the phenomenon in general. Repeatedly these mysterians employ an ostensive approach, in which we are asked to look at a ripe tomato, or taste it, or smell it, and are given the indexically expressed demand: explain that! (e.g., Chalmers Reference Chalmers1996).Footnote 9 Any attempt to say what “that” might be, is rejected — as Y1-Y4 is rejected in this objection. Often, the mere fact that words fail to convey the experience of seeing blue to the reader (as in the case of Jackson’s Mary, about which more below) is taken to disqualify them. By this logic, all of empirical science would be disqualified, for the mere word “Venus” hardly conveys the planet to the reader either. So a second response to the objection is to maintain the obvious, that Y1-Y4 do capture a significant portion, though not all, of what we find puzzling about the yellow quale.
A third issue needs addressing. The proper precedents of this type of objection share the following logic: when we “measure” the velocity of a bullet by measuring the displacement of a hanging sandbag into which it is shot (and calculating therefrom), we assume principles of classical mechanics (conservation of momentum) in so doing. Taking the calculated velocity as an observation demands circumspection. The observation is “theory-laden” (to use apt, albeit outdated terminology), infected, as it were, with the principles of classical mechanics. Therefore, any explanation of that velocity by classical mechanics is (partially) circular logically and methodologically. If this is the logical deep-structure of the objection, then it is coherent and substantial (if not decisive).
However, it clearly does not apply to the case at hand. As it happens, our explanandum is isolated from the explanans by a historical divide. The facts referred to in Y1-Y4 were ascertained long before the opponent processing theory appeared on the horizon. Of course, this temporal precedence is not itself the issue, but merely an indicator of it: the logical and methodological independence of explanans and explanandum. The explanandum is purely phenomenological in its origins and premises, while the explanans derives solely from observation of the nervous system and the behaviour of perceivers. It doesn’t get any neater than that. Thus our third reply to the objection.
Finally, there is the decisive issue of truth. We could, of course, leave this to the proper scientific authorities. However, I recommend that readers test Y1-Y4 for themselves. Some coloured felt pens, papers, and transparencies are the only supplies needed. Brightness (Y1) and similarity (Y2) orderings can be gotten by arranging coloured papers in various sequences and juxtapositions. By placing coloured transparencies over each other, colour chemistry (Y3 and Y4) can be observed. The viewing-tube permits observation of the hue-identity of brown and yellow. I also recommend the binocular-rivalry experiments of Du Tour. In 1860 he described gluing swatches of different colours back-to-back on a piece of cardboard and placing “the cardboard against my nose in a vertical plane and perpendicular to my face,” and then (by slightly crossing the eyes) letting the two visual patches merge. When I tried this, my results agreed with his, in corroboration of current theory; I can do no better than use his own words to describe my observations using blue and yellow: “I was unable to discern the least tint of green. That single patch I saw sometimes appeared blue, sometimes yellow … sometimes the patch appeared partly blue and yellow” (Wade Reference Wade1998, p. 150). Likewise for red and green. Other hues blended as the theory required.Footnote 10
There is an unfortunate tendency of philosophers, rational beings par excellence, to pay mere lip service to the fact that science is empirical. Its relationship to sensory experience is by no means accidental or eliminable (an issue to which we will return in considering Jackson’s “Mary” objection, below). As practicing scientists know, there is no substitute for seeing the data for oneself. Consequently, if you would learn whether Y1-Y4 really do capture significant truths about the yellow quale, you really must perform the simple experiments outlined — or else defer to those who do. So a fourth and final reply to the objection: look and see for yourself.
Objection 3: The explanation presupposes the identification of seeing qualia with the colour processing of the nervous system. This presupposition begs the question, indeed many important metaphysical questions, which must be answered for the explanation to succeed.
Reply: Treat the identity as merely hypothetical: a scientific hypothesis. Whether or not we should accept the hypothesis depends on its scientific merit, and prominent among its virtues is its explanatory power. We accept the identity of fire with rapid oxidation because it explains the phenomenon of fire, a step commonly termed “inference to the best explanation.” Accordingly, because the identity of neural colour processing with conscious colour processing explains our colour qualia, we have reason to accept it.Footnote 11 So empirical identities do not require independent metaphysical support. To put it in other terms, the explanation presupposes not metaphysical materialism, but rather methodological materialism.
According to methodological materialism (Foss Reference Foss1995, Reference Foss2000), scientific explanations enjoy the right to metaphysical modesty: whatever is methodologically sound (that is, empirically detectable and perchance controllable) may be accepted as real (that is, cited in theory and explanation), even in the absence of an explicit ontology and metaphysics. For example, gravity is explanatorily potent, and in this sense real, even though we do not know the answer to such elementary metaphysical questions as whether gravity is to be reckoned as an effect of mass or one of its properties. Newton thought gravitation an accidental property of mass (which he took to be equivalent to physical substance), while taking inertia as essential (along with extension, hardness, impenetrability, and inertia), since a body’s inertia is constant while its gravity is a force that varies (according to his famous law of inverse squares).Footnote 12 We now take “inertial” and “gravitational” to denote types of, rather than properties of, mass, which types we then go on to say are really identical. So mass is the central, unifying concept, with gravity designating one of its two sorts of expression, though whether as an effect of mass on space or other masses, or a spatial extended property of mass, or a dynamic aspect of mass, or whatever, we do not know — or care. It really does not matter, and that’s the point, albeit one that has not yet penetrated into the philosophical discussion of qualia, consciousness, and mind.
No matter how important metaphysics may have seemed to Newton and his audience, even then the elements of the old Aristotelian science, that remained by virtue of their conceptual inertia, were irrelevant to the methodological soundness, explanatory potency, and empirical reality of gravity. Newton’s formulae, and the actual methods for measuring the quantities they require, are the true essence of his natural philosophy. Everything else is accidental. If modern science is famous for anything within philosophical circles, it ought to be famous for its polite, principled, and liberating disregard for any previously assumed categories of metaphysics. The pace of liberation and its progress, at first slow (“I have no need of that hypothesis,” “Maxwell’s equations are his theory,” etc.) have since accelerated. In the past century, relativity required the (apparently inconceivable) merging of time, space, matter, and energy, while quantum mechanics merged thing and process and abandoned causality for statistical correlation. We could give details and further examples, but we hardly need to. The lesson taken to heart by methodological materialism is that science is metaphysically modest: given that a theory has empirical warrant via the measurement of its quantities (or observation of its terms, insofar as this is distinct), it needs no metaphysical warrant. Methodological materialism adds to the metaphysical modesty of science the thesis that metaphysics, inasmuch as it is a metaphysics of this observable universe, must answer to empirical research — not the reverse. That is materialism enough for anyone.
So the bio-psychological explanation of colour is likewise modest: it neither proposes nor presupposes a metaphysical system. Instead, it hypothesizes the identity of seeing colours and specific neural processing. What then, of the ontological status of the yellow quale, the thing that is seen? Certainly we cannot dismiss its reality, for it is seen (among such disparate things as rats, rainbows, and reason). It is methodologically sound: detectable and controllable, and that is enough to let us get down to work scientifically. This is not to say that philosophers should surrender these questions to scientists. Presumably qualia are intentional objects, like the dollar that the coin machine recognizes (whether it is there or not), and what could be more intriguing than their ontological status? If empirical science needs such things, that is an important fact. Like Socrates, we will scrutinize the beliefs of our age. But if we accept the gravitational explanation of the moon’s career through the heavens, logical equity demands we accept the neuroscientific explanation of qualia. That is the central thesis of this essay. Contrapositively, if we reject the explanation of qualia in principle, then we must reject scientific explanations in principle.
Objection 4: If we assume that the neural colour processing of human beings gives rise to human colour qualia, then it is easy to go on to explain why colours look the way they do. The hard problem is precisely that we do not have any scientific explanation of why neural colour processing gives rise to qualia in the first place. You address only the easy problem, ignore the hard problem, and so simply miss the point philosophically, as Chalmers so lucidly argues (1996).
Reply: We may begin by noting that Chalmers, the dean of qualiaphilic studies himself, identifies why colours look the way they do as part of the hard problem. In defining the hard problem, he presents two aspects: the very existence of consciousness, and “the specific character of conscious experiences. … Why is seeing red like this, rather than like that?” (1996, p.4, his emphasis). If we take Chalmers’s words seriously, then current bio-psychological texts have already broached the hard problem. If we do not take his words seriously, we must be forgiven for concluding that the hard-easy distinction is unclear. Either way, the objection is rebutted.
Well, suppose that we prefer the hard problem presupposed in the objection, rather than in the technicalities of Chalmers’s presentation of it. The objection, then, is that the explanation offered fails to explain why there are any visual qualia at all. But this is like objecting to the standard explanation of photosynthesis on the grounds that it does not explain life itself. Of course, such an objection is irrelevant, and it is not made one whit more relevant by pointing out that the explanation presupposes the existence of life in the first place. Every explanation presupposes the things referred to in its explanans, and of course it is perfectly legitimate to desire an explanation for these things as well. But this does not in any way undermine the explanation on offer — to say otherwise is to imply that nothing can be explained until everything is explained.
This objection, moreover, has its own presuppositions. It springs from a misleading image of the science of consciousness, in which consciousness is explained by a singular, breakthrough insight into the mental realm, like Newton’s universal gravitation in the physical realm. In contrast to the relatively simple phenomena of physics, mind is like life, a multi-faceted, many-splendoured phenomenon that is understood through comprehension of the numerous sub-phenomena of which it is composed. If you would understand life, you must read biology texts. As you master such topics as photosynthesis, cellular metabolism, Mendel’s Laws, mitosis, meiosis, mutation, evolution, symbiosis, selection pressure, etc., you will come to understand life itself — and will reject calls to collapse that understanding into a single concept or catchphrase. It makes as little sense to think that a phrase like “consciousness is a brain process” (or “consciousness is a functional state,” or whatever) will provide genuine understanding of mind, as to think that “life is auto-catalytic enzymatic chemistry” will provide genuine understanding of life — even if the phrase is true and incisive. Accordingly, if you would understand consciousness, you must read bio-psychology texts. The explanation of colour qualia is to be understood as one entry in that growing text, one piece of the puzzle of consciousness.
Finally, there really is no methodological boundary between the easy and hard problems. As biology teaches us, progress on easy problems is progress on the hard ones as well. To the extent that we understand why colours look the way they do, we thereby understand why things look like anything at all — in Chalmers’s terms, why there is any visual consciousness at all. Bio-psychological research has revealed the neural mechanisms whereby the objects around one, by virtue of the light they emit or reflect, are seen by us (i.e., the existence of visual consciousness), as well as how they are seen (i.e., the character of visual consciousness). Normally, we see the facing surfaces of these surrounding objects in such a way that each point on their surfaces is seen as having a specific location, brightness, and colour. The specific shapes we see are explained by the geometry of optics and perspective,Footnote 13 and seeing these shapes requires the neurophysiological mechanisms of brightness and colour. To see is to see colours, since black and white are the most primitive colours. All parties to the debate, I think, would admit this much. But, then, it takes hard work merely to define a boundary within nature herself that these explanatory methods not only do not, but cannot, cross (Chalmers Reference Chalmers1996, e.g., has to devote the first section of his book to its definition, and yet more words in Chalmers Reference Chalmers1997, to its defence from criticisms by Churchland Reference Churchland1995, Hodgson Reference Hodgson1996, Lowe Reference Lowe1995, and Searle Reference Searle1997). This sort of boundary is methodologically unstable: specifying, defining, and distinguishing what cannot be explained invites, rather than prevents, its explanation.
Objection 5: The final objection can be put more pointedly: to explain something is to show why it has to be, to reveal its necessity. But even if the neuroscientific mechanisms of colour you describe are entirely real, those who possess them may be unconscious zombies. Therefore, the scientific explanans does not necessitate its explanandum, so the explanation fails.
Reply: I propose a dilemma: either the objection has empirical content, or it does not. If the objection does have empirical content, then we are able in principle to detect zombies, and empirical research into the source of their unconsciousness is required before we can fully explain qualia. However, zombies are empirically undetectable by hypothesis: zombies are supposed to be empirically indistinguishable from you and me—in fact, this has been one of their defining features since their introduction.Footnote 14 Even you may be a zombie, no matter how sincerely you protest otherwise, and no matter how certain you are of your own consciousness. (If you can’t buy this, you can’t buy zombies!) Therefore, the objection has no empirical content. Hence, it is scientifically irrelevant.
To put it bluntly, the zombie objection amounts to little more than invoking the ghost of Hume to demand logical necessity, proof, of those who propose scientific understanding of the mind. Hume did, famously, refute Newton’s claim to have proven his natural philosophy. Newton did indeed model his Principia on Euclid’s Principia, as a stately parade of definitions, postulates, axioms, and theorems. Like Euclid, Galileo, Descartes, and most other great thinkers up to his day, Newton believed science (defined by Aristotle as consisting solely of universal and necessary truths) required (hence was susceptible of) proof — and empirical science was no exception. We do well to recall that geometry was understood by these scientists as proven (a priori) truth about actual space. Then Hume demonstrated that proof was restricted to relations between ideas, and matters of fact could not be proven. Scientists paid no attention, as the juggernaut of scientific progress accelerated. When Hume was belatedly vindicated by the rise of quantum mechanics and relativity theory at the start of the Twentieth century, the response of scientists was not to abandon the field, but rather to abandon more geometrico to the geometers, to give up any pretence to discovering necessity in nature — and then get back down to business. In the last analysis, it is facile to object to Newton on the grounds that he has not demonstrated the necessity of his physics. Just as zombies are possible, it is possible that angels, not the forces of gravity and inertia, move every single body in the universe. But that is no reason to reject Newton’s explanation. Even philosophers do not lie awake at nights worrying about what makes baseballs fly over the outfield fence. If Newton’s laws are assumed, then it follows — with all of the necessity logic and mathematics can muster — that well-hit balls will be home runs. So we accept these laws (along with our philosophical provisos) and get some sleep.
Likewise, given the identity of neural colour-processing and human colour vision, along with the principles pertaining thereto, colour qualia are necessitated. Of course, neither these identities nor the relevant scientific principles are themselves necessary. God would have some work to do to make them necessary, to use the image employed so aptly by the likes of Kripke and Chalmers. But that is just the point: science is not metaphysics, much less theistic metaphysics. It is metaphysically modest. Its assumptions are not metaphysical, but methodological. So, zombies are possible, but that is no reason to lose sleep worrying about why we see colours.
Lest this be misunderstood, I am not professing the abandonment of metaphysics, much less the subservience of philosophy to natural science. To the contrary, it is precisely on the basis of philosophical, and even metaphysical, considerations that I base my thesis: we already possess perfectly normal scientific explanations of qualia. The thesis is that scientific explanations of qualia are completely at par logically, metaphysically, methodologically – in short, philosophically – with other explanations. So scepticism about the explanation of qualia implies scepticism about scientific explanation generally. In the contrapositive, if we are not ready to embrace this more general form of scepticism, we have to accept the proffered explanation of qualia.
Objection 6: As Kripke (Reference Kripke1980) has shown, a posteriori identities of natural kinds, such as the identity of seeing yellow with a bit of brain processing, if true, are necessarily true. You have denied that the identity is necessary, from which it follows it is not true. So the explanation fails.
Reply: Kripke’s view springs from the idea that reference of natural-kind terms is fixed by the facts. Suppose it a fact that fire = rapid oxidation. Then, says Kripke, those who talked in 1666 about the Great Fire of London were referring to rapid oxidation, even though they intended nothing of the sort, given that Lavoisier’s oxidation theory still lay in the distant future. Given this view of the reference of the word “fire,” wherever there is fire, there is rapid oxidation. If the reference of words stays the same whether we are speaking about actualities or possibilities, then “fire = rapid oxidation” is true in all possible worlds, thus necessarily true. Now suppose that we accept Kripke’s theory of reference. We have not accepted natural necessities, since nothing is necessitated for the (extra-linguistic) world but rather for the language that describes it. It is the facts of the natural world that restrict linguistic and logical possibilities, not the other way around. There is no (so to speak, detachable) necessity that fire = rapid oxidation. All Kripke requires is that on the supposition that the identity is true, all hypothetical sentences of the form “if fire is not rapid oxidation, then …” are vacuous (apply only to the null world). Very simply, in assuming the identity, we have restricted ourselves to consideration of only those possible worlds where fire = rapid oxidation.
Now, still supposing Kripkean reference, and supposing in addition the truth of the hypothesis that seeing yellow is identical to the neural process outlined in our explanans, it follows that this identity is necessary in the Kripkean sense, after all, and the objection is disarmed. At this point Kripke argues against the necessity of the identity on the grounds that we can conceive various mechanisms whereby yellow is seen, an argument that methodological materialists are happy to grant, for this is merely to say that hypothetical identities do not restrict the logical possibilities of the world itself. It is only on the assumption that the identity is true that Kripkean semantics entails its necessity. No natural necessity is implied or presumed, particularly for the assumption itself. If other mechanisms are found, then we will happily modify our views to accept them and change science to include the new identities with the new mechanisms, provided that elimination of old concepts is not a better option. This was done in the case of the elements when isotopes were discovered, and likewise for photosynthesis, cellular metabolism, cellular replication, etc. So we may find ourselves qualifying the current theory as follows: for normal human beings, seeing yellow is the specified process. If we assume (true) this qualified identity, then, given Kripkean semantics, it is necessary, and once again the objection is disarmed. Should Kripkeans object yet again on the grounds of the conceivability, a reply of the same form will, yet again, suffice.
Objection 7: Necessity is not so easily avoided. In scientific explanations of such things as fire or photosynthesis, the explanandum logically supervenes upon its microphysical substructure: in any logically possible world in which the microphysical explanans obtains, the explanandum obtains as well (Chalmers Reference Chalmers1996). In granting the possibility of zombies, you admit that your explanandum does not logically supervene upon its explanans, which implies that it fails by the standards of ordinary reductive explanation.
Reply: In order to distinguish his form of necessity from Kripke’s, Chalmers employs a distinction between primary and secondary intensions. The primary intension of a term may be thought of as the set of normal criteria used by ordinary, non-expert speakers to fix its reference. For example, the primary intension of “fire” would be something like “the bright, hot, flickering thing that consumes wood or gasoline or other fuels, that is nowadays produced by matches and lighters, that was from time immemorial produced by lightning and volcanoes. …” It would seem nice to have a precise, or even a terminating, definition of the primary intension of fire, or of anything else, but this is impossible in practice. The primary intension just is the amorphous set of marks whereby the unaided senses fix a reference, that is, the pre-scientific, or folk, concept. By contrast, the secondary intension of “fire” (as used in this world) is rapid oxidation (assuming this true), its reference fixed, in Kripke’s manner, by the scientific facts. So, secondary intension cuts through the veil of appearances to the underlying reality.Footnote 15 But primary intension concerns the veil itself.
So, to say fire logically supervenes upon rapid oxidation is to say that in any possible world in which there is rapid oxidation, there is also fire in the primary intension of the term. Intuitively speaking, scientific fire necessitates folk-fire. It is in this sense that the explanans makes its explanandum necessary. By contrast, according to the objection at hand, there are possible worlds in which the neural mechanism of seeing yellow occurs, but the yellow quale (primary intension) does not. On this basis it is claimed that whereas science does explain fire, it does not explain seeing yellow. I will argue that fire does not logically supervene upon rapid oxidation, hence that logical supervenience is a misconception of scientific explanation. So even though seeing yellow does not supervene on its neural mechanism, it is nevertheless explained by it, exactly in the way that fire is explained by the mechanism of oxidation.
If the logical supervenience conception of scientific explanation were correct, it would be impossible to conceive of rapid oxidation without folk-fire. Yet it can be conceived. In rapid oxidation, electrons excited by the reaction drop down to lower orbits by emitting photons; these photons are the light and heat which are (partly) constitutive of folk-fire. One of the innumerable ways to have rapid oxidation without folk-fire would be for these photons simply not to be emitted. Rapid oxidation would not, then, be the hot, bright phenomenon designated by the primary intension of “fire.” Rapid oxidation would then not be self-sustaining, becoming an endothermic reaction rather than an exothermic one, so heat would have to be supplied to support it. But it would, for all that, be the rapid combination of oxygen atoms with those of wood or gasoline or whatever, atomic number being the identifying characteristic of atoms. So we have (re)conceived rapid oxidation as non-fiery, and have defeated the logical supervenience view of explanation.
The only legitimate defence against such a counter-example is to show that, appearances aside, the possibility described is not really conceivable. One way or another, a defence would have to show that it is inconceivable that the electrons not emit photons as they move to lower energy levels. That such a situation violates scientific laws as now understood is obvious, but surely the logical supervenience view is not predicated on the incredible notion that the laws of science (or worse, the current laws) are necessary truths. Furthermore, we need not even respond to demands for explanation: “But how come the electrons don’t emit photons?” Who knows? If nothing were conceivable unless explained, we would understand everything we can conceive. But in actuality we cannot.
But just to make the point more transparent, I have chosen the example so that the principle in question is one of the very many for which we have no explanation in the first place: we have no idea why electrons emit photons when they lose energy (or vice-versa). We know of no mechanism for the phenomenon. It is much like the phenomenon of seeing colours in Newton’s day. Newton identified light rays of different colours by their characteristic effects on conscious perceivers, but had no explanation for this effect. For us, electrons and photons are mysterious points in space, the tiniest conceivable (?) black boxes (i.e., point particles). “But if the electrons did not emit photons, energy would not be conserved!” Maybe, maybe not. It depends on how the rest of the story goes. But however it goes, conservation of energy is also a principle with no explanation. It is certainly conceivable that energy not be conserved; that is why it is scientifically interesting that it is conserved (except, perhaps, at the instant of the “big bang,” should it have been ex nihilo): of the non-denumerable multitude of possible ways the world might have been, the evidence indicates that all may be ignored except those in which energy is conserved. It is observation, not pure logic, that leads us to accept conservation of energy.
So it is possible for there to be rapid oxidation without folk-fire. Nevertheless, rapid oxidation explains fire. Therefore, we must conclude that it has not been shown that scientific explanation requires the specific form of necessity entitled logical supervenience, and the objection therefore is based on a false premise.
Those familiar with the recent history of philosophy of science cannot be surprised at the failure of logical supervenience, given the repeated failure and eventual abandonment of micro-reduction (Oppenheim and Putnam Reference Oppenheim and Putnam1958). The hope motivating micro-reduction was that the facts (both laws and particulars) of a (relative) macro-level could be deduced from those of its (relative) micro-level; in this way science could be unified via a series of reductions through different orders of size that finally rest in some future, ultimate microphysics. Alas, the deductions could not be achieved, given the disparities between laws, concepts, geometries, and logics at the different levels. What is novel in Chalmers’s approach is his attempt to sidestep the well-known deductive gaps in the syntactic domain by moving to the corresponding possible-world semantics. But if we cannot deduce fire from rapid oxidation (at least not without postulating their identity), why should we expect to observe that, throughout each and every possible world, wherever there is rapid oxidation there is fire? “Observation” (so-called) of possible worlds seems unlikely to yield more clarity than the corresponding syntax, much less the necessity clearly not present in that syntax.Footnote 16
Chalmers’s captivating arguments seem to bear this out. Why does he think that fire supervenes on rapid oxidation? Aptly enough, he often expresses his thesis concerning normal, successful scientific explanation by saying that the macrophysical supervenes upon the microphysical. What is his argument? What one expects at this point — and what is actually needed — is a specification of the microphysics of fire, and then an argument that this diminutive physics yields a folk-fire in all possible worlds. But no microphysics is ever mooted, and this surely must be reckoned a strange and revelatory (i.e., fatal) omission. In fact, the argument proceeds in the other direction: we start with the fire, and merely assume it is identical with some microphysical structure. Ignoring simpler cases like fire in favour of a wombat with two children, Chalmers reasons as follows:
The physical facts about our world will include facts about the distribution of every particle in the spatiotemporal hunk corresponding to the wombat, and its children, and their environment, and their evolutionary histories. If a world shared those physical facts with ours, but was not a world in which the wombat had two children, what could that difference consist in? … Such a world is inconceivable in principle. Even a superbeing, or God, could not imagine such a world. (1996, p. 73)
This Laplacean picture may well seem irresistible to the modern mind, but that hardly renders it the only picture conceivable. And God, as we shall see, can have other ideas.
Rather than specify the microphysics in question, this argument simply assumes that the wombat, its children, the environment, and their history going back at least to the dawn of life are nothing other than “distributions” of microparticles. Chalmers does not argue from a specified structure of microparticles, via physics, to a wombat. Instead, he effortlessly refers to a set of particles as those which are assumed to be the wombat. This is not physics, but physicalism, the doctrine that physics includes everything. Intriguingly, this will make it most attractive to those most disinclined to accept its ultimate conclusion: the inaccessibility of consciousness to physical science. Laplace, who was both physicist and physicalist, at least realized one needs to invoke a superhuman intelligence to realize the physicalist vision. It is only at the level of folk-microphysics, so to speak, that by a wave of the hand we may specify a wombat-identical system of particles. Of course, working under the assumption of this identity, the wombat logically supervenes on the particles — but this is to assume precisely what is at issue. Even if it is so in “our world,” as his argument assumes, why would it be so in all possible worlds? Chalmers’s argument begs the question. Logical supervenience is not demonstrated, but assumed. That is, Chalmers assumes physicalism.
If, on the other hand, the supposed set of wombat-making microparticles were actually specified, God would have little difficulty in nevertheless denying us a wombat in its primary intension (or folk-wombat). Now that Chalmers has brought God into the discussion, it is instructive to return to the massively simpler case of fire. God need not go so far as to alter standard electron-photon interactions in order to get rapid oxidation without folk-fire. All S/He need do is ensure that none of the energetic photons emitted in rapid oxidation ever interacts with human retinas. Under their quantum aspect, all such interactions are enjoined merely by quantum probabilities. It would be very, very improbable that so many photons fail to react with so many nerve endings, but God could pull it off by simple decree. Quantum probabilities could even be preserved (over the universe as a whole) by having photons overly reactive elsewhere. We would see wood, gasoline, and so on being consumed, and things (even our own bodies) being heated and sometimes charred, but we would never experience the bright flickering that is definitive of folk-fire. Rapid oxidation would be invisible, like nuclear radiation. God could do this without even altering physical law. It would, moreover, be even easier for God to deny us folk-wombats in a world with the same physics as our own (wombats being scarcer than fire).
It is quite remarkable that, according to Chalmers, physicalism is so easily conceived for fire and wombats that anyone can do it, while conceiving a complete physicalism, including consciousness, is so difficult that not even God can do it. God can conceive a physical duplicate of this world, identical down to the last particle. In this possible world there is a person indistinguishable from you, reading a copy of this essay. That person is also certain that s/he is conscious, and would (or does) provide just the same evidence and argument for this as you would (or do). If God were a committed physicalist, and so could even conceive of complete physicalism, that person would be conscious, and you, even assuming you are conscious, could be that person. But, according to Chalmers, if you are that person, then — despite what you may think — you are unconscious, so God logically cannot be a physicalist. S/He can be a physicalist when it comes to wombats, but not when it comes to you. God is constrained to add to the world something beyond the ken of physical science to get you and your consciousness. This ambitious claim, I submit, simply begs the question. It cannot be supported by the argument from the failure of consciousness to logically supervene on physics, because it is the ultimate basis of that argument. So we are left with the naked contention that a completely physical universe, one that includes consciousness, is inconceivable. Given that we do not even have a way of telling whether or not we are in that purely physical world, not even in principle, the contention that it is impossible is stripped even of any perceptible significance to us, which is about as naked as naked can be. Surely it has no scientific significance, despite the blithe talk of microphysics.
In summary, no sufficient reason has been given to think scientific explanation requires logical supervenience. So, no disparity has been demonstrated between scientific explanation in general and the explanation of qualia.
Objection 8: It is possible that some people experience colour spectra that are inverted relative to those of others: where you see yellow, they see blue, etc. So where, according to your explanation, there is a yellow quale, they in fact experience a blue quale. Therefore the explanation fails.
Reply: I propose a (by now familiar) dilemma: either this possibility has empirical content, or else it does not. Consider the first option. We are already familiar with circumstances that change the colours one sees. It has been known for centuries that jaundice makes things look yellower (a temporary intra-personal change), for example, and Dalton discovered red-green colour-blindness in the nineteenth century (a permanent interpersonal difference). Empirical investigation of such conditions can explain and enumerate their effects, with the result that we will come to know the differences in colour qualia between different perceivers, and whether anyone’s spectrum is inverted relative to anyone else’s. This interesting, empirically meaningful query poses no objection to the present explanation, but instead calls for further research.
If, on the other hand, inverted spectra have no empirical content, then there is no scientific problem. Imagine two people, Laverne and Shirley (Gibbons Reference Gibbons2005), who are identical as far as their nervous systems process colour. Nothing we do or can do could ever reveal the least difference between them. Yet we are to suppose that one of them has an inverted spectrum: fire hydrants look to Laverne the way that grass looks to Shirley and the rest of us. Yet Laverne says that fire hydrants are red, and, all things considered, knows that fire hydrants are red just as well as any of us do—so what grounds do we have for saying she is mistaken when she says they look red?Footnote 17 How desperately must one dislike physicalism to make a stand against it on such iffy grounds? We might just as well suppose that somebody might undergo a pain-pleasure inversion, but then like pain, find it soothing, etc., while hating pleasure, being disturbed by it, etc. Such a speculation would hardly be a challenge to psychology. We might equally imagine that, of two functionally identical computers in front of us, one has its machine code inverted, zero for one (or on), one for zero (or off). Given this inversion, it reads all of its programs and input devices invertedly, and sends inverted output to its monitors and actuators. No empirical investigation could ever reveal any difference, by hypothesis, for the inverted computer. If we now object that computer science does not explain computation because of the possible inversion of machine code, we cannot complain if we are dismissed with the observation that science is not in the business of ruling out every possible supposition—especially those that could never make any empirical difference. The same reply should suffice in the case of the supposition of an inverted spectrum.
Objection 9: You have explained only the accidental properties of the yellow quale, but have left its essence untouched.
Reply: What could be more essential to a yellow quale than the way it looks? The essence of something (insofar as there is any agreement on a topic of such antiquity and, hence, thorough disputation) is the set of properties that make it what it is, rather than something of a different sort. Generally speaking, what makes qualia what they are is what they appear to be. In the case of colour qualia, the sense of “appearance” in question is literal: look at something yellow; the colour you see is the yellow quale. Yellow qualia are precisely those that look yellow, and not those that look red, green, etc. Its essence is its look. That is what needs to be explained — and is what is taken by some to be an insuperable barrier to scientific explanation of consciousness. However, that is precisely the explanandum of the explanation at hand. So the explanation is on target.
Sometimes this simple fact is obscured by the circuitous terminology used to refer to qualia, such as “raw feels” (Ryle Reference Ryle1949), or simply “feels,” or “what it is like to be x,” where x takes values such as “a bat,” or “see yellow” (Nagel Reference Nagel1974). Unless we read these phrases charitably, we will be misled. Chalmers, for instance, says that “what it means for a state to be phenomenal is for it to feel a certain way” (1996, p. 12, his emphasis), and in various places demands explanation of why seeing a colour feels the way it does. This literally indicates he is talking about the emotive, tactile or proprioceptive affects of seeing the colour, such as why seeing red might make one feel alarmed or queasy. But the affects of seeing colour is not our topic — nor his, presumably. Similar comments apply to the Nagelian formulation. What is it like to see red? Maybe it is fun, maybe it is boring, maybe each time is different — but that is not the colour qualia problem. Presumably, then, in the case of colour qualia at least, Nagelians like Chalmers will agree that what they mean by “what it feels like too see yellow” is the way yellow looks.
Objection 10: Qualia are intrinsic properties, but you have addressed only extrinsic properties, and therefore miss the mark. Your proposed explanation addresses only the relational, hence extrinsic, properties of yellow: its brightness (its closer similarity to white than to black), its atomicity (its not being composed of other hues), its place in a colour space, and so on. You have left untouched what yellow is in itself, and so have left untouched the real puzzle of qualia.
Reply: The explanatory demand just stated would rule out all scientific explanation whatever. We do not feel puzzled about what keeps the moon in the sky, even though all of the concepts used to explain this phenomenon are relational. The mass of the moon, for instance, is an essential part of the explanation of its path through the heavens, and mass is defined in terms of a relationship, the ratio of resistance to acceleration. Acceleration in turn is a relationship among relationships, the ratio of the difference in velocity at two different times relative to the difference between those two times. Velocity is the ratio of the difference in spatial position at two times relative to the difference in time. And now we have traced our definitions to the concepts of space and time, and surely the relativity or time and space is old news. In brief, scientific concepts are inter-defined, and the meaning of any scientific concept is given by its specific relationship to the others. The mere possibility of measurement requires such relationships, such as the relationship between temperature and the expansion of mercury in a thermometer, or the relationship between time and the processes of clocks. The very logic of explanation, scientific or otherwise, requires that (on pain of circularity) the explanandum be distinct from the explanans, which entails that explanation must involve the relationships (logical and mathematical) between explanans and explanandum. To demand that a scientific explanation of qualia deal specifically with their intrinsic properties is to set up a criterion that would rule out all explanation whatever.
It is not even clear that the notion of intrinsic property makes sense.Footnote 18 An intrinsic property is supposed to be one that something has independently of anything else, or aside from its relationships to them. But independence itself is a relationship. The absence of any given relationship is, logically, just another relationship: the fact that 3 is not equal to 4 (the absence of equality) is just as relational as the fact that it is equal to the square root of 9 (the presence of equality). To put it paradoxically, the very concept of intrinsic property reveals them to be extrinsic.
Putting pure logic aside and turning to epistemology, it is obvious that we cannot possibly know about any property except insofar as this property also has extrinsic relationships to us and our epistemic faculties. Even if we suppose that the intrinsic property (or properties) of the yellow quale is not ruled out by logical paradox, we could only know about such a property insofar as it was causally involved in the production of this knowledge. When we say “I see a yellow quale,” this physical report must have the yellow quale among its causal antecedents in order for its expressed claim to be known. And so we are faced, once again, with a (hopefully familiar) dilemma. To suppose that, by virtue of being intrinsic, the properties of the quale are causally inert, is to suppose it is forever unknowable.Footnote 19 On the other hand, to suppose that it is the cause of our knowledge of it, is to suppose that it is only by virtue of its extrinsic properties that it is known, since causality is relational, hence extrinsic. Therefore, elementary epistemological considerations tell us that if any qualia are known, they are known solely by virtue of their extrinsic properties. So if they are ever to be understood, that is, explained, it must be by virtue of their extrinsic properties. Hence this objection can evade the explanation of qualia only at the cost of evading the topic altogether.
Objection 11: Since you raise the issue of epistemology, you should know that it has already been employed decisively against you by Jackson (Reference Jackson1982, Reference Jackson1986). Imagine a scientist, Mary, who completely understands the explanation of colour qualia you have outlined, in its full theoretical background moreover, but who has never seen colours due to being sequestered since birth in a black and white environment. Despite Mary’s complete theoretical understanding, she has not the least knowledge of what colours look like, hence no knowledge of colour qualia as you define them. This is confirmed by the fact that when she finally is allowed to see red, she learns something new. Of course, the same thing applies to yellow.
Reply: Actually, yellow is different. As we have seen, the yellow-blue opponent process is not independent of the light-dark process, as Mary would surely know. Even though she had never seen either hue before, she could not be fooled into accepting a paradigm shade of blue in place of a paradigm shade of yellow, or vice versa. So Jackson’s argument thrives only because it has been restricted to, as Wittgenstein put it, a narrow diet of examples. It will not do here to object that, since it still works with the red-green continuum, the Mary argument should not be faulted for its failure with the yellow-blue continuum. If it worked at all, it would show that Mary cannot even in principle know what colour qualia are like. So given that she does know what some of them are like (indeed, arguably half of them, the yellow-blue half), the argument fails. If some science may trespass the barrier to inform Mary (or us) about qualia (i.e., qualia-as-such), then the barrier is not a barrier in principle. An airtight barrier with some leaks is not an airtight barrier.
Nevertheless, more interesting than yellow’s exception to the Jackson argument is why it works when it does work. I propose the following: (i) Jackson’s justly famous argument relies on seldom-recognized relationships that must obtain between any form of empirical knowledge and the perceptual experience of the knowing subject (or community); (ii) these relationships obtain for all scientific knowledge, not just for scientific knowledge of qualia; and (iii) they have been prejudicially applied to qualia in Jackson’s argument, though in fact qualia are not special in this regard. An example will explain what I mean.
Consider first Frank, who is an absolute expert in astrophysics and knows all that there is to know about the structures of solar systems, galaxies, and the universe as a whole. In addition to such mega-structures, he also knows about the micro-structure of heavenly bodies including the nuclear fusion that heats the stars, the collapse of matter in black holes, and the relative stability of planets. However, he has been kept, from birth, in a metrically deficient environment. He has never once been allowed to see a ruler or metre stick, to know his own height or the size of the apartment in which he is imprisoned or the number of molecules in a glass of water. Nor does he know the weight or mass of anything around him, or its temperature. He is not permitted to see clocks or witness for himself the passing of one day into the next; to know how old he is or the standard marks of age in a human being. Admittedly, it is difficult to prevent Frank from having any personal experience of measurement, but then it is likewise difficult to prevent Mary from having any experience of colour: she must never be allowed to rub her eyes (thus stimulating her retinal cones); her colour processing systems must develop, despite never being used, and yet not atrophy, etc. Frank knows such things as the size of the Earth, its distance to the sun, the time it takes light to traverse the galaxy, as well as the size of the hydrogen atom, the rate at which they burn in stellar cores, and the duration of stellar collapse. But there are things he still does not know, as is proven by his surprise when he is shown a metre-stick one day. “So that’s how big a metre is!” he exclaims. He goes on to say that at long last he can truly understand what he knew merely in theory before, and so on. More of the same follows as he is given a watch, a scale, etc. (though the information they provide could have been inferred from knowing the length of a metre, or conversely, which is why they were forbidden in the first place).
One thing will be readily apparent, I trust, in the case of poor Frank: the knowledge of which he is deprived prior to seeing the metre stick is not about qualia, but about astrophysics. What Frank’s case shows is that every form of empirical knowledge presupposes some connection with the experience of the knower. That is precisely what is required in order that the knowledge be empirical. Exactly what experience is required, and how it must be related to what is known, has been the subject of investigation and dispute from the first glimmerings of philosophy until today. It is obviously a complicated matter, involving such things as faith in proper authority (we cannot all look into the telescopes and microscopes, or operate mass-spectrometer detected gas chromatographs, for ourselves), the peculiarities of individual histories (those lucky enough to have seen aardvarks know them better, in a certain, hard-to-define way, than those who have not), and so on.
Tye admits that when Mary first sees colours, “in an ordinary, everyday sense, Mary’s knowledge increases” (p. 18) because she gains access to colours via first-person “phenomenal concepts,” whereas before she had access merely by way of the third-person concepts of physical science.Footnote 20 He says this despite also asserting that “given her exhaustive knowledge of the physical facts, the contents of her thought-types before and after [seeing colours] remain unchanged” (p.19). Like Churchland (Reference Churchland1988) before him, Tye concludes that what Mary knows stays the same, thus rejecting Jackson’s challenge to physicalism while accepting that how she knows it changes. Though I have a lot of sympathy with this approach, I believe it avoids Jackson’s argument on a mere technicality. We cannot deny that part of the change in how Mary knows colours is that she knows them better after she has seen them (as Tye’s admission indicates)—in which case Jackson’s premise that Mary learns something by seeing colours is still materially accepted even if formally rejected.
I suggest instead that scientific explanations require us to locate ourselves in the scientific model employed, just as we must first locate ourselves on a map before we can use it to get to our destination. Some maps, like those installed on university campuses, include a dot marked “You are here” to help with this process. Until you know where you are on the map, you cannot use it to get around. It is just as important to relate the other conventions of the map to your own empirical senses: to know what represents the buildings, what represents the parking lots, and the distances in terms of your own experience. In the case of scientific models, we must coordinate our natural (mainly innate) manifest model (in which directions, e.g., are normally represented as in front or in back of us, to the left, right, etc.) with the (mainly learned) scientific model (in which directions, e.g., are normally represented as East, West, North, South, etc.), which in turn requires sensory access to the quantities, concepts, and entities used in the models. Frank (partially) locates himself in the scientific model of the universe by seeing how long a metre is. Likewise, Mary does improve her epistemic situation, not by learning something new but by improving the coordination between her manifest and scientific models of colour. For a more detailed explanation of this “keying” process, see my (2000), esp. pp. 69-96.Footnote 21
Most importantly, scientific knowledge is about the very same world in which we are physically located as embodied beings and to which we have epistemic access solely by virtue of our senses. Constricting that access constricts our possibilities for knowing the world. That much is obvious, though there has not been much philosophical research into this fact and many uncertainties remain. To take the extreme case, someone who was denied all sensory contact with the world would not have any empirical knowledge of it. The empirical world would always remain a mere theoretical posit for such an individual. Is it for this reason that Jackson does not suppose the simpler case, that Mary is blind? In such a case it would be improbable that Mary could, as it were, picture the behaviour of light along with the numerous tiny structures of the nervous system needed for vision, much less be able to picture seeing itself. Overhauling the scientific theory of vision to make it accessible to the blind is no simple matter, and its potential uncertain. Supposing that Mary were from birth (reversibly) blind, instead of just being colour deprived, would bring to the fore the necessity of specific sorts of empirical experience to sorts of scientific knowledge. This might in turn lead us to reflect that scientific models need to be keyed to our own experience in order to provide us empirical knowledge, and that this keying process (roughly a coordination of indexically specified experience with scientific quantities, concepts, etc.), while not a species of coming to truly believe a proposition, is nevertheless essential for empirical knowledge.
To put it picturesquely, our various senses provide us various windows on the world, and closing one of them constricts our ability to comprehend the world. Jackson’s Mary argument involves just such a constriction. Once this constriction is removed, Mary is better able to key her scientific models to her own experience, and thus is more able to see and understand the world through her scientific model. Specific sorts of experience are more relevant to certain scientific models than others. Seeing a wombat is particularly relevant to biological models, but not to those of nuclear physics. Likewise, actually seeing colours is relevant to anyone who, like Mary, might be called a specialist in colour vision.
In any case, it would be wrong to conclude from the case of poor Frank that there is a special domain, the domain of the measurable, into which science cannot, even in principle, venture. Likewise, it would be wrong to conclude from the case of Mary that there is a special domain, the domain of human sensory experience, into which science cannot, even in principle, venture. Accordingly, I propose that the lesson Jackson draws from Mary’s case is completely off the mark: her case implies nothing about qualia or the limits of physical science as such. It does have something to teach us, on the other hand, about the general conditions of empirical knowledge. What we may conclude is that, just as internal relationships among concepts are essential to scientific knowledge (as sketched in the reply to the previous objection), so too are there essential relationships between those concepts and the experiences of the knower (or community of knowers).
Objection 12: You have admitted the main point that so-called “qualia freaks” are pressing against physicalism: that in merely seeing something we learn something. You qualify the assertion by redefining this learning as “improved empirical access” or whatever, but the point remains. But, as Michael Tye (Reference Tye2000, pp. 21-42) has shown, explaining just what it is that is learned (or is gained epistemically) here is impossible. By assuming such an explanation is needed you have fallen prey to the illusion that there really is an explanatory gap in Levine’s (Reference Levine1983) sense—and so have unwittingly played into the hands of your opponents.
Reply: I accept Tye’s conclusion that such an explanation is not needed in order to maintain the logical coherence of physicalism, and agree that in this sense the “explanatory gap” is an illusion. However, I think Tye wrongly concludes that no such explanation is possible, and so he is forced to retreat to the position that its absence is not a logical objection to identifying phenomenal states with brain states. Though his final position is perfectly sound from a logical point of view, I think that the request for an explanation to fill the apparent gap is reasonable on empirical and methodological grounds, even though it is not demanded on logical grounds.
Tye offers an argument reminiscent of Nagel (Reference Nagel1974) that the way colours look cannot be explained since “third-person, objective concepts” employed in explanation (p. 31) do not make contact of the right sort with first-person “phenomenal concepts” based on “phenomenal qualities.” Like Nagel, Tye begins with the conviction that the “intuition that there is a direct and immediate access to such [phenomenal] qualities (under normal conditions) needs to be preserved” (p. 28), but then goes on to recast Nagel’s view in terms of a neo-Fregean sense-reference distinction. Plausibly taking Fregean senses to arise from different “modes of presentation” (p. 18, e.g.) of a given referent, Nagel’s distinction between first-person and third-person concepts can be reconstructed as follows: science provides us a third-person mode of presentation (sense 1) of our sensory states in terms of brain processes, while those very states provide us a first-person mode of presentation (sense 2) of themselves as phenomenal qualities. Thus there is only one reference, the sensory state, but two senses: 1. the third-person (scientific) sense, and 2. the first-person (phenomenal quality) sense.
From this point of view, the demand that qualia be explained scientifically amounts to the demand that one sense (the first-person sense) of a given reference (the sensory state) be deduced from another (the third-person sense). But this is like demanding that the Clark Kent’s mild-mannered newspaper reporting be deducible from Superman’s bold use of superpowers to achieve justice: there is no reason whatever to think that one mode of presentation can be deduced from another. Add to this the fact that access to the first-person sense is necessarily restricted to the first-person, hence that, for all any of us know, “absent qualia are also epistemically possible” (p. 31), and we go from not having a guarantee, to being unable ever to dispel logical doubts about any such deduction. Even so, thinks Tye, we need not panic: “There is no significant question here for the physicalist. Only one state exists, conceived of in two ways, and that state must be self-identical” (p. 33). He then goes on to say that we have excellent empirical reasons for accepting the identity in question—and here, of course, I agree—and hence for accepting his neo-Fregean analysis (which, relevantly, bears a strong resemblance to Paul Churchland’s view in his 1988, cast in terms of the distinction between intensional and extensional properties).
I have no objections to this analysis so far as it goes, but I do think that the demands that may reasonably made on science here are greater than Tye allows. Certainly there is nothing in Frege’s analysis of empirically based identities that requires that one sense be deducible from another, and such a lack does not automatically impeach the identity in question. On the other hand, even modes of presentation are open to scientific investigation. There is nothing to prevent us from explaining on scientific grounds why, for example, Venus has two modes of presentation. If someone asks how it could possibly be the case that the Morning Star, which is seen only in the morning, could possibly be one and the same thing as the Evening Star, which contrariwise is accessible only in the evening, that person has not made an unreasonable request. Tye’s essentially correct (but perhaps too subtle) argument runs up against the fact that science typically does explain not only the way things are, but the way they look or seem as well.
Part of the reason we accept that the Earth moves is that science explains why it nevertheless seems to be still. In other words, science explains not only within the realm of reference, but within the realm of sense as well. There was a time before the explanation of the apparent immobility of the Earth when Copernicus and Galileo nevertheless accepted the helio-centric model. They were not irrational or unscientific to do so: they had perfectly good empirical grounds to accept the model, and hence that the Earth moved, and to treat the apparent immobility of the Earth as a problem to be solved rather than as a fatal objection to helio-centrism. On the other hand, the request that this appearance be explained, and explained scientifically, that is, in a manner consistent with the helio-centric model itself, was perfectly reasonable, even though both Copernicus and Galileo went to their graves without satisfying this request. We had to wait until Newton’s physics in order to understand why Earth seems motionless (when standing still we will neither see any motion of the Earth nor feel any forces of acceleration).
However, Newton’s explanation of Earth’s apparent motionlessness was a real scientific achievement, and one entirely relevant to whether we should accept the hypothesis that Earth orbits the sun. Tye is wrong to suggest that Copernicus, Galileo, and Newton simply could have rejected this demand for explanation as a “cognitive illusion.” Indeed, our more recent neuroscientific discoveries concerning the mechanisms of the inner ear for sensing motion are part of the ongoing project of modern science, and in this way neuroscience is in the business of completing the work of Copernicus, Galileo, and Newton. Furthermore, the apparent motionlessness of the Earth is not by any means an isolated case. To pick just one more example at random, given that objects are numerically identical with swarms of tiny particles in nearly empty space, we need to understand why we refer to objects via concepts that have the contrary sense of solidity and fullness. We accept the identity of physical objects with micro-particles in large part because we have micro-physical explanation of why our hands cannot pass through tabletops and our eyes can see neither the particles of which hands and tables are comprised nor the spaces between them. Likewise, if the identity of seeing yellow and the state of our nervous system when we see yellow is ever to be fully acceptable to us, we have to have an acceptable explanation of why things would look yellow when we are in that state. Admittedly, the explanation of why yellow looks the way it does takes us far deeper into the workings of the nervous system than the explanation of why the Earth looks motionless, but otherwise they are logically and methodologically equal. Tye is right that the identity of sensory states with neural states is not defeated by the explanatory gap, but wrong to think the gap unbridgeable, and wrong to dismiss the goal of bridging it as a cognitive illusion.
Objection 13: Matthen has convincingly shown that it is entirely accidental that a given sensory experience is attached to any particular property we may sense (2005. pp. 241-3). It is a mere historical accident that the sensory classification of yellow involves the experience of yellow. “Levine’s complaint [the explanatory gap] obviously implies that there is something that needs to be explained about why a particular sensory classification leads to a particular experiental signal”, but “the spirit in which such questions are asked is misguided. They are no different from asking why ‘bat’ stands for bat rather than mat or cat… . . [But there] is no appropriateness of signal to property beyond that which is endowed upon it by historical accident” (p. 242). So, if your argument were correct, it would have shown a necessity in nature that simply cannot be there. Therefore, your argument is not correct.
Reply: I have a lot of agreement with Matthen’s view, which I see as saving the kernel of truth that motivates Tye’s objection. Like Tye, Matthen’s philosophy benefits enormously from being well-informed by empirical science, and I do not want to give the impression that I reject either of their views wholesale. Indeed, I see Matthen’s account as an ingenious and very attractive rethinking of traditional philosophical concepts in the epistemology and metaphysics of perception, one that brings them into agreement with the state of the art in the empirical sciences relevant to human perception—and that is mainly consistent with the spirit, if not every letter, of Tye’s account. But I note that Matthen does not say that such explanations are impossible, and I would propose that there is a proper spirit in which they may be requested.
Locke used to muse that it was entirely possible for God to make the sensation of blue be the normal effect of seeing yellow objects, hence that the actual conjunction of yellow sensations with yellow objects could not be explained scientifically but must instead be attributed to the imponderable will of God. We can do much better than that now by appealing to the vicissitudes of evolution rather than to the vicissitudes of divine will. From this point of view, Matthen is right to say that the reason yellow looks the way it does is the result “of a chance arrival by natural selection at one equilibrium point rather than another. There is no appropriateness of signal [yellow qualia] to property [yellow surface] beyond that which is endowed upon it by historical accident” (p. 242). I agree.
But there is more at issue here than “appropriateness,” however Matthen conceives of it. It seems that our colour sensory processes might have been wired up differently, and even that it might have been the case that the neural state caused by yellow objects be the one that is currently caused by blue objects, and vice versa, just as Locke speculated, though we do not need to make the huge metaphysical assumption of theism in order to arrive at this conclusion. We simply need observe, as Matthen does, that various sorts of colour processing, with different sorts of neural wiring, have evolved among different species of animals. So, from the point of view of evolutionary theory, we may simply treat our own wiring as an accident, and hence the look of yellow things as an accident. Usually we cannot trace specific facts of a given phenotype to its genotype, nor trace the genotype to specific mutation events and selection events that are now, alas, lost in the depths of time. So when it comes to evolutionary explanations, we must be happy with generalities rather than with specifics.
Nevertheless, there are no accidents from a scientific point of view. And while we may be content to dismiss questions about why we are wired up the way we are to the category of historical accident under the general heading of evolution, it cannot be an accident that activation of the wiring for seeing yellow results in the fact that we do in see yellow, rather than, say, blue. Given that we are wired up as we are, it is not an accident that yellow is a bright colour and blue is a dark colour. Given our wiring, yellow has to look like yellow, and not like blue. That is a far more noteworthy fact than any of the remaining gaps in perceptual science that we may point to. In other words, we already have some good scientific explanations of qualia, so the spirit of antipathy toward neuroscience on the grounds of a so-called explanatory gap is indeed misguided. On the other hand, filling in explanatory gaps increases our understanding, and this is always a good thing. Or so my philosophy tells me.
3. Conclusion
The philosophy of mind deserves its own subject matter. The inherited wisdom of the preceding millennia is that this is automatically guaranteed, because mind cannot be known in the same way as other things. Typically this view was the cornerstone of an accommodating metaphysic, usually one in which mind exceeded the ontological and epistemological resources of the physical. In the past few centuries, however, science has far outstripped metaphysics in explanatory power, by relying on a methodological materialism rather than on, as is so often supposed, a materialist metaphysics. Perhaps because this is not adequately realized, the most prominent arguments within contemporary philosophy of mind against the scientific explanation of consciousness, and qualia in particular, would, if correct, count against scientific explanation generally. Properly understood, these arguments are really a critique of modern science, and of part of the skeptical tradition within the philosophy of science. Proponents of these anti-scientific views would do well to realize that it is not just psychology, or consciousness studies, that they are challenging, but science in general. They are, of course, welcome to do so — but their arguments are not part of the philosophy of mind as such. Mind, like life or electrons, poses special problems for those scientists and other philosophers who seek to understand it. Evading the gaze of science altogether, in principle, is not one of them.
