(review of C. McGinn, The Problem of Consciousness), The Times Literary Supplement, May 10, 10, 1991.
Colin McGinn, The Problem of Consciousness, Blackwell, 1990, 30 pounds, 213 pages, Index.
These are heady times for the sciences of the mind. The pace of discovery is quickening, thanks to the mountain of data provided by the new brain-imaging technologies, but thanks even more to the computer simulations that have expanded and disciplined our imaginations, dramatically enlarging the logical space of models that can be investigated. We can now seriously consider hypotheses that a few years ago were simply unframable--"inconceivable", a philosopher might have been tempted to say. These computer-expanded powers are being vigorously exploited by a new generation of theorists and experimentalists. In some quarters the first symptoms of gold rush fever have been detected.
In other words, it's a perfect season for naysayers, and philosophers have risen to the occasion. The most radical is Colin McGinn, former Wilde Reader of Mental Philosophy at Oxford, who has recently taken a position at Rutgers University in New Jersey. The Problem of Consciousness is a collection of eight essays, two of which have not previously been published. McGinn's central thesis is that the problem of consciousness is systematically insoluble by us (Martians or demigods might have better luck). Our brains just weren't meant to get a grip on this tough problem, but--there, there, it's all right--we mustn't draw the conclusion from the fact that we can't understand it, that the mind is intrinsically mysterious. After all, whoever promised that we should be able to understand all possible good science?
McGinn is quite aware of how outrageous this doctrine will appear to many people, and he takes pains to place his pessimism in the best light.
I predict that many readers of this chapter will find its main thesis utterly incredible, even ludicrous. Let me remark that I sympathize with such readers: the thesis is not easily digestible. But I would say this: if the thesis is actually true, it will still strike us as hard to believe. (p.21)
Chalk up one victory for his prediction: I find his thesis not just incredible and ludicrous. As a fellow philosopher, I find it embarrassing. It is not that I disagree with McGinn about the possibility in principle that there are phenomena that will forever defy human understanding, but just that I find him arriving at his pessimistic verdict about consciousness after such a paltry canvassing of the opportunities. His thesis about the likely limitations of our brains would be uncontroversially true if it weren't for our clever trick of expanding the powers of our naked brains by off-loading much of the work to artifacts we have designed and built just for this purpose. The brains we were born with are no doubt quite incapable of grasping long division--let alone calculus or photosynthesis--without the aid of pencil and paper or chalk and blackboard. We have to work to acquire some of our concepts, but we don't have to do all the work in our heads, thank goodness. One might think, then, that in order to defend a thesis about the outer limits of our powers, one should at least take a peek at the concepts made available to those who have armed themselves with the new technology. A quick survey a current cognitive science would reveal, moreover, hot-spots of enthusiasm here and there that would benefit from a wet blanket judiciously thrown by a philosopher. This is not McGinn's agenda, however. He has figured out our limits from first principles, and he confidently declares without direct examination that there is no salvation to be found in those quarters. So far as I can tell, he arrives at this judgment without having gone to the trouble of becoming acquainted with the new concepts.
Compare consciousness to another notoriously baffling case: quantum physics. The late Richard Feynmann, who probably grasped it as well as anyone, is often quoted: "I don't understand it; nobody understands it." This from a virtuoso explainer, one of history's most resourceful understanders. But even Feynmann, who was also an International Grand Master in the chutzpah department, didn't say: "I don't understand it, and nobody ever will." (He, by the way, spent many of his last days exploring the power of huge parallel super-computers to expand the imaginations of physicists.)
What support does McGinn offer for his striking conclusion that "there is something terminal about our perplexity" (p.7)? He draws his main inspiration from two philosophical sources, Thomas Nagel (formerly at Princeton, now at NYU) and Jerry Fodor, now his colleague at Rutgers. All three live in Manhattan and are no strangers to the Port Authority Bus Terminal. Perhaps this helps to explain their shared pessimism, for it does appear that we are witnessing the birth of a new school of philosophy: New Jersey Nihilism. Nagel's naysaying is renowned. His 1974 paper, "What is it Like to be a Bat?" advanced the claim that it is impossible for us to know the answer to the title question. McGinn takes himself to be building on Nagel's foundations, expanding his claims, hardening the line. Fodor's nihilism is less often recognized, but has been emerging with increasing vigor in recent years.Endnote 1 Here, for instance, is Fodor on why science cannot explain the central Cartesian core of the mind:
A recent issue of Scientific American (September, 1979) was devoted to the brain. Its table of contents is quite as interesting as the papers it contains. There are, as you might expect, articles that cover the neuropsychology of language and of the perceptual mechanisms. But there is nothing on the neuropsychology of thought--presumably because nothing is known about the neuropsychology of thought. I am suggesting that there is a good reason why nothing is known about it--namely, that there is nothing to know about it. (The Modularity of Mind, MIT Press, 1983, p. 119.)
McGinn relies on Fodor to provide the entering wedge of his argument: what Fodor calls epistemic boundedness and McGinn calls cognitive closure. "A type of mind M is cognitively closed with respect to a property P (or theory T) if and only if the concept-forming procedures at M's disposal cannot extend to a grasp of P (or an understanding of T)." (p.3) (Don't be misled about the apparent rigor of this definition; the author A never puts it to any use U in any formal derivation D.) The idea is plausible enough: "What is closed to the mind of a rat may be open to the mind of a monkey, and what is open to us may be closed to the monkey." Monkeys, for instance, can't grasp the concept of an electron, McGinn reminds us. Fodor in turn got his idea of epistemic boundedness from Noam Chomsky, who divides all matters of human puzzlement into soluble problems and insoluble "mysteries". The human mind has its limits, and for Chomsky, the problem of free will, for instance, is simply off limits. This is doctrinally convenient, maybe, but rhetorically unstable, to say the least. In other moods, both Chomsky and Fodor have hailed the capacity of the human brain to parse, and hence presumably understand, the official infinity of grammatical sentences of a natural language. If we can understand all the sentences, can't we understand the sentences that best express the solutions to the problems of free will or consciousness? I think we should be unimpressed by the example of the monkey, to whom the electron is out of bounds, for not only can it not understand the answers; it can't understand the questions. The monkey isn't baffled, not even a little bit.
Of all the technologies that have expanded our powers of conception, the first is the greatest: language itself. But for Chomsky and Fodor, there has always been a need (and now we can see what it is) to deny, or at least minimize, the extent to which language is seen as a gift that enhances the powers of the naked brain. For Chomsky, what our capacity to speak shows is that we have an awesome innate "language acquisition device"--all the power is already in the brain, in effect, and learning one's native tongue is just a matter of letting the local environment set a few parameters: Dutch vs. French vs. Chinese. In case one had any doubts about whether Chomsky was just speaking of an innate talent for syntax, Fodor pushed the doctrine into semantics as well, arguing for the thesis that there is no such thing as acquiring a new concept by learning a language, or expanding one's language. All the concepts we can ever come to express are already innate in our brains at birth! Yes, Aristotle had the concepts of airplane and electron innately ensconced in his brain; he just never got around to using them. By passing the buck to biology (and then conveniently ignoring biology) Chomsky and Fodor protect the mysteries of mind from an unseemly unravelling.
Unlike the monkey, we understand the unanswered questions about consciousness. McGinn's book is, among other things, an expression and discussion of the questions, which he gives every evidence of understanding quite well. For McGinn to have a convincing case for human cognitive closure, he should provide an empirical example of some creature, human or otherwise, who can definitely understand some question, but be definitively incapable of understanding the answer. He may think he has provided such cases: "The man born blind cannot grasp the concept of a visual experience of red, and human beings cannot conceive of the echolocatory experiences of bats." (p.9) But these presumed facts are just assumed, not argued for; they are neither established nor obvious, although McGinn treats them as if they were.
But whether or not there are good arguments to support the general thesis about human cognitive closure, McGinn offers a specific reason for thinking consciousness is forever outside our ken, and it has nothing to do, he claims, with how big the "frontal lobes" of our descendants might become. He argues that consciousness has what he calls a hidden structure, and this structure is systematically inaccessible from either of the two available routes: the first-person point of view of soliloquy or introspection and the third-person point of view of science. For science to explain consciousness, it would have to create an intermediate level at which to describe the hidden structure--something our science couldn't do.
Neither phenomenological nor physical, this mediating level would not (by definition) be fashioned on the model of either side of the divide, and hence would not find itself unable to reach out to the other side. Its characterization would call for radical conceptual innovation (which I have argued is probably beyond us). Since it would not be characterized by concepts familiar from either side of the psychophysical nexus, however, extended, it would not simply raise the same old problem again in a new form. The operative properties would be neither at the phenomenal surface nor right down there with the physical hardware; they would be genuinely deep and yet they would not simply coincide with physical properties of the brain. Somehow they would make perfect sense of the psychophysical nexus, releasing us from the impasse that seems endemic to the topic. They really would explain how it is that chunks of matter can develop an inner life. (p.102-3)
I quote this passage at length because it is in fact an excellent description of exactly the set of concepts that are being developed in cognitive science--at the "software" or "virtual machine" level of description favored by Artificial Intelligence, but also increasingly exploited in cognitive psychology and computational neuroscience. McGinn has carefully--and correctly--deduced just the properties that scientific concepts must have if they are to offer a genuine explanation of consciousness, but then he neglects to look to see if any such concepts have been developed by the relevant sciences! He mentions various topics and researchers in cognitive science in passing, but apparently thinks a detailed examination would be unable to shed light on his problems.
There is one exception: he pauses to give "a brief flavour of the empirical findings" regarding the phenomenon of blindsight, and gets them wrong (pp.110-111). Blindsight subjects have had portions of their visual cortices destroyed, leaving large scotomata or blind areas in their visual fields; they are usually aware of nothing that occurs within the boundaries of the scotoma (swiftly moving stimuli are a common exception), but if cued in a "forced choice" situation (in which the experimenter tells them both when to guess, and which two categories to choose between), subjects can sometimes guess well above chance, while still denying any visual experiences on which to base their guesses. The discriminable categories are very limited, the differences between the stimuli must be huge and simple, and, once again, the subjects must be cued or prompted to make a guess. McGinn, however, neglects to note these severe limitations on this guessing capacity, leaving the uninitiated reader to suppose that, perhaps, a blindsight subject might recognize a friend or read a newspaper headline held in the blind area, all the while enjoying no visual "qualia." Summing up a hypothetical case, he says, "She appears to others to be (partially) sighted, but she takes herself to be blind. Behaviorally, she can function much like a sighted person; phenomenologically, she strikes herself as blind." This is at best an exaggeration, and a little later he builds on it: "Besides, let us be naive for a minute, do blindsight patients not look very much as if they are having visual experiences when they make their surprising discriminations? . . . They do not look the way people look when there is nothing experiential going on."(p.112) On the contrary, it is precisely because they have to be prompted to make their guesses that they do look the way people look when there is nothing experiential going on. This may seem a minor point, but in fact the plausibility of McGinn's interpretation of blindsight as support for his thesis depends crucially on the reader's jumping to such an exaggerated and oversimplified picture of the behavioral talents of blindsight subjects. It permits McGinn to draw a boundary between two sorts of properties in a place that his view requires but the actual facts about blindsight oppose.
McGinn tells us at one point that his argument for a certain contrast presupposes a certain conception of natural kinds--Kripke's and Putnam's--and adds: "Those who reject that conception are unlikely to appreciate the intended contrast; they should not read on." (p.127)) Disobeying his instructions, one can read the whole book in a conditional spirit: if one restricts oneself to the assumptions, concepts and methods that McGinn favors, what are the results? McGinn's philosophical method is to reflect informally on what seems to him to follow from how the issues seem to him. In one crucial two-page stretch of his argument (pp.8-9), I count seven occurrences of "seem", each of which introduces a claim that would be challenged by one who has been brainwashed by cognitive science (to put it in the negative light that McGinn's method suggests). If the "seems"-operators were removed, McGinn's sentences would be turned into blunt assertions that would definitely stand in need of further support. Why does McGinn put everything in the "seems" mode? To avoid objections? After all, one might say, if McGinn is only reporting how things seem to him, how can one quarrel with that? One needn't. Indeed, one can even agree with him. In a backhanded way he has proven his case: armed only with the methods and concepts of traditional philosophy of mind, one cannot explain consciousness. But we've known that for a long time.
1. For an overview, see my "Granny's Campaign for Safe Science," in B. McLaughlin and G. Rey, ed., Fodor and his Critics, Blackwell, 1991.