Living on the Edge

Daniel C Dennett

In a survey of issues in philosophy of mind some years ago, I observed that "it is widely granted these days that dualism is not a serious view to contend with, but rather a cliff over which to push one's opponents." (Dennett, 1978, p.252) That was true enough, and I for one certainly didn't deplore the fact, but this rich array of essays tackling my book amply demonstrates that a cliff examined with care is better than a cliff ignored. And, as I have noted in my discussion of the blind spot and other gaps, you really can't perceive an edge unless you represent both sides of it. So one of the virtues of this gathering of essays is that it permits both friend and foe alike to take a good hard look at dualism, as represented by those who are tempted by it, those who can imagine no alternative to it, and those who, like me, still find it to be--in a word--hopeless.

The seven essays arrange themselves in such a way as to span the cliff edge handily. At one extreme, Clark and Sprigge are well over the edge, hovering, like cartoon characters held aloft by nothing but the strength of their convictions. It would be a crime to disillusion them. In the middle are Foster, and Fellows and O'Hear, utterly unpersuaded by my version of functionalist materialism, and willing to defend dualist (or apparently dualist) positions positively and vigorously, without begging the question against my alternative. Then there are the critics, Lockwood, Seager and Siewert, whose sympathies lie (in the main) with the others, but who do not commit themselves to any solution to the problems, dualist or otherwise, but concentrate more on flaws they think they detect in my arguments.

The idea of detecting flaws in my arguments must seem risible to Clark and Sprigge, who sometimes find it sufficient rebuttal simply to paraphrase one of my claims and append an exclamation point to it. But even this is useful; it goes to confirm one of my main claims: some serious thinkers find it impossible even to entertain my hypotheses. Not only do their incredulous dismissals testify to the vigor of the horses I am beating (something about which doubts have been expressed in some quarters), but they also provide independent benchmarks for re-calibrating my responses to the most persistent objections. For instance, my fictional Otto has been assailed as a stooge by some critics, in spite of the fact that his speeches are all, in fact, tightened up versions of actual objections raised against the penultimate draft. Here Otto finds friends galore. Lockwood, bless him, even describes him as my philosophical conscience!

All of the essays provide valuable clarifications and innovations--there is not a workaday or routine exercise in the lot--and I am proud to have provoked such a variety of contributions to our vision of these issues. There is a good deal of useful overlap, with several themes of mine attacked from slightly different angles, and I think the best way of exploiting this is to start with the most radical, over-the-edge (if not over-the-top) opposition, and work my way back to solid ground, occasionally skewing the order to take advantage of converging lines of attack.

Stephen R. L. Clark

Why may I not insist, against Dennett, that I do indeed know that I intend and feel, or that I know it better than I can possibly know the truth of any neurological theory he propounds?

In the main, Clark leans on Searle, and in one of the few passages in his essay that comes to grips with an argument of mine, he rather seriously misrepresents what I say in rebuttal of the Chinese Room. He says:

Dennett simply composes a set of witty conversations, which, he says, might be reproducible by computers responding solely to syntactical cues. He then encourages his readers to believe that such conversations are 'proof' that such (wholly imaginary) computers 'understand' as well as we.

it is no longer obvious, I trust, that there is no genuine understanding of the joke going on. Maybe the billions of actions . . . produce genuine understanding in the system after all. If your response to this hypothesis is that you haven't the faintest idea whether there would be genuine understanding in such a complex system, that is already enough to show that Searle's thought experiment depends, illicitly, on your imagining too simple a case, an irrelevant case, and drawing the 'obvious' conclusion from it.(p.438)

The affinity between Searle and self-professed dualists like Clark has always been an unholy alliance (if I may put it that way). Searle has always insisted that his position is not dualistic at all, that he is a good materialist, that the presence or absence of a soul or spirit has no bearing on the behavioral competence of a physical body so far as he's concerned. Those who think that the presence of "spirit" or "soul" gives us powers to act that no computer could even mimic should thus be indifferent to Searle's strange thesis, but so desperate are they, I suppose, to find champions who will keep evil AI at bay, that they forgive Searle for conceding their chief disagreement to the opposition. This has always posed something of a diplomatic problem for Searle, who hotly denies he is a dualist whenever challenged by us hardheads, but seems to have been less eager to alert his cheering dualist supporters to the embarrassing fact that they have entirely missed his point.

The other interesting feature of Clark's essay is his criticism of Dawkins' memes on what purport to be biological grounds: disanalogies between memes and genes. Dawkins does not make a genotype/phenotype distinction for memes, nor are there identifiable loci for memes. Neither of these are clearly shortcomings of the meme concept--rather than just differences--nor does Clark give any reason for supposing these disanalogies couldn't be removed if there were grounds for doing so. In fact, in a commentary (forthcoming a) on a new paper by Dawkins (forthcoming), I note that something like the genotype/phenotype distinction might yield a readily achieved improvement in his concept. Moreover, a functional (not anatomical) notion of locus for memes is clearly definable--whether or not it would prove useful is an interesting question. But that is not what Clark really objects to about memes; what he really objects to is that "It is just this sense of a divine intellect that is missing in meme-theory, and with it any respect for truth." But as Clark must know, this is a question-begging assertion, for some of us view this absence of the divine intellect as a prerequisite for any acceptable theory of meaning and truth, and respect for truth is the source of our conviction.

T. L. S. Sprigge

I am grateful to him for coming to the defense of "what Dennett abusively calls 'figment'," for he thereby renders explicit, for all to see, the covert reasoning that I have otherwise had to impute before criticizing. Here is his argument: First he draws a threeway distinction illustrated by

(1) my image of my friend

(2) my friend as I currently imagine him

(3) my friend as he really is

and goes on to claim that "the important thing here is that one should distinguish components of consciousness from the object these components function to set before me (in ways which differ significantly in perception from in thought and imagination)."

I agree; the actual components of whatever-it-is that is the medium of conscious content must be distinguished from the intentional objects thereby constituted or represented by those vehicles (whatever they are) of content. Sprigge goes on:

Of course if I think in a merely verbal way of a blue cow no component of consciousness will be blue in any sense at all, but if I imagine a blue cow my image, though it is not the blue cow I am imagining, has a certain quality which I am prepared to call 'blue' myself, though pedants may argue about its proper label till the cows, of whatever colour, come home.

We agree that people can have beliefs about the components of their consciousness--he expresses quite a few such general beliefs of his own, as he notes. In such instances, the components themselves are the intentional objects of those (reflective) beliefs. Now the question that must divide us is this: are they merely intentional objects or are they also real? It depends, of course, on what you believe about the components you are thinking of. Whatever you think your own components-of-consciousness are, whether you are right, I claim, is always an open, empirical question, and the answer is never obvious--it only appears to be obvious to those who think they have privileged access to the "inherent" nature of these components.

Consider the room full of Marilyn Monroe pictures. Here there is no doubt at all what the heterophenomenology is: a vast, regular array of identical, high-resolution photos of Marilyn Monroe. That is the object of your (unreflective) consciousness. Now reflect on the conscious experience itself: What components does your consciousness have, in virtue of which this is its object? Are there details among the components or are the components quite minimal? If the intentional objects of the experience at a certain moment include, say, thirty high-resolution images of Marilyn Monroe, does it follow that there are components of your consciousness for each of these thirty images? Notice that you are not in a position of authority to answer this question.

Sprigge speaks the "medium" in which even abstract thinking is conducted, and I agree that for there to be content, there must be vehicles of content--a medium in which that content is embodied.Endnote 1 I say the medium (in human beings) is neural processes, and hence the real qualities of these components are the qualities that neural processes have in the normal exercise of their functions. Being blue is not one of them, but encoding-blue-in-virtue-of-such-and-such-functional-properties can be. Sprigge would perhaps object that the medium of consciousness certainly doesn't seem to be neural processes, but if he did, I would suggest that this is because he is himself failing to make the distinction he thinks I ignore: he is mistaking the content for the medium. (This point will become clearer below, I trust, in my discussion of Seager.)

I consider Sprigge to have paid me a backhanded compliment for my arguments against qualia: he proposes to overcome them by jettisoning the standard appeal to what he calls the Humean denial of necessary connections between distinct existences. He says "once one realizes that something can both be a distinct quality of experience with its own inherent nature and also be necessarily related to certain behavioral dispositions one is released from this unpalatable choice." Sprigge is right about one thing: all the qualia arguments in the literature share the assumption that there is a contradiction in any view which says that there is a necessary connection between qualia and behavioral effects on the one hand (functionalism, "logical behaviorism") and that their identity is "intrinsic" or "inherent" on the other. I'm not known for my appeals to the sanctity of philosophical traditions regarding necessity and possibility, but when someone declares that a revolution in metaphysics is the way to evade my objections, I consider myself to have hit a nerve.

John Foster

I am similarly delighted to see that Foster thinks my arguments have secured one of my main conclusions: "Dennett's reasoning is impeccable: There is no way of preserving forms of non-cognitive presentation within a materialist framework." Of course we draw opposite conclusions from then on: I say so much the worse for "non-cognitive presentation" and he says so much the worse for materialism. Foster is not alone in reaching this verdict. One of the main themes of my book is that it is harder to be a good materialist than most casual materialists have thought, and it is been fascinating to me to see how many closet dualists I have driven into the open. This is progress no matter who "wins," for it obliges materialists to take more seriously the issues that have underlain the dualist impulse all along, while also offering dualists a perspective on how they might achieve at least some of the goals they care about without acquiescing in a theoretical cul-de-sac.

Foster shows that I was wise not to claim to have offered a definitive refutation of dualism. I am quite sure that no such refutation is possible, but that is faint praise for dualism. We can imagine lots of non-starters of which that is true, such as the theory that says that everything is just five minutes old but arranged just as if the universe was some billions of years old. Foster himself is prepared to join me in dismissing epiphenomenalism as a non-serious (though irrefutable) theory.

My main objection to dualism is that it is an unnatural and unnecessary stopping point--a way of giving up, not a research program. That is quite enough for me. Foster eventually confronts this objection, and asks why I should want to avoid dualism at all costs? Only because that is my way of keeping the scientific enterprise going. It is a self-imposed constraint: never put "and then a miracle happens" in your theory. Now maybe there are miracles, but they are nothing science should ever posit in the course of business. Temporarily unexplained whatchamacallits are often a brilliant stopgap--Newtonian gravitational attraction as "action at a distance" before Einstein, Mendelian genes before the discovery of the structure of DNA--but positing something which one has reason to believe must be inexplicable is going too far. At one point Foster discusses the prospect of treating the fundamental physical laws as "probabilistic": "We could then think of the interventionist causal role of the non-physical mind as that of selecting between, or at least affecting the probabilities of, these physically possible states." Now if, like Roger Penrose (1989), Foster supposed this "interventionist causal role" was explicable in principle, he would be advocating, like Penrose, a non-conservative, expansionist materialism, but he goes on to deny this, and he notes, correctly, that this would be enough to make the view anathema to me.

This raises Foster's main point: isn't this way of characterizing the difference between unacceptable dualism and tolerable expansionist materialism vacuous or question-begging? Why, Foster asks, should the dualist be required to explain things more deeply than the materialist? I'd pose a more lenient demand: that the dualist offer any articulated, non-vacuous explanation of anything in the realm of psychology or mind-brain puzzles. Since I am simply proposing a constraint on what sort of theory to take seriously, it really doesn't matter to me (except as a matter of communicative convenience) whether the term "dualism" is defined in such a way as to permit varieties of dualism to meet the constraint. Indeed, Nicholas Humphrey (1992) declares that his position is, in a certain sense, a kind of dualism, and yet since it undertakes to meet the demands of objective science, I consider it radical, but eminently worthy of attention now--not a theory to postpone till doomsday. And if Penrose were to declare that his position, too, was really a sort of dualism, and if this understanding of the term caught on, I'd want to shift nomenclature and find some new blanket pejorative for theories that tolerate "and then a miracle happens."

This sheds light on Foster's claim that my "prior rejection of dualism" is "the basis" for my denial of the inner theater.

What enables Dennett to represent his functional approach as correct is that, given the falsity of Cartesian dualism, there is no possibility of finding a 'central conceptualizer and meaner' to be the subject of the irreducible cognitive states and activities which our initial (anti-functionalist) intuitions envisage; and without such a subject, there is no serious rival to an account of cognition along functionalist lines.

Yes, one could say that I diagnose dualism as a sort of false crutch for the imagination; it gives people the illusion that they can understand how there could be a "central conceptualizer and meaner." A case in point: Foster asks

If the mind is a non-physical substance, what is its intrinsic nature? But it seems to me that the dualist has no special problems here. To begin with, I do not myself see why the dualist needs to admit that there is anything more to the nature of the mind than what introspection can reveal.

You argue very persuasively that there aren't hundreds of high-resolution Marilyns in the brain, and then conclude that there aren't any anywhere! But I argue that since what I see is hundreds of high-resolution Marilyns, then since, as you argue, they aren't anywhere in my brain, they must be somewhere else--in my nonphysical mind! (p.359)

Foster's account of my view is largely accurate and sympathetic, in spite of his deep opposition to it, but there is one point where he slightly misrepresents it. He is right that I argue against any such thing as "distinctively presentational, non-cognitive awareness" but it is not quite right to say that my view is "that consciousness, whether sensory or introspective, is purely cognitive--a matter of acquiring beliefs or making judgments," since I grant that contentful states of all sorts occur, and have effects that are in their varying degrees and ways constitutive of consciousness. The disruptive effect of pain, for instance, is not "cognitive" in itself; it is surely one of the most weighty factors in the family of dispositional factors that make something pain. In my discussion of Lockwood, I will review and expand upon the elements of my view that make Foster almost right.

Fellows and O'Hear

It is always reassuring to encounter accurate summaries of one's views by one's critics, and Fellows and O'Hear come up with some expressions that actually improve on the source text. For instance their version of the Cartesian Theatre is non pareil:

In the Cartesian Theatre, what seems to me is. If I seem to be a unitary self, then I am. If I seem to be seeing a bent stick or a pink elephant, then I am seeing such things--not in the external world, it is true, but in my private theatre nonetheless.

They go on to attribute to me the "thesis that there is no more to experience than thinking, or that seeing is believing." This echoes Foster's reading (as just noted) and, again, I will defer my response of this crucial point to the discussion of Seager, and concentrate first on their alternative and its difficulties.

Dennett will say to us that our thought that we are more than zombies is an illusion we suffer from owing to a defective set of metaphors we use to think about the mind.

Exactly right.

We hope to show, by contrast, that Dennett's replacement metaphors leave out something which is essential to us as human beings.

They see exactly the point of my rejection of Searle's Chinese room, and they accurately recount my argument against it, but say it misses the point. Searle's argument has to do "with the ascent from formal or syntactic operations to semantical ones." Their discussion betrays a variety of naive misconceptions about computer science in general and AI in particular, but let's concentrate on the conclusion, not the path: programmes, they say, are intrinsically syntactical and only extrinsically semantical or meaning-bearing; "symbol shuffling by itself does not give any access to the meanings of the symbols." Access to what or to whom? I would say (deliberately adopting Searle's slanted terminology) that symbol shuffling is what makes there be something that could have access to, want access to, need access to, the meanings of symbols. They say "computer programmes need minds to read them, hence minds cannot be computer programmes merely" This just begs the question, of course, but it hints at the bottom line of their objection:

Dennett cannot, then, on pain of circularity, say that it is just another text or set of texts which gives texts meaning. . . . At least some texts will have to have the meaning-conferring properties of selves and agents.

Philosophers often maneuver themselves into a position from which they can see only two alternatives: infinite regress versus some sort of "intrinsic" foundation--a Prime Mover of one sort or another. There is always another alternative, which naturalistic philosophers should look on with favor: a finite regress that peters out without marked foundations or thresholds or essences. Here is an easily avoided paradox: every mammal has a mammal for a mother--which implies an infinite genealogy of mammals (which cannot be the case). The solution is not to search for an essence of mammalhood that would permit us in principle to identify the Prime Mammal, but rather to tolerate a finite regress that connects mammals to their non-mammalian ancestors by a sequence that can only be partitioned arbitrarily. The reality of today's mammals is secure without foundations. (For more on this theme in my work, see Dennett, forthcoming b.) In this instance, the solution is to show via a finite regress (or progress, if one works "bottom up') how it can be the case that "intrinsically syntactic" mechanisms ultimately compose systems that deserve to be called semantic engines, capable of "seeing [each other's] outpourings in semantic terms."

The last trump played by Fellows and O'Hear is an unflinching defense of the reality of the self: "Each of us can say that unless . . . there was a genuine sense of an 'I' as the centre of my experience, there would be no way to fix or locate a particular set of thoughts or texts as the place from which I operated." What puzzles me about this argument is that it seems quite obvious to me that everything they say about indexicality and confusion about location applies to zombies--we just have to add scare-quotes to avoid begging the question. Zombies are presumably as subject to the disruptive disorders of Multiple Personality Disorder and scatter-brainedness as we are, after all, and when they manage to rise above these afflictions, it is because they have a well-designed "sense of an 'I'." A zombie can "wonder" where he is, "discover" that the hand in the yellow glove is not his own hand after all, and "recognize" himself in a mirror. I don't see that Fellows and O'Hear have offered any reasons for dismissing the solution to the problem of indexicality I proposed: you are what you control. It's as simple as that. Or as I put it: "How come I can tell you all about what [goes] on in my head? Because that is what I am; . . . a knower and reporter of such things in such terms is what is me." (p410)

Finally, after an excellent summary of my position, they say that I fail to explain thereby "how it is that Otto could be said to be mistaken unless there was an Otto who was, in the sense Dennett wants to rule out, the subject of the deception." Let's consider this objection in more detail. Suppose Otto were a mere zombie. Then he would be mistaken in his pseudo-beliefs (the pseudo-beliefs he "expresses" in the speeches I have given him). If an unconscious zombie can have a "belief", he can have a mistaken "belief". Fellows and O'Hear, like Searle and many others, have already conceded that there would be a certain explanatory robustness and systematicity to scare-quoted talk about the "beliefs" and other pseudo-mental states of zombies. It falls to them therefore to show that there is a further problem with the attribution of particular sorts of beliefs--e.g., indexical beliefs or higher-order beliefs or mistaken beliefs--and I see no grounds being given.

Michael Lockwood

The most strikingly pre-emptive criticism I have encountered to my theory is simple: I have (quite obviously) not explained consciousness at all; I have left out consciousness; I have "side-stepped" the central puzzle.Endnote 2 I have missed the whole point. Now this would be a rather strange sort of neglect, even by the standards of neuropsychology--to set out to write a book explaining consciousness, and to write a book that actually accomplishes something but nevertheless entirely overlooked the project it putatively set out to cover. If I told you someone had written a book entitled The Arab-Israeli Conflict Solved, and that it somewhat surprisingly neglected to mention the fact that the Arabs and Israelis dispute the right to certain tracts of land in the Middle East, you would probably conclude that the author must be insane (or post-modern, if that means anything different).

But so alien is my explanation of consciousness to many readers that they do not even recognize it as a flawed explanation, or a refuted explanation of consciousness--they don't see that I have even tried to tackle what they consider the nub. Sometimes the confidence with which this view is expressed amuses me. It reminds me of an encounter I just learned about from a colleague. A team of medical educators recently returned to Boston from Mexico, where they had been engaged in a project to teach the principles of birth control to uneducated women in remote areas. Armed with all the latest audio-visual equipment, they had held large groups of these women spellbound with their videotapes of microscopic sperm wriggling their way towards an ovum, computer-animated diagrams of conception, and so forth. After one presentation, one of the spectators was asked her opinion. "It's really very interesting how you people make babies," she replied, "but here we don't do it that way. You see, our men have this milky fluid that comes out of their penises . . . " It is almost embarrassingly obvious to Sprigge and Clarke, for instance, that I could not be talking about consciousness--the consciousness they know so well--and so they are tempted to conclude that I must be a very different sort of animal!

Which is harder to credit: that I would write a book that didn't even try (in spite of its title), or that they (and so many of them!) would fail to see an attempt as an attempt? I could point to those who do see my book as offering not only an explanation but a good one, but the other side will clearly view them as taken in by my slippery rhetoric, lulled to sleep by tricks and examples. Somebody is missing something big; who is missing what?

Faced with such a curious question, I find Lockwood's essay a godsend, for he sees exactly what my attempt at explanation is, and sees that it is such an attempt, and such a radical one that he can scarcely believe that I mean it. But he gives me the benefit of the doubt, thank goodness. His first line of attack concerns the consciousness of animals and infants. This is a frequently voiced objection to my theory of consciousness as a culture-borne virtual machine: isn't my theory refuted by the obvious fact that animals bereft of culture (and, of course, newborn human infants) are conscious? Lockwood, appealing as so many do to Nagel's "what it is like to be" formula, says:

Consciousness in this sense is presumably to be found in all mammals, and probably in all birds, reptiles and amphibians as well.

It is the "presumably" and "probably" that I want us to attend to. Lockwood gives us no hint as to how he would set out to replace these terms with something more definite. I'm not asking for certainty. Birds aren't just probably warm-blooded, and amphibians aren't just presumably air-breathing. Nagel confessed at the outset not to know--or to have any recipe for discovering--where to draw the line as we descend the scale of complexity (or is it the cuddliness scale?). This embarrassment is standardly waved aside by those who find it just obvious that there is something it is like to be a bat or a dog, equally obvious that there is not something it is like to be a brick, and unhelpful at this time to dispute whether it is like anything to be a fish or a spider (to choose a few standard candidates for the median).

Fellows and O'Hear put the same point somewhat more circumspectly:

animals and human infants seem to be conscious perfectly well without the mediation of any culturally acquired 'software'.

How, though, could we ever explore these "maybes"? We could do so in a constructive, anchored way by first devising a theory that concentrated exclusively on human consciousness--the one variety about which we will brook no "maybes" or "probablys"--and then look and see which features of that account apply to which animals, and why. There will still be plenty of time to throw out our theory if and when we find it fails to carve nature at the joints, and we might just learn something interesting.

Forget culture, forget language. The mystery begins with the lowliest organism which, when you stick a pin in it, say, doesn't merely react, but actually feels something.

Bats, like us, have plenty of relatively peripheral neural machinery devoted to "low level processing" of the sorts that are routinely supposed to be entirely unconscious in us. And bats have no machinery analogous to our machinery for issuing public protocols regarding their current subjective circumstances, of course. Do they then have some other "high level" or "central" system that plays a privileged role? Perhaps they do and perhaps they don't. Perhaps there is no role for such a level to play, no room for any system to perform the dimly imagined task of elevating merely unconscious neural processes to consciousness. Lockwood says "probably" all birds are conscious, but maybe some of them--or even all of them--are rather like sleepwalkers, or non-zimbo zombies! The hypothesis is not new. Descartes notoriously held a version of it, but it is Julian Jaynes (1976) who deserves credit for resurrecting it as a serious candidate for further consideration. It may be wrong, but it is not inconceivable--except to those who cling to their traditions as if they were life-rafts.

And what of the "one great blooming, buzzing confusion" of infant consciousness? (James, 1890, p.462.) Well, vivid as James's oft-quoted (and misquoted) phrase is--a rival on the philosophy hit parade for Nagel's formula--it manifestly presumes more than any cautious investigator would claim to be able to support. That the inchoate human brain is unorganized to some degree is not in doubt; that this incipient jumble of competing circuits is experienced as anything at all by the infant is the merest presumption. It may be, and then again, it may not. The standard working assumption appealed to by Lockwood and Fellows and O'Hear doesn't let us consider these as open hypotheses, in spite of the considerable scientific grounds for doing so. At least some animals and infants seem to be conscious in just the way we adults do, but when we adopt an investigative strategy that first develops an articulated theory of adult human consciousness, and then attempt to apply it to other candidates (as I do in the last chapter), it turns out that appearances are misleading at best.

In particular, the very idea of there being a dividing line between those creatures "it is like something to be" and those that are mere "automata" begins to look like an artifact of our traditional presumptions. Since in the case of adult human consciousness there is no principled way of distinguishing when or if the mythic light bulb of consciousness is turned on (and shone on this or that item), debating whether it is "probable" that all mammals have it begins to look like wondering whether or not any birds are wise or reptiles have gumption. Of course if you simply will not contemplate the hypothesis that consciousness might turn out not to be a property that sunders the universe in twain, you will be sure that I must have overlooked consciousness altogether, since I entertain and even defend this hypothesis.

Lockwood recognizes that my defense of the scarcely credible hypothesis involves denying "the reality of the appearance itself." Like Siewert, whose views I will turn to next, Lockwood appreciates the pivotal role of the example of blindsight in my campaign against "real seeming." The standard presumption is that blindsight subjects make judgments (well, guesses) in the absence of any qualia, and I use this presumption to build the case that ordinary experience is not all that different.

Dennett's position, in effect, is that it is only in degree that normal sight differs from blindsight. Normal sight carries with it far greater confidence in the corresponding judgments, and is of vastly greater discriminative power; but there is, in the end, no qualitative difference between that and blindsight.

So what are we supposed to be doing? Simply generating new judgments and checking the old ones against them? Surely not. Judgments, Otto would insist, are too anaemic, too high-level, too intellectual to do duty for the substance of sensation and perception.

What are judgments, then, if they are not to be modeled on sentences expressed to oneself? Haven't I myself on occasion called them propositional episodes? Yes, and I beg to remind my fellow philosophers that propositions, officially, are not the same as sentences in any medium, and as abstractions they come in all "sizes." There is no upper bound on the "amount of content" in a single proposition, so a single, swift, rich "propositional episode" might (for all philosophical theory tells us) have so much content, in its brainish, non-sentential way, that an army of Prousts might fail to express it exhaustively in a library of volumes.

Is it "remotely credible" that "seeing is believing"? Lockwood's Otto is incredulous because he has fallen for some covert (or "sophisticated") version of the following: Seeing is like pictures, and believing is like sentences, so since a picture is worth a thousand words, seeing could not be believing!

 If you think that the contrast between "merely verbal" and "imagistic" (Sprigge) secures a distinction between (informational) content and quality, you are dismissing a major theoretical option without trying properly to imagine it. (For more on this theme, see Dennett forthcoming c.)

Charles Siewert

 Siewert, in his scrupulous, ingenious essay, shares with the other authors the reluctance to abandon qualia (or, in his terms, "visual quality") and, like Lockwood, he sees my discussion of blindsight as a weak link in my chain of arguments. I have claimed that there is no gradual story that can be coherently told that takes us from actual blindsight to zombiehood. Siewert accepts the challenge:

But now if we can imagine a sort of minor loss of consciousness with conscious-like responsiveness intact in the case of unprompted blindsight, then why not suppose this sort of loss gradually augmented, so that the variety and extent of consciousness diminishes finally to nothing, while the behavior remains that of a conscious human being? If there is no conceptual obstacle to this, we arrive in piecemeal fashion at the notion of a totally unconscious morpho-behavioral homologue to ourselves--the dreaded zombie. . . .

Let us first catalogue the differences that have to be traversed as we move from actual blindsight to the target of zombiehood (for the moment, just partial zombiehood--visual zombiehood). Actual blindsight subjects need to be prompted. They claim they see nothing (in their blind fields), and moreover, they don't spontaneously volunteer any judgments, or modulate their nonverbal actions on the basis of visual information arising from the blind region. Actual blindsight subjects exhibit sensitivity to a very limited or crude repertoire of contents: they can guess well about the shape of simple stimuli, and very recently (since Consciousness Explained was published) evidence of color-discrimination has been secured by Stoerig and Cowey (1992), but there is still no evidence that blindsight subjects have powers beyond what can be manifest in binary forced choice guessing of particular widely-dispersed properties. (No one has yet shown delicate color discriminations in blindsight--or even the capacity to tell a red cup on a green saucer from a green cup on a red saucer, for instance.)

My contention is that what people have in mind when they talk of "visual" consciousness, "actually seeing" and the like, is nothing over and above some collection or other of these missing talents (discriminatory, regulatory, evaluative, etc.); I don't know where to "draw the line"--I leave that to those who disagree with me--but I should think that any believer in visual properties is going to become embarrassed at some point in the traverse from actual blindsight to partial zombiehood. Let us look at Siewert's path. He asks you to imagine having a blindsight scotoma, but noticing that "you are on occasion struck by the thought, as by a powerful hunch or presentiment, that there was something just present (say, an X) in the area corresponding to your deficit." This thought, though conscious, would not be an instance of a conscious visual experience, however accurate and reliable, he claims, and he supposes that I would say it is a "terrible mistake" to claim to be able to imagine this prospect. But I agree that it is readily imaginable. I agree that if I found myself having such hunches, and grew to rely on them, I would still be unlikely to consider them instances of visual consciousness--but only because they are (as imagined by Siewert) so poor in content: an X or even an X suddenly moving left to right, and currently just about there or even a pink X. The paradigmatic presentiment is a content-sparse propositional episode, while vision is paradigmatically rich.

Siewert finds this line of mine unpersuasive. He even sees as "evasive" what I consider to be an essential move. Let me review the bidding: we're talking about an intuition-pump transition, excellently oriented and posed by Siewert, and I have drawn attention to a feature that is, I claim, doing the dirty work: the tacit assumption that the "amount of content" or whether a discriminative talent is "high-grade" (Siewert's term) makes no difference. This is the knob we must turn this way and that to see what happens. Let's try, slowly.

Can I imagine having a presentiment, lacking all visual quality, but with a full serving of visual content?

(A) It suddenly occurred to me that there was a wad of crumpled paper lying on the floor, shaped remarkably (if viewed from this angle) like a sleeping kitten, except that (it suddenly occurred to me) the sun was glinting off the edges just so, and this led me to have the further hunch that if I squinted, the wad of paper seemed to be exactly the same color as my bedspread over there. . . but of course there was nothing visual about my experience--I'm blind!

If I found myself in the imagined predicament, I might well panic. In a weak moment I might even convert, and give up my own theory, But that is just to say that in fact I not only can imagine that my theory is wrong; I can even imagine myself coming to believe that it is wrong. Big deal. I can also imagine myself having the presence of mind in these bizarre straits to take seriously the hypothesis that my own theory favors: I'm deluding myself about the absence of visual quality. That might even seem obvious to me.

So it is not, as Siewert realizes, a simple question of what he or I can and can't imagine. I have argued against the familiar idea among philosophers that blindsight offers a clean example of visual function without visual quality--a secure first step towards taking the concept of zombies seriously. I don't consider myself to have given a conclusive a priori argument against this idea, but just to have offered a plausible alternative account that explains (I claim) the same primary phenomena--and the secondary phenomena: the tendency of philosophers to overlook my account.

Siewert sees that there is an ominous stability (ominous by his lights) to my position, and he diagnoses its dependence on an epistemological position of mine he calls "third-person absolutism." As one who thinks absolutism of any sort is (almost!) always wrong, I heartily dislike the bloodcurdling connotations of this epithet, but I think he's got my epistemological position clear. Whatever the position is called, it is not a rare one. It is, in fact, the more or less standard or default epistemology assumed by scientists and other "naturalists" when dealing with other phenomena, and for good reason. As he notes, he has yet to work out the details of a defense of an alternative that doesn't slide into solipsism or something equally bad.

While waiting for him to compose a justification for his novel epistemology, with its "distinctive warrant", I propose a pre-emptive strike--at any rate a glancing blow. According to Siewert's neutral epistemology, certain things are conceivable that are not (or not clearly) conceivable according to standard "third-person absolutism". And so . . .? Would this show that these things are actually possible, or would it show that this novel epistemology is too lenient? Should science take these newly conceived possibilities seriously? Why? The "neutrality" of his proposed vantage point midway between traditional first-person authority and traditional third-person objectivity is fragile. At least the old infallibility doctrine had a certain self-supporting chutzpah in its favor.

William Seager

Seager's essay is constructed around the attempt to present me with a dilemma: "either his verificationism must be so strong as to yield highly implausible claims about conscious experience, or it will be too weak to impugn the reality of actual phenomenology." Like others (e.g., some of the commentaries in Behavioral and Brain Sciences on Dennett and Kinsbourne, 1992, and Baars and McGovern, forthcoming), Seager wants to diagnose the central move of my theory as a more or less standard verificationist power play--too strong to command assent. In the months that have intervened since my book went to press, I've composed several corrective passages that put my apparent arch-verificationism in better light. (Those darn Multiple Drafts--there's just no keeping up with them!) They are tailor-made, it turns out, to meet Seager's objections, so, with apologies for self-quotation, I will repeat them below, since I think it is valuable to get all this point and counterpoint together in a single place.

As Seager shows very clearly, in his careful discussion of the color phi case, the difference between his H1 and H2 is that while experience is "generated" in both cases, this happens either before (Orwellian) or after (Stalinesque) the binding of the mid-trajectory color shift. And he notes that my account "does not require consciousness at all"--that is, does not require the "generation of experience" as a separate component in the manner of H1 and H2. He sees this "disturbing", since it seems to him to imply that all is dark inside. But of course! We know that to be true. There is nothing "luminous" (as Purvis put it) going on in the brain.

The conviction that this extra, well-lit process must occur is, of course, a persistent symptom of Cartesian Materialism, and Seager's view is illuminated (if I may put it that way) by Lockwood's ingenious dramatization. Lockwood imagines a version of my Multiple Drafts Model that retains the Cartesian Theater, as a stage (or at least a "pool of light") in which an "avant garde" play is performed, complete with inconsistent flashbacks, revisions, and tampered "instant replay" video. This does indeed preserve, as Lockwood claims, all the curious temporal features I used to disparage the Cartesian Theater, without abandoning the presentation process. I welcome this elaboration, for it lays bare the fundamental problem: Lockwood's troupe is not avant garde enough! Why should they bother with the boring and bourgeois ritual of actually presenting the play (with actors in costume, etc.) when all they really have to do is send their real-time script-revisions directly to the critics, libraries, and subscribers? The content is all there, in time to have its apposite effects (their play will seem to have been performed), and you save a fortune on lighting! There has to be some extra role for presentation in this special medium, and Lockwood offers nothing but "common sense" in favor of the need for such a process or such a medium.Endnote 4

Seager tries to fill this gap by developing what he considers an embarrassment, if not a reductio. He formulates

(H3) There is conscious experience

(H4) There is no conscious experience, but (false) memories of conscious experience are being formed continuously . . .

and notes that it follows that I cannot distinguish these. He is right; H3 and H4 are just a different way of stating the apparently rival hypotheses that you are conscious or that you are a zimbo who only mistakenly thinks he's conscious, and my conclusion is indeed that the apparent difference between these hypotheses is an artifact of bad concepts. I should have made this more explicit in the book, but it took a critic, Bruce Mangan (forthcoming), to distill the essence of the point. Consciousness, he proposes, is "a distinct information-bearing medium":

He points out that there are many physically different media of information in our bodies: the ear drum, the saline solution in the cochlea, the basilar membrane, each with its own specific properties, some of which bear on its capacity as an information medium (e.g., the color of the basilar membrane is probably irrelevant, but its elasticity is crucial). "If consciousness is simply one more information-bearing medium among others, we can add it to an already rather long list of media without serious qualms."

And here is Mangan's hypothesis: when it comes to prosthetic replacements of media, all media are fungible in principle except one: the privileged central medium of consciousness itself, the medium that "counts" because representation in that medium is conscious experience. What a fine expression of Cartesian materialism! I wish I had thought of it myself. Now neurons are, undoubtedly, the basic building blocks of the medium of consciousness, and the question is: are they, too, fungible? The question of whether there could be a conscious silicon-brained robot is really the same question as whether, if your neurons were replaced by an informationally-equivalent medium, you would still be conscious. Now we can see why Mangan, Searle, and others are so exercised by the zombie question: they think of consciousness as a "distinct medium", not a distinct system of content that could be realized in many different media. . . . The alternative hypothesis, which looks pretty good, I think, once these implications are brought out, is that, first appearances to the contrary, consciousness itself is a content-system, not a medium. And that, of course, is why the distinction between a zombie and a really conscious person lapses, since a zombie has (by definition) exactly the same content-systems as the conscious person. (Dennett, forthcoming d)

. . . certain sorts of questions about the British Empire have no answers, simply because the British Empire was nothing over and above the various institutions, bureaucracies and individuals that composed it. The question "Exactly when did the British Empire become informed of the truce in the War of 1812?" cannot be answered. The most that can be said is "Sometime between December 24, 1814 and mid-January, 1815." The signing of the truce was one official, intentional act of the Empire, but the later participation by the British forces in the Battle of New Orleans was another, and it was an act performed under the assumption that no truce had been signed. Even if we can give precise times for the various moments at which various officials of the Empire became informed, no one of these moments can be singled out--except arbitrarily--as the time the Empire itself was informed. Similarly, since You are nothing over and above the various subagencies and processes in your nervous system that compose you, the following sort of question is always a trap: "exactly when did I (as opposed to various parts of my brain) become informed (aware, conscious) of some event?" Conscious experience, in our view, is a succession of states constituted by various processes occurring in the brain, and not something over and above these processes that is caused by them.(Dennett and Kinsbourne, 1992b, p.235-36.)

This point is even better illustrated in response to Seager's discussion of dreams. He astutely observes that much of the theoretical apparatus of my 1976 paper, "Are Dreams Experiences?" foreshadow the analyses in Consciousness Explained--right down to my apparently outrageous suggestion, way back then, that one might dream a dream backwards but remember it back-to-front, a bit of elbow room for the brain that is not just possible in principle, but (I now claim) necessary in practice. What about forgotten dreams? Is it the case that no test could reveal whether we had them or not? No, there are lots of imaginable tests that could determine whether or not, while you slept, a particular narrative was activated, composed, re-activated, rehearsed, etc.--while remaining entirely inaccessible to waking recollection in the morning. What would be beyond testing is the apparent distinction between this all going on entirely unconsciously and going on in the consciousness of dreams.

So I can appeal to the findings of sleep researchers (of the future--as Seager says, the REM findings are nowhere near enough) to remove "forgotten dreams from the realm of the unverifiable"--but the price, which I for one will gladly pay, is that by "dreams" we have to equivocate (apparently) between conscious dreams and unconscious (zomboid) dreams. I am not sure why he thinks I hold there would be no way of investigating these hypotheses. I have already stipulated that all the various contents in all the narrative threads can be (in principle) identified, and their vehicles traced, timed, and located, so there will be no bar at all to the discovery of what Seager calls episodes of narrative spinning. There will be less reason than ever for calling them conscious, of course, and this was the germ of truth in Norman Malcolm's notorious claims.

This still seems verificationist, of course, but the appearance is misleading, and I now have a new way of clarifying my position, thanks to Lockwood. In a debate with me at Amherst College some months ago, Michael came up with a wonderful phrase (which appears in slightly revised form in his essay in this journal): consciousness, he said (with an air of reminding his audience of the obvious) is "the 'leading edge' of . . . memory." "Edge? Edge?" I replied, "What makes you think there is an edge?" and my response to him on that occasion has since grown into a separate paper, "Is Perception the 'Leading Edge' of Memory?" (forthcoming e). It also provoked me to compose yet another little story, which I have used to stave off this misconception in another reply to critics (Dennett, forthcoming d):

You go to the racetrack and watch three horses, Able, Baker and Charlie, gallop around the track. At pole 97 Able leads by a neck; at pole 98 Baker, at pole 99 Charlie, but then Able takes the lead again, and then Baker and Charlie run ahead neck and neck for awhile, and then, eventually all the horses slow down to a walk and are led off to the stable. You recount all this to a friend, who asks "Who won the race?" and you say, "Well, since there was no finish line, there's no telling. It wasn't a real race, you see, with a finish line. First one horse led and then another, and eventually they all stopped running." The event you witnessed was not a real race, but it was a real event--not some mere illusion or figment of your imagination. Just what kind of an event to call it is perhaps not clear, but whatever it was, it was as real as real can be.

There would be a finish line if there were, in the brain, a transduction of information into a new medium, but I have argued that there is no such transduction. The functions or competences that together compose what we think of as definitive of consciousness eventually come to apply to some of the various contents that float by in our brains; it is access to these functions, and nothing else, that puts contents into our streams of consciousness (in contrast to our streams of unconsciousness). There is a stream of consciousness, but there is no bridge over the stream!

I have tried, in these responses, to repay in kind the respect these critics have paid to my book. Better than ever I appreciate how hard it is to make oneself take seriously views one finds outrageous, how easy it is to be tempted by cheap caricature. These critics have set a good example, time and again coming up with keenly observed and constructively expressed versions of doctrines of which they are deeply skeptical. As one who is all too often deeply disappointed and embarrassed by the way my fellow philosophers snipe at each other, I would like to express my deep satisfaction with the way this encounter has come out.

When I have pounced with glee on telling turns of phrase in my opponent's essays, I hope I have managed to be as fair as they have been with me. My belief is that it is in relatively casual and unguarded choices of expression that we philosophers tend to betray what is really moving us, so opportunistic "pouncing" is an ineliminable part of philosophical method. What is required to keep it from deteriorating into cheap debating tricks and sea-lawyering is, on the side of the pouncer, a proper attention to the principle of charity, and on the side of the pouncee, a willingness to listen, to entertain the other side's points before composing rebuttals--or (wonder of wonders) concessions. It is a pleasure and an honor to count these philosophers as not just the loyal opposition, but as fellow investigators on what must be, in the end, a common project.

1.For more on the "medium" of consciousness see below, in the discussion of Seager.

 2.In his review in Science, Dale Purves (1992) claims I "sidestep" the question of consciousness in my book because what consciousness is is a "luminous and immediate sense of the present, about which we are quite certain." He never attempts to unpack that metaphor of luminosity, and in the end he allows as how "metaphor is not enough"--I quite agree.

3. Or perhaps that variety of temporal lobe epilepsy for which pronounced "philosophical interest" is known to be a defining symptom. See, e.g., Waxman and Geschwind, 1975

 4.He suggests that I am wrong about "filling in" and cites Ramachandran's recent research as bearing on this. I welcome the attention drawn to Ramachandran's work, for in fact it ends up supporting my view, not undermining it. The issues are much too involved to do justice to here, but Churchland and Ramachandran (forthcoming) present the attack in great detail, and I reply in kind in three papers (Dennett, 1992, forthcoming a and c).

Akins, K., forthcoming, "What is it Like to be Boring and Myopic?" in B. Dahlbom, ed., Dennett and his Critics: Demystifying Mind, Oxford: Blackwells.

Baars, B., and McGovern, K., forthcoming, "Does Philosophy Help or Hinder Scientific Work on Consciousness?" in Consciousness and Cognition.

 Churchland and Ramachandran (forthcoming), "Filling In: Why Dennett is Wrong," in B. Dahlbom, ed., Dennett and his Critics: Demystifying Mind, Oxford: Blackwells.

Dawkins, R., forthcoming, "Viruses of the Mind," in B. Dahlbom, ed., Dennett and his Critics: Demystifying Mind, Oxford: Blackwells.

 Dennett, D. C., 1976, "Are Dreams Experiences?" Phil. Review, April, pp. 151-71.

 Dennett, D. C., 1978a "Current Issues in the Philosophy of Mind," American Philosophical Quarterly, October, pp. 249-61.

 Dennett, D. C., 1987, The Intentional Stance, Cambridge, MA: MIT Press.

 Dennett, D. C., 1992, "Filling in vs. Finding out: a ubiquitous confusion in cognitive science," H. Pick, P. Van den Broek, D. Knill, eds., Cognition, Conceptual, and Methodological Issues, Washington, DC: American Psychological Association.

 Dennett, D. C., forthcoming a, "Back From the Drawing Board," in B. Dahlbom, ed., Dennett and his Critics: Demystifying Mind, Oxford: Blackwells.

 Dennett, D. C., forthcoming b, "Self Portrait," for S. Guttenplan, ed., Companion to the Philosophy of Mind, Oxford: Blackwell.

 Dennett, D. C., forthcoming c, "Seeing is Believing--or is it?" in K. Akins, ed., Perception (Vancouver Studies in Cognitive Science, vol 5), Oxford: Oxford Univ. Press.

 Dennett, D. C., forthcoming d, "Caveat Emptor" (reply to my critics) in Consciousness and Cognition.

 Dennett, D. C., forthcoming e, "Is Perception the 'Leading Edge' of Memory?" in A. Spadafora, ed., Memory and Oblivion, Locarno Conference, Locarno, Switzerland, October, 1992.

 Dennett, D. C., and Kinsbourne, M., 1992, "Time and the observer: The Where and When of Consciousness in the Brain," Behavioral and Brain Sciences, 15, pp.183-200.

 Dennett, D. C., and Kinsbourne, M., 1992b, "Escape from the Cartesian Theatre" (reply to commentators), Behavioral and Brain Sciences, 15, pp.234-47.

 Humphrey, N., 1992, A History of the Mind, London: Chatto & Windus; New York: Simon & Schuster.

 James, W., 1890, The Principles of Psychology, Cambridge, MA: Harvard University Press (1983 edition).

 Jaynes, J., 1976, The Origins of Consciousness in the Breakdown of the Bicameral Mind, Boston: Houghton Mifflin.

 Mangan, B., forthcoming, "Dennett, Consciousness, and the Sorrows of Functionalism," in Consciousness and Cognition.

 Penrose, R., 1989, The Emperor's New Mind: Concerning Computers, Minds, and the Laws of Physics, Oxford, Oxford Univ. Press.

 Purvis, Dale, 1992, "Consciousness Redux," Science, 257, pp.1291-2.

 Stoerig, P., and Cowey, A., 1992, "Wavelength Discrimination in Blindsight," Brain, 115, pp.425-44.

 Waxman, S. G., and Geschwind, N., 1975, "The interictal behavior syndrome of temporal lobe epilepsy," Archives of General Psychiatry, 32, pp.1580-86.