News and views about philosophy, the academic profession, academic freedom, intellectual culture, and other topics. The world’s most popular philosophy blog, since 2003.
People should understand what precipitated my remarks here at Leiter Reports. Also, a few replies and comments on the conversation:…
As far as I am concerned, I am open to being convinced that AI is good for the profession. However,…
Kaufman seems to imply that proponents of the “important truths” view must maintain that the truths AI might help us…
Wallace’s distinction between the value of the philosophical process and the value of its product suggests a further complication, one…
It seems to me—no philosopher that I am—that, having reread Prof. Kaufman’s essay above, one might reasonably ask what, if…
I’m not sure whether an interactive, online logic textbook counts as an open-access philosophy book, but for those who might…
For a slightly more positive assessment, JS Mill: “The real advantage which truth has, consists in this, that when an…
An enjoyable and provocative read, but is it fair to the consciousness "deniers"? What say you knowledgeable readers?
Would love to see a response by Alex Rosenberg ("There is no first-person point of view"):
http://www.alexrose46.com/posts/2016/9/13/why-you-dont-know-your-own-mind
I'm not knowledgeable enough of a reader to comment productively, but I do have a question. Strawson writes near the beginning, "The Denial began in the twentieth century and continues today in a few pockets of philosophy and psychology and, now, information technology." The rest of the article presents examples from philosophy and psychology (and physics), but not information technology. Evidently, it is a reference to recent discussions of AI, machine learning, and robotics. But why would an electrical engineer or computer scientist need to bother at all with the question of the existence of human consciousness? Are there some AI experts who imagine introducing consciousness into machines? If so, do their silly ambitions prompt other AI experts to deny consciousness altogether?
Well, I'll claim to be less knowledgeable than you and anxiously await a discussion among philosophers re. Strawson's really interesting piece. But it seems that a reasonable answer to your question is that the kind of "nothing but"-ism described by Strawson in his article is somewhat rampant in large sectors of the AI community, with "nothing but" corresponding to some hardware and the algorithm du-jour (e.g. "deep learning" in multi-layered perceptrons). And in these cases the errors in terms of those described by Strawson would seem even more egregious than the garden variety materialist naturalist who claims consciousness can't be material based on current understanding of material reality, re. restricting the understanding of material reality in a rather extreme sense.
No, it's so ridiculously unfair that it's either in bad faith or displays a near-total failure to actually read Ryle, or Wittgenstein, or Dennett, or… with any charity, or attempt to understand the arguments being made. But (a) the article is a terrible starting point for an actual positive account of functionalist, or behaviorist, accounts of consciousness, and (b) there are others far better qualified to give that account.
In lieu of something longer, though, a quick observation: Strawson states that physical science can describe only the structure of matter, but cannot describe its nature, and so can't explain how matter can be conscious. Without further explication of the structure/nature divide, that argument works equally well to establish that physical science is in principle unable to explain how matter can be pizza. Strawson gives no such argument that I can see in this article, only a table-thumping appeal to intuition.
Does David Wallace mean to say that there is a thing called pizza that is not a collection of our experiences of pizza and it is also not obviously a clump of matter? And we can ask whether the real pizza is wholly material? I imagine experiences of pizza. I imagine an obviously material swarm of corpuscles (or their moder day equivalent) that can cause me to have experiences of garlicy sauce and soft bread and warm cheese. But the real pizza I am asked to imagine is none of these things, thought it may turn out to be wholly explained in material terms. What is this real pizza?
I believe this is some sort of pizza Platonism that is wholly unjustified. 🙂
The "nothing but"-ism to which the AI community subscribes is correct, because computation *is* nothing but material. Consequently, I don't get why the AI community challenge (per Strawson) to the existence of consciousness poses a threat. To me, it seems beside the point.
It can't be sillier than his own panpsychism.
A point on machine consciousness: whether consciousness is ontologically distinct from the physical and whether implementing the same computational stuff as goes on in the human brain in a computer would get you a conscious computer are in fact separate questions.
Even (most, especially nonreligious) dualists or panpsychists don't doubt that conscious experiences *depend* on the brain; the question is just about the nature of the dependence (i.e. is it causal or constitutive (dualists say causal, and so I guess do some panpsychists); if it's constitutive, do the most basic facts at the bottom of the chain of constitution themselves involve conscious (some panpsychists say they do.) And the accounts they give of the rules taking you from particular brain states to the particular conscious experiences that depend on them (whether causally or constitutively) sometimes entail that any functional duplicate of the brain will in fact be conscious, regardless of the material it's constructed out of. For example, dualists sometimes think there are primitive laws of nature that take you from the information contained in a system to the instantiation by that system of conscious experience type properties ('phenomenal properties' under one use of that famously confusing piece of jargon.) And 'information' is usually defined causally, rather than in terms of intrinsic properties of the information carrier, so that at least a perfect silicon duplicate of a normal subject a normal environment will have the same experiences if it gets the same inputs from it's sensors. David Chalmers, by far the most distinguished contemporary dualist has defended the view that an AI that was a functional duplicate of a human being would almost certainly be conscious.
Given this, just because an AI scientist is confident that making a computer run the right programme will bring about consciousness, they're still not necessarily committed to any particular view on the metaphysics of consciousness, still less to the views Strawson dislikes.
And a second point on 'qualia': If qualia are meant to be intrinsic properties of experiences, then lots of people who would never dream of seeing themselves as consciousness-deniers deny they exist. For example, sense-datum theorists who identify qualia with properties of sense-data, so that they 'feel of seeing red' really picks out a property of the surfaces of sense-data are 'consciousness-deniers' on that definition of qualia, which seems weird. (Strawson clearly thinks of Broad and Russell as allies, so maybe he doesn't mean to say that anyone who denies the existence of qualia conceived of a properties of experiences themselves is a consciousness-denier.) Moving to more sensible views than sense-datum theory, lots of contemporary work in the philosophy of perception claims that 'qualia' are really just properties of the perceived objects, and not properties of experiences; so that what you identify when you focus on the 'feeling of pain', is actually just damage to the body for example. (In cases of hallucination, on some of these views, qualia are 'merely intensional objects', just like the fountain of youth can be the intensional object of my thought without actually existing.) The motivation for these views being the famous 'transparency of experience', the observation that when you try and pay attention to the 'feelings themselves' what you actually pay attention to seems to be spatially located and external. (On this, try Gilbert Harman's The Intrinsic Quality of Experience, amongst others.)
A very cool and startling paper on the second point is Alex Byrne's Color and the Mind Body Problem, which argues that the traditional mind-body problem should be replaced with a problem about how physical surfaces, and other perceptible external things, manage to instantiate the qualities that we immediately perceive:http://web.mit.edu/abyrne/www/color_and_mind-body_prob.pdf
Are there really any consciousness-deniers? Perhaps there are, but it's not clear to me that Wittgenstein, Ryle, and Dennett count as such. (Strawson mentions W only obliquely and R not at all. Dennett is his stalking horse, it would appear, and clearly Dennett does not deny consciousness, even in his heterophenomenological tracts.) It's true that they all deny the central psychological importance and explanatory priority of raw feels, first person perspectives, what-it-is-likes, etc. They might even deny that these things are stable enough to be scientifically investigated. (Doesn't Nagel play into that sentiment with his view from nowhere?) But this does not merit the award of The Great Silliness, does it? Strawson writes: 'to seem to feel pain is to be in pain.' Why is it SILLY to deny this—or at least to say that it is at best analytic, and hence superficial?
This is similar in character to G. Strawson's book from a few years back. The same problem arises: it is an absolute caricature of what Dennett actually believes. Dennett doesn't *deny* consciousness – rather he says, IMO correctly, that when one investigates it, one discovers it really isn't at all like what one thought. Strawson, by contrast, seems to think that by continuing on with "this is the most certain thing in the world" card is doing exactly what Dennett says leads nowhere. Dennett encourages one to "bracket" this (sort of like Husserl) judgement and attempt to find other approaches.
Also, Dennett, like the Churchlands and others accused of "denying the obvious", realizes there are other constraints on what we say about such problems that should be taken seriously – like conservation laws. We now know the amount of "slack" in our knowledge such things, and it is extremely small – far too small to have anything other than matter in the broad sense involved.
I too see nothing about "information technology", which these days is sort of my field (though artificial intelligence is not). This just seems to be an aside – which is a shame because there are very profoundly important problems philosophers and others should discuss even with "multilayer perceptrons".
A denial of a clear fact can be sillier than a strange explanation for that fact .
(But I’m not a panpsychist)
You can acknowledge what is clearly a fact, but deny that anybody knows what it is, precisely—deny that it is clear in this sense. That's not silly; it can actually be profound. Dennett is somewhere in this territory; not in denial territory. (See comments 9 and 10.) It's probably a bit silly to go overboard and deny the fact, but hardly Greatly Silly.
It’s fair to point out that I have read this line of argument from GS quite at least twice previously. But I think that about the almost unbelievable absurdity of eliminativism he’s basically right. Note that this doesn’t mean that anyone actually holds this position, just that it would be very foolish.
I think lots of ‘common sense’ views about consciousness are wrong (as does GS, if you read him on the self) but the existence of our conscious experience is as certain as GS says.
I’m less convinced by the intrinsic character of matter argument but I think that it’s an interesting sketch to be filled in (it is no prima facie sillier and has no less evidence for it than lots of things that physicists say about physics! e.g. much of string/M theory). Panpsychism in Strawson’s sense I think is a perfectly valid position – no more daft than plenty of other views philosophers hold and not in the same stupidity league as eliminitavism (again, if anyone holds this).
I’ll get his book when it’s out; I’ve enjoyed his others.
No, the "nothing-but" refers to consciousness (i.e. "nothing-but" computation); assuming that computation *is* nothing but material is certainly reasonable enough. If the inference is correct, i.e. consciousness being nothing but computation and computation being nothing but material then consciousness is nothing but material, in my feeble understanding, Strawson doesn't necessarily disagree with that, but he feels that evidence for it is far from established by current physical theories (and is currently only accomplished by somehow denying the existence of subjective experience).
The Denialists are rarely as denialist as their critics claim. Even in "Quining Qualia," Dennett doesn't deny that we have conscious experiences. He denies that there are qualia, that qualia are somehow indispensable for thinking about consciousness.
Critics will say that this comes to the same thing. They'll say that if you deny that there are qualia, you ipso facto deny something we know more intimately than anything else. But is that true? "Qualia" hardly belongs to the commonsense picture of conscious experiences. We might say, "I know what it's like to taste Marmite," but that doesn't obviously commit us to philosopher's English, like, "I'm acquainted with the qualia that characterize the tasting of Marmite." Qualia are theoretical posits, and as such, they have to earn their keep to be useful for the theory. And in "Quining Qualia," Dennett does (in my opinion) a good job of showing that they *don't* earn their keep, that they're *not* useful, because the concept doesn't hold up under scrutiny, and can't do its job.
Is denying that there are qualia the same as denying the existence of experiences? You might think that the right response to a qualia quiner is, "If you have to ask, you'll never know," but that's conflating qualia — the theorist's posit — with experience, which is what most of us take to know intimately. Imagine that someone is skeptical of aggregative principles in consequentialism; they might say, "What are these hedons, and does the very idea of adding them even make sense?" It would be pretty churlish for a consequentialist to respond, "What, you deny the existence of pleasure?" No — it's just skepticism about a theoretical entity.
That is sometimes true of Dennett. Depends on the day or the passage we’re looking at. He’s brought up a number of things that could complicate our views of certain conscious phenomena such as Orwellian vs Stalinist revisions, different interpretations of the change of the coffee tasters tastes, and so on. None of that is silly and is worthy of some consideration. But when he claims a good materialist must deny that the afterimages I see are in fact colored or deems the painfulness of pain is sheer illusion— well, I have to part ways with him.
Thanks. We seem to be on the same page, more or less. But do you really want to say that afterimages are things that you see, or that these things—the afterimages themselves—are in fact coloured? Anyway, that's what Dennett rejects, and I'm with him on that. I'd be surprised if Galen Strawson wanted to say anything else.
I see, thank you. My apologies for dwelling on the IT aspect to the detriment of attention to philosophical issues (e.g., qualia, the Stanford entry on which confuses me, due in no part to it). To return then to your comment (3), what is rampant in large sectors of the AI community is a belief that progress in computation will eventually arrive at reproduction of consciousness. Contrast pizza making, where no amount of progress will persuade anybody that the result is consciousness or even consciousness-like (excepting poetic instances of likeness, for which anything is like anything else). The crux may be the force of the word "reproduction" as I'm using it. When we say animals reproduce, we mean that the result is more animals. When we say an artist reproduces a painting, we mean that the result resembles the original, even if we also acknowledge that the result is, trivially, another painting. AI scientists strive to make machines that perfectly resemble consciousness, but either confuse the result with regeneration of consciousness or dismiss the difference as negligible. It's as if I deem this exact reproduction of a painting by Raphael as good as, and therefore essentially identical to, the original. But now I'm begging the question, not to mention heading down a path of no interest to Strawson in the NYRB piece. For AI, perfect resemblance to consciousness might very well be all there is to consciousness, because there's no original to hold up to it in protest.
" But do you really want to say that afterimages are things that you see, or that these things—the afterimages themselves—are in fact coloured?"
No, and I'll raise you by suggesting that neither is anything else, at least in the sense you're using the word.
I think that counter-intuitive assertion is implicit in Chirimuuta's theory as described in Outside Color when tweaked with a little neurology. (She may not agree – I hope to find out soon.)
Re GS's article, the adage about glass houses and stones seems apropos.
Good. But I would have to say my after-images are colored— some red, some green, some purple. I guess I’m not a good materialist.
Some mitigation:
If you don’t like the word ‘see’, perhaps taking it as factive and world-directed, then I visually experience them. We don’t have to think about after-images as “things” either. I don’t suppose after-images are mental particulars instantiating phenomenal properties. More likely they are best construed as collections of phenomenal properties being instantiated by the subject or the subject’s visual field.
Comment #20 was supposed to be posted as a reply to Mohan. My bad.
If your intuition runs in the direction of the defence of the concept of qualia, and you'd
like some decent arguments in support, David Chalmers should be your go-to. I'm afraid that
Galen Strawson merely offers a pugnacious "How dare you deny what I know!" schtick that is
quite embarrassing intellectually. I heard him perform this at the UNC Chapel Hill Philosophy
Colloquium a few years back and it left a bad taste in my mouth. He's just accusing those who
disagree with him of bad faith.
We’ll certainly AI people have the right to claim that the endpoint of their efforts will be consciousness, sort of like one of the “regulative assumptions” of the pragmatists, i.e. assuming the possibility of success seems a reasonable assumption for engaging in any goal oriented activity. But that doesn’t mean that anyone else has to regard that as a reasonable assumption particularly if what it currently entails seems unreasonable.
I think that Strawson is right that Dennett (and perhaps Ryle and some others, but not Wittgenstein) denies the reality of consciousness. The point is not that he rejects the notion of qualia, which surely is questionable, but that what – if anything – he puts in its place is as transparently a mere functionalist ersatz as, say, a logical construction out of sense data is for a pizza. What I found striking about Strawson's piece is that neither he nor most of his critics seem even to consider the idea that both consciousness and the pizza might have a physical microstructure without being identical to it. As long as such common sense realism is ruled out from the start, I suppose the debate will continue to go back and forth between the silliness of consciousness denial on the one hand and that of panpsychism and other forms of consciousness fundamentalism on the other.
Sorry for the late response; I hadn't been monitoring this thread over the weekend. I think Dennett would agree that you have visual experience as of a coloured after-image, but he doesn't want to reify the thing you see. You agree that after-images aren't mental particulars, and that's half way there, but you reify the visual field, and he wouldn't (I think). Still, you get the idea.
I don't think that the dispute between Chirimuuta and colour-realists is relevant to the issue about eliminativism.
—–
KEYWORDS:
Primary Blog
Leave a Reply