About Rationally Speaking


Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

Tuesday, March 26, 2013

I just don’t get it. One more on Mary and the Zombies


by Massimo Pigliucci

Can science explain consciousness?, asks a recent New York Times column by philosopher Gary Gutting. Well, it depends on what one means by “explain,” “consciousness,” and “can.” No, I’m not trying to play semantic games, but I keep being frustrated by philosophers who think they have discovered something profound, while in fact they are just playing with words. I don’t get it, and it doesn’t help my life long quest to bridge the gap between science and philosophy, since it makes it far too easy for scientists to mock and dismiss the latter.

Gutting’s column is about two classic thought experiments in the philosophy of mind: one concerns Mary the neuroscientist who grows up without color perception and then suddenly experiences color; the other regards the possibility of beings, “Zombies,” who look and talk like you and me and yet have no consciousness, nobody’s “up there” in their brains, so to speak. The first thought experiment is due to Frank Jackson, the second to David Chalmers.

Both situations have been discussed at Rationally Speaking before (see links above), and Gutting summarizes the basics of each in his post too. What I am concerned with here is what we are supposed to take from these philosophical exercises, particularly when it comes to the question of what science can and cannot (potentially, if not now) tell us about consciousness.

Here is what Gutting suggests we should learn from Jackson’s and Chalmers’ speculations, respectively:

At that moment, Mary for the first time experiences the color red and now knows what red looks like. Her experience, it seems clear, has taught her a fact about color that she did not know before. But before this she knew all the physical facts about color. Therefore, there is a fact about color that is not physical. Physical science cannot express all the facts about color.

If a zombie-twin is logically possible, it follows that my experiences involve something beyond my physical makeup. For my zombie-twin shares my entire physical makeup, but does not share my experiences. This, however, means that physical science cannot express all the facts about my experiences.

Let’s unpack both conclusions carefully, starting with Mary. The (apparent) force of Jackson’s hypothetical situation relies on an equivocation about the word “fact.” Mary had studied all the scientific facts about how the brain works, but those facts do not equate with having the first person experience of actually seeing color, just like I can study all the facts involved in, say, learning how to swim, and yet feel surprised by my body’s reactions the first time I actually touch the water and begin to swim. Mary’s experience is not a “fact” in the same sense as a scientific understanding of sensorial experience is made of “facts” [1]. What is happening to Mary is that her understanding of what it means to see color is being augmented: before she began perceiving color, her understanding was limited to scientific facts; afterwards it is being complemented by the experience of actually seeing color. To say that science cannot “explain” the latter is, again, to equivocate on what it means to “explain.” Scientists actually have a pretty decent, if partial, explanation of color perception, one that details all the neural components that have to be in place for such experience to take place. But when we switch to the first-person feeling of perception we consider something that is, by its very nature, not a matter of scientific explanation at all. In other words, to charge science with failure of “explaining” first person experience, when one really means the feeling resulting from, not the mechanism that allows, such experiences, it so make a category mistake.

What Jackson’s thought experiment shows is not that science fails to explain consciousness, but rather that there are certain phenomena that are simply outside the domain of science (I know, I know, in some circles such a possibility is considered heresy!), such as first person experiences. To confuse the two would be like blaming the New York Yankees for never having won an NBA tournament: of course not, they are a baseball team, not a basketball team.

Now to Chalmers’ zombies: contra Gutting’s summary above, nothing concerning the physical world follows at all from considering my the possibility of my zombie-twin. That is because at best Chalmers has shown that there is no logical contradiction between talking and interacting like a human while at the same time not having conscious experiences. But logical constraints are too weak to tell us much about physical constraints, and even the latter only set boundaries within which biological phenomena evolve in part as the result of contingent events. What Chalmers (and Gutting, and several others) don’t seem to understand is the simple proposition that the problem of consciousness is a biological, not a logical one. In a sense, this too is an issue of category mistake: confusing logical with physical and contingent possibilities. In this case too it isn’t that philosophy shows that science cannot do something, but rather that the domain of science does not extend to that of logic, which is eminently philosophical. (Of course, scientific reasoning is still bounded by logic.) Yet another case of blaming the Yankees for something they couldn’t possibly do...

None of the above should be construed as a dismissal of thought experiments (which, incidentally, have a place also in science), but rather as a clarification of what thought experiments are about. Hypothetical scenarios simply do not have the power to discover new facts about the world. If that were the case, then we wouldn’t need empirical science, or at the very least the role of empirical science would be greatly diminished in comparison with what it actually is (I’m looking at you, string theorists...). Rather, thought experiments are helpful to bring to the fore our assumptions about certain matters, as well as to unpack the logical entailments of those assumptions. When well done, they represent good philosophy (or good science, depending on the experiment). But when one uses them beyond their scope to reach grand conclusions about what practitioners of another field can and cannot do, one ends up just looking silly.

————

[1] Just like, say, a mathematical “object” is not the same kind of thing as a physical object like my laptop computer. And that’s quite irrespective of one’s opinions on the ontology of mathematical objects.

110 comments:

  1. Professor Pigliucci,
    What you are saying is that prominent philosophers, who are professors in prestigious universities and published famous papers, are making basic mistakes such as category mistakes and misunderstanding how thought experiments work. You are essentially saying that they are incompetent. How is that possible? You'd think that 1. They know these mistakes when they see one (don't they have to teach logic 101?), 2. Somebody has pointed out the mistakes to them before, and 3. Don't their books and papers have to be reviewed? How can the reviewers not pick up the mistakes?

    The reason why I wrote the above is that I agree with you, not because I don't agree with you. It appears to me that the arguments are clearly flawed. But if that is true, why are they taken seriously?

    ReplyDelete
    Replies
    1. Another favourite of mine is The Chinese Room. Why John Searle is taken seriously I honestly can't understand. It's a terrible argument and it misses the point completely.

      Delete
    2. DM, Is this because you agree that Searle mistakenly treated the brain as the bearer of intelligence instead of the brain being intelligence's causal basis?

      Delete
    3. @Baron P

      Something like that, yes. I would subscribe to the "virtual mind" counter-argument.

      Delete
    4. 1. Biases and motivation should not be seen as evil villains. Prof. Pigliucci is more-or-less rebuking professors who are, first and foremost, human. The of this piece is also human (as are you and I for that matter). We should not assume a capability of recognizing these mistakes implies they *will* recognize them. For the same reason, we also should not assume that since the objections can be presented plainly and you and I can see them clearly, they are by any means obvious or correct. These issues are complicated and the categories (for which there have been categorical errors) are ill-defined throughout the history of philosophy.

      2. People may have also praised them for their observations before, perhaps negating the impact of the criticisms. Even if the objections have been presented in no uncertain terms, it possible their opinions would not change, especially with such a long heritage of metaphysical observation on *their* side.

      3. I have no details or insights for 3. I fear that philosophical review is a little broken, because in the absence of directly observable evidence, paradigm may sometimes function like opinion (and vice versa).

      Delete
    5. DM, re either the systems or virtual mind reply, I think in both cases there'd have to be an assumption on your part that the brain causes intelligence. But there are scientists who would argue that the brain is a system that's been evolved to make much better strategic use of the intelligence that life forms were either "born" with or somehow stochastically acquired. In other words we use it but we didn't cause it.

      Delete
    6. I agree that the Chinese Room is a terrible thought experiment. But I don't agree that the Mary or Zombie thought experiments are terrible. If they have errors, the errors are much subtler and more interesting, and no one on this site has yet succeeded in clearly demonstrating them.

      I feel it important to add that the ad hominems on display here are largely uninformed and anti-intellectual. Because we find a certain philosophical argument obnoxiously difficult to dissolve, we start tarring the original innovators of those arguments as religious zealots and dogmatists? The irony is almost painful, in light of the fact that it is we who are expressing an unargued faith in the falsehood of the conclusion while largely refusing to argue the point, and Chalmers and Jackson who are actually trying to reason their way to the right conclusion. Sure, their reasoning may be flawed, but until I start seeing less rhetoric and more actual counter-arguments here, I won't be impressed by the knee-jerk dismissals.

      As it happens, Jackson HAS changed his mind. Years ago. The guy's a physicalist now. The reason we're still talking about his argument is that even though Jackson has joined the ranks of the many physicalists who are convinced there must be something unsound about it, even he has a hard time diagnosing and articulating exactly where the problem is coming from. Our challenge is not merely to assert that we're so confident of physicalism that there has to be something or other wrong with the argument; rather, our challenge is to figure out what precisely that something is, and to defend our diagnosis.

      A lot of you treat Massimo's response as though it is both decisive and childishly obvious, treating it as an opportunity to bemoan the foolishness of philosophers, or of humanity in general, for not seeing such an obvious problem in the argument. But, um, 'Science isn't in the business of explaining first-order facts' certainly isn't obviously true (in fact, I'd lean toward saying it's false). And it certainly isn't obviously adequate as a defense of physicalism (on the face of it, a dualism of first- and third-order facts is tantamount to dualism, not to physicalism). At the very least, more work needs to be done. I don't mean to suggest that we should despair of dissolving the problems; but I do want us to start actually putting in the hard work of doing so, rather than just patting ourselves on the back for being smart enough to not accept any arguments made by a Rival Tribe.

      Delete
    7. @Robby, "If they have errors, the errors are much subtler and more interesting, and no one on this site has yet succeeded in clearly demonstrating them."
      IF they have errors, and you haven't found them either, how would you know that if the errors are there, they're much subtler? Wouldn't that be a rather faulty tautological argument? Like saying that any errors are apparently too subtle for us to find so far, therefor if there are any, they are too subtle for us to find.


      Delete
    8. @Robby

      >I feel it important to add that the ad hominems on display here are largely uninformed and anti-intellectual.<

      Not sure if this is aimed at me, but for what it's worth I do have a high opinion of Chalmers in particular, and Jackson seems pretty reasonable from what I can see. Both of these thought experiments are useful in examining intuitions and beliefs, and I did not mean to denigrate them in any way, short of disagreeing with me.

      John Searle, on the other hand, drives me crazy, because he believes and frequently states that he has found an iron-clad knock-down argument against Strong AI, and he plainly hasn't. It might be easier to endure if I could argue with him face to face, but when I read him or hear him making this pronouncements it drives me up the wall.

      >A lot of you treat Massimo's response as though it is both decisive and childishly obvious, treating it as an opportunity to bemoan the foolishness of philosophers, or of humanity in general,<

      This would not be my attitude, but I do think there is an interesting point to be made here with regard to Massimo's attitude to philosophy.

      Massimo, as a professional philosopher, is often quick to point out how useful philosophers are at expertly dissecting conceptual problems and how poor scientists and laypeople such as Michael Shermer often are at considering philosophical questions.

      This is slightly irritating to laypeople such as myself, because it leads me to suspect that Massimo would take my opinions and arguments more seriously if I had a PhD.

      It is therefore amusing to consider this in the light of his mockery of the arguments put forth by his fellow professional philosophers. If professional philosophers can be so wrong, or can disagree so vehemently, then perhaps having a PhD in philosophy does not in fact make you much better equipped to think clearly. This is a gratifying view to hold if you don't have a PhD in philosophy and have no likelihood of ever getting one (due to work and family commitments).

      Delete
    9. I agree that the right explanation might turn out to be surprisingly simple. But I don't think this would qualify the explanation as 'obvious'; obvious explanations don't require years of careful study and debate to figure out and justify, even if in retrospect they can be formulated with great concision.

      Delete
    10. I am confused by what is going on here. Smart people can disagree with each other. Massimo himself said it many times. But in this case Massimo is asserting that other philosophers are deeply mistaken about the logical consequences of thought experiments. In other words, philosophers like Chalmers are illogical and they simply can't see how illogical they are. How can serious professionals be so wrong about the tools that they use? It's like saying a respected mathematician does not understand induction. He or she continues to make mistakes when he uses induction, does not acknowledge errors in his proofs, and teaches students to use induction wrong. That really seems hard to believe to me.

      Although what Massimo wrote seems to make lots of sense to me, I am compelled to conclude that Massimo cannot possibly right. He must have misunderstood the arguments made by other philosophers. That is uncomfortable to me too. You'd think philosophers understand each others arguments clearly.

      It appears to me that the only way to resolve this is that philosophy as a whole does not agree what conclusions thought experiments can logically draw. That makes things even worse because philosophy supposedly knows how to use logic. I am really confused.

      Massimo, if you get a chance, can you try to explain to me how prominent philosophers (Chalmers, a professor at ANU, which is not too shabby) can possibly be so wrong and how he manages to publish papers and books containing, as you said, basic logic mistakes. If that is true, in your opinion, should his papers be retracted?

      I understand that people in every field can become prominent for reasons unrelated to the quality of their work, but what you are saying is that Chalmers is making simple, basic mistakes can that be easily refuted, and yet his arguments are still being taken seriously. This is unimaginable in any field, in science or in humanities, that I know.

      Delete
    11. I apologize if my other comments unduly distracted from the substance here. Ultimately these ad-hominem questions aren't what we should be talking about at this stage. To even begin to evaluate the competence of philosophy as a whole, or of a whole subdiscipline of philosophy (like phil. of mind or phil. of science), we'd first need to be extremely well-acquainted with a lot of the specific philosophers in those fields. At this point most of the posters here aren't even well-acquainted with Chalmers and Jackson's zombie and Mary's Room arguments (e.g., they don't understand which specific fact Jackson thinks Mary learns after leaving the colorless room, and they don't understand which kind of possibility Chalmers thinks conceivability can tell us about), much less with their oeuvres as a whole. So let's focus more on unpacking the specifics of the arguments themselves, and where they and Massimo disagree. Very little technical knowledge is required for this discussion, just a willingness to read a few papers written by Chalmers, Jackson, and their interlocuters; so it will be more productive to discuss the substance than to fret about levels of expertise.

      I take it Massimo accuses Jackson of "equivocation" and a "category mistake" because Massimo thinks there is a basic category distinction between first- and third-person facts, and that either our actual epistemic standards do, or our ideal epistemic standards ought, focus only on predicting, explaining, and physically describing the latter.

      But Jackson, of course, doesn't accept such a duality as a presupposition; indeed, he would probably respond that to assume an irreducible duality of facts like this is to assume dualism itself, which Jackson (as a former a-posteriori epiphenomenal dualist, current a-posteriori physicalistic monist) wants to resist. Jackson is only guilty of "equivocation" or confusing the two categories if we presuppose that he accepts Massimo's picture of proper epistemic methodology, which artificially restricts the legitimate ways one can inquire into the nature or basis of first-person facts. But we'll at least want some sort of positive argument for this artificial restriction, if we're to even be able to assess the success of Massimo's response.

      Indeed, it almost sounds like Massimo is not defending physicalism so much as saying it's not scientists' *fault* that physicalism is so hard to justify: "What Jackson’s thought experiment shows is not that science fails to explain consciousness, but rather that there are certain phenomena that are simply outside the domain of science[....] To confuse the two would be like blaming the New York Yankees for never having won an NBA tournament: of course not, they are a baseball team, not a basketball team." But the point of Jackson's argument isn't *who's to blame* for the evidence that seems to favor dualism. The point of Jackson's argument is that the prima facie evidence for dualism seems hard to dissolve or explain away.

      So this response to Jackon will need some more unpacking and defending. (Perhaps Massimo has done so elsewhere! I'm only evaluating the presentation here.) His objection to Chalmers' argument, on the other hand, is simply off-base. Nowhere does Chalmers confuse or conflate logical and nomic possibility. Indeed, Chalmers thinks zombies are nomically (i.e., physically, biologically...) impossible, which would be incoherent if he conflated nomic and logical possibility, given that he thinks zombies are logically possible. He's talking here only about logical possibility, and the brand of physicalism he seeks to refute is simply the thesis that our world's facts logically supervene on its microphysical facts. As far as I can tell, Massimo might very well agree with Chalmers that (logical-supervenience) physicalism is false; perhaps their only disagreement is about whether this is an important or interesting result? It's hard to tell!

      Delete
    12. Robby Bensinger, thank you for the reply. Could you recommend a minimal reading material that would give me a basic but accurate picture of the Mary argument and the zombie argument? I have a feeling that the two arguments must be presented inaccurately by Massimo and by Gutting and I'd like to find out. But then again, if professional philosophers like Massimo and Gutting can't get the subtle points right, what hope do I have?

      Delete
    13. Massimo's and Gutting's presentation of the arguments isn't particularly unusual or inaccurate, but they are imprecise. In the case of the Zombie Argument in particular, that ends up causing a lot of confusion about what kinds of possibility and conceivability are being invoked.

      Both of these arguments have been heavily discussed in philosophy of mind for over 30 years now, so we can fruitfully avoid some blind alleys by noting distinctions past writers have drawn and responses they've made. Regarding the Knowledge Argument:

      - Nida-RĂĽmelin 2002/2009, "Qualia: The Knowledge Argument". http://plato.stanford.edu/entries/qualia-knowledge/
      --- If you're going to read just one article, make it this one.

      - Nagel 1974, "What Is It Like to Be a Bat?" http://cutonthebiasworkshop.files.wordpress.com/2011/05/nagel-1974-what-is-it-like-to-be-a-bat.pdf
      --- The most important precursor to Jackson's Knowledge Argument.

      - Jackson 1982, "Epiphenomenal Qualia". http://www.sfu.ca/~jillmc/JacksonfromJStore.pdf
      --- The original presentation of the 'Mary's Room' scenario comes from p. 130.

      - Jackson 1986, "What Mary Didn't Know". http://home.sandiego.edu/~baber/analytic/Jackson.pdf
      --- Responds to a lot of common misconceptions about the argument. Short.

      Lewis 1990/2004, "What Experience Teaches". http://faculty.washington.edu/bonjour/PHIL463/Lewis.pdf
      --- Responds to a lot of common misconceptions about the argument. Argues that Mary doesn't acquire new propositional knowledge, though she does acquire new abilities, new know-how.

      Conee 1993. "Phenomenal knowledge". http://www.tandfonline.com/doi/abs/10.1080/00048409412345971
      --- Argues that Mary doesn't acquire new propositional knowledge, though she does acquire a special new kind of acquaintance.

      Dennett 1991. Consciousness Explained.
      --- Claims around p. 400 that Mary would already know everything about color phenomenology in advance. Explained at greater length in Dennett 2006, "What RoboMary Knows": http://ase.tufts.edu/cogstud/papers/RoboMaryfinal.htm

      - Byne 2004/2010. "Inverted Qualia". http://plato.stanford.edu/entries/qualia-inverted/#LocInvSpeSce
      --- Suggests in §§2.2, 2.3.1 that Inverted Qualia arguments made to date are very problematic. (This lends some credibility to Dennett's 'blue banana' scenario and potentially undermines variants like Marianna's Room.)

      Robinson 1993/2004. "Dennett on the Knowledge Argument". http://books.google.com/books?hl=en&lr=&id=HYyubHBrVUcC&oi=fnd&pg=PA69#v=onepage&q&f=false
      --- Short.

      Delete
    14. sciation - I think your concern is a healthy one. There are a couple of aspects to this issue that I think make it prone to this kind of disagreement.

      The first is that the zombie thought experiment calls upon a strong intuition. From the Stanford Encyclopedia:

      "The intuitive appeal of the zombie idea can be overwhelming. Those who exploited it in the 1970s typically assumed without argument that zombies are not just conceivable but possible".

      So even if your reasoning and argumentation are perfect, you're starting with a strong intuition, and that can lead to wildly different viewpoints. I personally do not share the intuition, so I'm in immediate, *intuitive* disagreement with Chalmers.

      The second thing is that Pigliucci making an accusation of a "category mistake". This kind of mistake is actually extremely easy to make, due both to the way we conceptualize and the way we use language. And even articulating the exact nature of a particular category mistake can be very difficult. For example, if you spoke of an atomic force as being a cause and compared it directly to an economic force as a cause, you could be said to be committing a category mistake (depending on what you were arguing). But try to explain precisely what's different, and you'll find it's harder than it seems.

      Delete
    15. More Knowledge Argument material:

      Chalmers 2002. "Consciousness and its Place in Nature". http://consc.net/papers/nature.html
      --- Surveys views on phenomenal consciousness in general, including their positions on the Knowledge Argument (but also the Zombie Argument). The physicalist views I listed above are all Type-A or Type-C.

      Loar 1997. "Phenomenal States (Second Version)". http://www.nyu.edu/gsas/dept/philo/courses/consciousness97/papers/loar.html
      --- A defense of Type-B Materialism. Mary does acquire new knowledge when she leaves the colorless room, but only insofar as she acquires a redescription of old knowledge using importantly new concepts. The 'directness' of these new concepts' reference makes them psychologically autonomous.

      Stoljar 2005. "Physicalism and Phenomenal Concepts". http://web.gc.cuny.edu/cogsci/private/Stoljar-phenomenal-concepts.pdf
      --- Gives a more broad overview of the phenomenal concepts strategy, and argues that it isn't relevant to the Knowledge Argument after all.

      Chalmers 2003. "The Content and Epistemology of Phenomenal Belief". http://consc.net/papers/belief.html
      --- §2.1 responds specifically to phenomenal concepts theorists who seek to explain the special character of phenomenal concepts by indexical/demonstrative means.

      Chalmers 2006. "Phenomenal Concepts and the Explanatory Gap". http://consc.net/papers/pceg.pdf
      --- Argues that the Type-B theorist's phenomenal concepts strategy cannot succeed, because 'concept zombies' will produce the same problems as the original phenomenal zombies.

      Delete
    16. Robby & Asher, much appreciated.

      Delete
  2. The confusion seems to be about the difference between (1) a static description of color perception, i.e. an algorithm (2) a dynamic state sequence of color perception, i.e. computing the algorithm that is describing color perception using external computational hardware (3) using your own computational hardware to compute the algorithm of color perception.

    The difference between #1,2,3 seems to be that each level removes some uncertainty over logical beliefs about complicated algorithms.

    ReplyDelete
    Replies
    1. We have "computational hardware"?

      Delete
    2. What I mean is that thought experiments such as Mary's room indicate that Mary knows (1) all abstract facts about the brain which are related to color perception. Maybe she (2) even runs experiments where those abstract facts about color perceptions are being empirically observed in a brain other than her own brain. What she lacks is (3) the implementation of those facts, about color perception in neural networks, in her own neural network.

      Each of those steps provides additional knowledge about color perception. Which means that Mary really is lacking some physical information as long as she does not reach step #3.

      Delete
    3. Why is step #3 required for any physical knowledge? When I write a computer program, am I missing any physical facts about that program in virtue of not running the program on my own brain? Couldn't I gain complete physical knowledge about ordinary programs just by studying the algorithm (and its implementation in other computing systems), even if I didn't myself implement it?

      If not, then what sorts of true propositions would I be missing out on? If so, then how does the color-perception algorithm differ from ordinary algorithms?

      Delete
    4. Consider the difference between an agent who likes vanilla ice cream and an agent who likes chocolate ice cream. You can't learn all physical knowledge about what it means to be an agent who likes chocolate ice cream if you are an agent who likes vanilla ice cream as long as you do not alter your utility-function in such a way as to favor chocolate ice cream.

      The reason being that there will always remain some uncertainty about the behavior of agent's favoring chocolate ice cream if you do not turn yourself into an agent favoring chocolate ice cream and thereby become a perfect predictor of how such an agent would act. As long as you favor vanilla ice cream you will always remain an (slightly) imperfect predictor.

      Delete
    5. "You can't learn all physical knowledge about what it means to be an agent who likes chocolate ice cream if you are an agent who likes vanilla ice cream"

      I'll assume for the moment that 'liking chocolate' and 'liking vanilla' are mutually exclusive. Even if so... why not? This sounds like the ice cream version of motivation internalism, the doctrine that you can't understand moral claims unless you're motivated by them. In the context of ice cream, it sounds even less plausible. The person who hates ice cream and the person who loves ice cream might in principle have access to exactly the same voluminous physical information; I see nothing in principle that could permanently block some propositional knowledge about the physical nature of ice cream from either one.

      "The reason being that there will always remain some uncertainty about the behavior of agent's favoring chocolate ice cream if you do not turn yourself into an agent favoring chocolate ice cream" - So your diagnosis of the Mary thought experiment is that Mary doesn't learn a new piece of knowledge, 'that's how it feels for people to experience redness'; she just acquires more confidence in something she already knew. Is that right? Because this just doesn't seem to be accurate, and it certainly doesn't dissolve the original intuition. Intuitively, Mary wouldn't even be able to entertain thoughts about redness before she first experiences it; the issue is not that she is uncertain about something, it's that she is totally clueless about the property in question's even being a property at all. And this cluelessness is startling if we think that the physical facts logically entail all our world's facts.

      The only information I can think of that could possibly be barred to the non-vanilla-lover is indexical information like 'I know I'm the sort of person who likes vanilla.' And, sure, indexical information like that is indeed barred to certain people in virtue of their preferences. But only trivially so, since the facts in question aren't facts at all with respect to this agent. You can't know something that isn't so. This sounds like so much wordplay, like arguing against God's omniscience on the grounds that Mark Twain knows something God doesn't -- namely, he knows 'I am Mark Twain', whereas God knows the distinct fact 'He is Mark Twain'.

      In the original thought experiment, the relevant information Mary learns isn't indexical information about her own mental states, and it isn't a new form of confidence or justification in something she already knew (cf. Jim Pryor's response to the thought experiment); it's general psychological information about everyone else's mental states. Indexical information seems trivial to us, whereas novel phenomenal information like 'leaning what it's like to experience redness' seems substantial and significant. Why should that be?

      Delete
  3. A word: The Eastern Masters of old believed no word could explain the truth, whereis I see the truth can be spoken it simply cannot be heard.

    And as for science explaining anything, they have yet to hear or see!

    =

    ReplyDelete
  4. Hi Massimo,

    I agree that the arguments don't establish dualism, but I think we need to do more work to get there than you suggest.

    It seems that a big part of difference between you and the dualists (Chalmers and old Jackson) is that they think that all of biology and psychology can be ontologically reduced down to the microphysics. If one is a pluralist (as, I gather, you more or less are), then one will respond to the failure to reduce consciousness to physics with a shrug: very little can be reduced to physics, you will say.

    You say, "But when we switch to the first-person feeling of perception we consider something that is, by its very nature, not a matter of scientific explanation at all."

    I think old Jackson will respond to this -- with justification -- by saying that this amounts to a denial of physicalism. You're agreeing with him. Of course, the notion of explanation is highly epistemic, which is why Chalmers puts more weight on the modal argument, which is more directly ontological.

    Chalmers presents several arguments, which I think should be taken seriously, for the claim that ontological reduction requires logical supervenience and that in nearly all cases except consciousness we do get logical supervenience on the microphysics. He argues that all biological facts logically supervene on the microphysics. And I agree with him about this -- as long as we're careful about how we understand the supervenience base.

    So, again, you're agreeing with Chalmers that consciousness fails his criterion of physicalism: It can't be ontologically reduced to physics because it doesn't logically supervene on the physics. Your disagreement lies in the question of whether other special science facts can be ontologically reduced to physics.

    My own view is that the criterion of logical supervenience is probably OK. I think Chalmers is within his rights to ask for an explanation of the what goes wrong when we think we're conceiving of a zombie (just as I could explain to you how you are misconceiving of the situation when you suppose you are conceiving of square circle).

    I actually think I can point out the mistake that dualists are making here. It isn’t that we should use biological necessity rather than logical necessity when we’re testing for ontological reduction. It is instead that we need to make sure we are properly accounting for the physical dynamics and constraints in the microphysical supervenience base. This is far more difficult than Chalmers supposes.

    At bottom, Chalmers is imagining a behaviorally identical twin, but he has not guaranteed (and cannot guarantee) that he is conceiving of a physically identical twin (where such a twin would also have to be dynamically identical to him).

    ReplyDelete
    Replies
    1. I mostly agree with you, Peter. But it sounds like your objection to Chalmers' thought experiment mostly comes down to: 'Physical systems are too complicated for us to be really carefully imagining them. If we just knew more about the nitty-gritty details of brains and minds, we'd suddenly realize there was a hidden contradiction in our thought experiment from the get-go.' That may be so, but Chalmers' response will then be to ask for at least a schematic element of the physical system that he's neglecting. You don't need to fully explain consciousness, but you do need to at least give some general example that hypothetically sounds like it could (a) bridge the explanatory gap, and (b) explain away our initial intuition to the contrary. The latter, (b), is important because we don't ordinarily think that high-level properties fail to logically supervene just because they're really really physically complicated.

      Delete
    2. Robby:

      I agree that dualists are still going to insist that I haven't bridged the explanatory gap, but I'll be happy if the argument gives someone antecedently disposed towards physicalism a way of dispelling the zombie intuition, even if it won't be enough to win over the committed dualist.

      Still, I think the argument has more to offer than you suggest, and I think it does offer a way of explaining away our zombie intuition. The argument isn’t simply pointing out that the physical details are really complicated; it’s also pointing out that unless we make some really strong assumptions, we have no way imagining what a “physically identical world” will be like at all.

      We can easily imagine a behaviorally identical world – since everything will look like it does in this world. But is such a world also physically identical? Only if we’ve included all the relevant causal-dynamical details. But have we?

      Notice that we can hold the behavior of objects fixed while imagining away various causal properties (mass, electric charge, etc.). So this makes it obvious that we can imagine worlds that look like our world even though they are missing physical properties.

      Now I ask you whether, when you are conceiving of a behaviorally identical world (sans consciousness), you can be confident that your conception is holding fixed all the dynamical features that produce the behavior you’re imagining. It seems to me that this line of reasoning should lead all of us to question whether we have really succeeded in conceiving of a zombie world.

      Notice too that the notion of a “physically identical world” isn’t even well defined in the absence of the causal closure of physics. Suppose vitalism were true: would a world that is a “minimal physical duplicate” be one in which birds continued to fly around (because we’re keeping the locations of the physical particles the same as in the actual world)? Or would it be a world in which birds dropped dead and decomposed (since there’s no elan vital in that world)?

      If it’s a world in which the birds still fly, then our conceptions aren’t going to be a good guide to whether life is physical. If it’s a world in which they drop dead, then can we really be sure that our “zombie twin” will exhibit pain reactions, color discernment, etc.?

      So I think that the contradiction that shows zombies are logically impossible is the following: The zombie is both supposed to have all the same causal features that I have (because it is physically-dynamically identical), and it is supposed to lack a causal feature that I have (because it lacks qualia, which are causal features of the actual world).

      An epiphenominalist will deny this of course, but I hold that anyone who is not antecedently committed to epiphenominalism should find zombies inconceivable. (I discovered after the fact that John Perry makes a very similar argument in his Knowledge, Possibility, and Consciousness.)

      Delete
    3. "The argument isn’t simply pointing out that the physical details are really complicated; it’s also pointing out that unless we make some really strong assumptions, we have no way imagining what a 'physically identical world' will be like at all."

      I have no way of mentally simulating thousands of H₂O molecules or genomes, and I'm not particularly fond of epiphenomenalism. Yet I share Chalmers' intuition that it's incoherent to imagine water varying while H₂O is held constant; and I share his intuition that it's incoherent to imagine complex biological facts like 'Anne is a parent of Seth' varying while biochemical facts about Anne and Seth's genetic history and structure is held constant; and I share his intuition that it is coherent to imagine phenomenal facts varying while neurological facts are held constant.

      The challenge for your view is to either show how I'm misinterpreting my own intuitions in these three cases (and ones like them), or to explain why it isn't a problem if my intuitions about the logical supervenience of the phenomenal radically and incorrigibly differ from my intuitions about the logical supervenience of the biological and the chemical. Since all three of these cases are too complicated to simulate in any detail, and since I do strongly suspect that physicalism is true, neither my computational limitations nor any prior commitment on my part to some brand of dualism/epiphenomenalism can explain away the intuitions in question.

      "Notice too that the notion of a 'physically identical world' isn’t even well defined in the absence of the causal closure of physics."

      But Chalmers does usually assume that the physical is causally closed. He isn't presupposing epiphenomenalism; he's just noting that, if we don't go out of our way to assume either it or its negation, then something like epiphenomenalism seems to fall out of how we naturally think about the phenomenal and the physical.

      "Suppose vitalism were true: would a world that is a 'minimal physical duplicate' be one in which birds continued to fly around (because we’re keeping the locations of the physical particles the same as in the actual world)?"

      Vitalism claims that the physical isn't causally closed; there are nonphysical vital essences that indispensably produce some of the behaviors we see in physical systems around us. So, presumably, if we subtracted these vital essences one day we'd suddenly see our cells fall to pieces, their metabolism shutting down or "complex adaptive behavior and reproduction" ceasing to function: http://consc.net/papers/facing.html . (Similarly, Descartes explicitly denied the possibility of zombies, or functionally identical 'automata,' on the grounds that a purely physical system couldn't possibly instantiate complex behaviors like language. So Cartesian dualism is very unlike Chalmersian dualism. Descartes is vulnerable to a vitalism analogy, but Chalmers isn't.)

      Since a vitalist cannot accept the conceivability of 'life zombies' (functionally identical non-living things), the analogy to Chalmers' view is extremely tenuous. As Chalmers notes, vitalists are only concerned with explaining functions, behaviors, causal relations. The difficulty with Chalmers' view is that it permits epiphenomenal and quasi-epiphenomenal views, on which the functional, behavioral, and causal is exactly what isn't at issue.

      Delete
    4. "An epiphenominalist will deny this of course, but I hold that anyone who is not antecedently committed to epiphenominalism should find zombies inconceivable."

      I don't understand what this means. Zombies should either be conceivable or inconceivable, for everyone. (At least, for everyone with unimpaired imaginative faculties.) Maybe epiphenomenalists are the only ones who can be both informed and deluded into thinking they're conceiving of zombies, but that isn't the same as saying they can really conceive of functional duplicates of us that lack phenomenal consciousness.

      Personally, I think zombies are conceivable. That's because I'm an eliminativist; I think they're actual. Other physicalists also think they're conceivable -- type-B physicalists who employ the Phenomenal Concepts Strategy. If all of those physicalists are deluded about the conceivability of zombies, then I'd very much like a better argument for why and how they're deluded than just 'well, they have to be deluded somehow or other, because otherwise the dualists win!'. That may be true, but if the dualists have the better arguments overall, then win they should. Epiphenomenalism is not an automatic reductio ad absurdum, any more than mathematical platonism is.

      (And indeed, these two views are very similar in a lot of ways. I'm surprised Massimo isn't more keen on epiphenomenalism, since it provides an interesting route by which one might allow numbers to causally impact our epistemic (phenomenal) states while preserving the causal closure of the physical. Of course, this would have some paradoxical down-sides as well... but it's at least worth considering.)

      Delete
    5. "The challenge for your view is to either show how I'm misinterpreting my own intuitions in these three cases (and ones like them), or to explain why it isn't a problem if my intuitions about the logical supervenience of the phenomenal radically and incorrigibly differ from my intuitions about the logical supervenience of the biological and the chemical."

      That seems like an awfully large burden. Our intuitions are very often naturally wrong and their sources almost by definition (difference between intuition and other types of beliefs) very difficult to pin down.

      Would you agree that your third intuition ("coherent to imagine phenomenal facts varying while neurological facts are held constant") implies some sort of causal efficacy outside the physical? That would seem to be the starting point for challenging the intuition.

      Delete
    6. I don't think it's too big of a burden. If anyone were able to construct a toy model of a similar error people make in some less complicated domain, and show how it might plausibly be analogous to this particular error, I would also count that as making a lot of headway. The problem of qualia is so severe in part because it seems so unique; analogies like vitalism are completely off-base, so part of what's needed here, if you feel unable to attack the problem directly, is a better supply of analogies.

      I'm not sure what you mean by "causal efficacy outside the physical". It seems to me that the original thought experiments neither assume that there are causally efficacious non-physical things, nor assumes that there aren't.

      Delete
    7. Robby:

      "I share his intuition that it is coherent to imagine phenomenal facts varying while neurological facts are held constant."

      That's an odd claim for an eliminativist. What's varying?

      My argument is meant to target someone who is antecedently disposed to think that qualia are real and produce causal effects (that are not overdetermined).

      So my target is type-B materialists, and people (e.g., Kim) who were disposed towards phsyicalism but have concluded that the conceivability argument undermines their position.

      I threw together a quick flow chart of how I see the argument going here. The relevant point is just that there is path for physicalists to follow (holding that qualia are causes) that allows them to conclude that zombies are not ideally conceivable.

      We mistakenly found them prima facie conceivable because were imagining behavior without conceiving properly of the relevant causal-dynamical facts.

      No time for more now. Best wishes, P.

      Delete
    8. In the thought experiment, you're imagining someone (Z) physically identical to you (A) but without subjective experience. So the phenomenal fact that is varying is the presence of subjective experience. The physical facts are the same.

      Your intuition is that it's coherent to imagine this. That entails that what's causing the subjective experience for A is not dependent on anything physical, since anything physical is held constant for Z.

      Delete
    9. Okay, that flow chart is awesome. It almost reads like a horoscope.

      My favorite part is, "You will need to offer independent arguments for your dualist position". A lot of time is spent trying to establish dualism. If you found a gap, fill it already.

      Delete
    10. "That's an odd claim for an eliminativist. What's varying?"

      The main question you're asking is how eliminativists can even meaningfully talk about 'qualia'. I can't define qualia extensionally, by listing a bunch of things that are intuitively phenomenal and a bunch of things that aren't, because I think the extension is empty. So I rely on two alternatives to fix the meaning of 'qualia'. First, 'qualia' are defined as the posits of theories like Chalmers'. They are whatever sorts of properties would satisfy his descriptions -- introspectively vivid, autonomous, transparent, primitive, ineffable, monadic objects of direct acquaintance by subjects of cognition. Insofar as we can coherently talk about such things, we can coherently talk about them varying independent of the physical; so the zombie thought experiment is coherent. But since we have no powerful reason to believe in such things, we should conclude that we ourselves are zombies.

      A more interesting approach, not open to most eliminativists but open to me, is to define qualia as the satisfaction conditions not of the philosophical theories of people like Chalmers, but of the ordinary quasi-perceptual meta-representations we all undergo of our experiences qua phenomenal. My eliminativism is idionsyncratic in that I think there is something about our perceptual experience that really does prima facie indicate an irreducible subjective component. Since I am more confident in physicalism than I am in the accuracy of my brain's meta-representations, I conclude that this irreducible component isn't really there; but we can still coherently imagine a world in which the component *was* there, and an epiphenomenalist world of this sort would indeed, I think, satisfy Chalmers' intuitions.

      "My argument is meant to target someone who is antecedently disposed to think that qualia are real and produce causal effects (that are not overdetermined)."

      Sure. But I wouldn't endorse eliminativism if I thought it were possible for non-eliminative physicalist accounts of phenomenal consciousness to succeed. :) I'm not pushing back against your suggestions because I think eliminative and non-eliminative physicalism are both equally good strategies for responding to Chalmers, and I just feel like being mean to the non-elminative strategies today. I'm pushing back because I suspect non-eliminative, non-dualist strategies are completely doomed. I think type-B materialism is doomed, and the Phenomenal Concept Strategy fails. To be convinced they aren't doomed, I'd need to see a lot more flesh put on the bones you provided above.

      "We mistakenly found them prima facie conceivable because were imagining behavior without conceiving properly of the relevant causal-dynamical facts."

      Chalmers agrees. He just thinks those causal-dynamical facts are psychophysical, not purely physical. He doesn't think the Zombie argument shows that qualia are in fact nomically autonomous. He thinks it only shows that qualia, as we conceive them, would need their lack of autonomy stipulated as a brute fact, rather than it falling out elegantly from our concept of qualia as it does, e.g., from our concept of water, of parenthood, and of every other macroscopic fact we've looked at.

      Delete
  5. What I wrote* to Gary Gutting's post (and got some positive recommends!):

    "I assert that I am the only one who has phenomenal consciousness (PC). Gary Gutting and everyone else are zombies. They say they have PC, but I don't believe it. Even if I scan their heads with an fMRI machine and see that their brain processes appear to work very much like mine, I still assert they are zombies. How can they prove me wrong?"

    I meant it as an ironic view of the concept of PC.

    * http://opinionator.blogs.nytimes.com/2013/03/22/mary-and-the-zombies-consciousness-revisited/?comments#permid=24

    ReplyDelete
  6. When I am in great emotional pain, I often experience this pain in a physical way: it manifests as an acidic-feeling soreness that radiates from my left shoulder all the way down the back of my arm and to my pinky and ring fingers. I read an article (I'll have to see if I can find it later) that explained how the brain processes emotional pain by piggy-backing off of existing structures that process physical pain, and this was suggested as a possible mechanism by which this phenomenon occurs.

    I bring this up, because it suggests a way for Mary to experience the color red without having to see it with her eyes. Assuming her brain's anatomy is intact, it should be possible to, via probes or whatnot, stimulate the part of her brain that responds to color - thus causing her to experience 'seeing red.' She would not be aware that the color itself is actually called red, having had no basis for comparison with the real thing - but the experience would be real nonetheless, and completely physical.

    This is the main problem I have with Chalmers - he is arguing from ignorance, and claiming that because we don't understand how this works *right now,* it much be inexplicable beyond some sort of dualism.

    This business with hypothetical zombies is nonsense. So what if it's hypothetically possible (which I very much doubt on principle anyway)? Many things are hypothetically possible, but we don't claim them to be *actually* possible. Regardless, though, the zombie argument assumes in the beginning the very thing it's trying to prove: dualism. You have to already believe in duality in order to accept the premise that zombies are hypothetically possible.

    Ultimately the problem stems from thinking about "consciousness" as a *THING* rather than a *PROCESS*. There is no such *thing* as 'consciousness' in the same way that there is no such *thing* as 'running.'

    ReplyDelete
    Replies
    1. Erik: Although I too reject Chalmers' conclusion, I have to say that you're being unfair to his arguments, and clearly haven't taken the time to fairly or rigorously assess them. Chalmers has written hundreds of pages carefully, and quite lucidly, unpacking different notions of conceivability and possibility. Whatever you think of his metaphysical leanings, the guy's a pro, and a lot of his work can be extremely useful to physicalists if only because it's so ruthlessly precise and unequivocal. I recommend checking out in more detail his presentation of the arguments in 'The Conscious Mind' (1996), which rests on no ambiguities about 'hypothetical' v. 'actual' possibility (by which you might mean nomic vs. logical possibility, or you might mean conceivability vs. possibility simpliciter). If you're left unsatisfied after 'The Conscious Mind,' check out a paper Jackson and Chalmers co-authored in 2001: http://faculty.arts.ubc.ca/maydede/mindsem/texts/Chalmers-Jackson_ConcetAnalReductiveExp.pdf

      As for the question of whether Mary could 'cheat' by triggering a red experience in her own consciousness, this is irrelevant, as Jackson already recognized in 1986: http://home.sandiego.edu/~baber/analytic/Jackson.pdf

      "The knowledge argument does not rest on the dubious claim that logically you cannot imagine what sensing red is like unless you have sensed red. Powers of imagination are not to the point. The contention about Mary is not that, despite her fantastic grasp of neurophysiology and everything else physical, she could not imagine what it is like to sense red; it is that, as a matter of fact, she would not know. But if physicalism is true, she would know; and no great powers of imagination would be called for. Imagination is a faculty that those who lack knowledge need to fall back on.

      "[...T]he knowledge Mary lacked which is of particular point for the knowledge argument against physicalism is knowledge about the experiences of others, not about her own. [...] Before she was let out, she could not have known facts about her experience of red, for there were no such facts to know. That physicalist and nonphysicalist alike can agree on. After she is let out, things change; and physicalism can happily admit that she learns this; after all, some physical things will change, for instance, her brain states and their functional roles. The trouble for physicalism is that, after Mary sees her first ripe tomato, she will realize how impoverished her conception of the mental life of others has been all along. She will realize that there was, all the time she was carrying out her laborious investigations into the neurophysiology of others and into the functional roles of their internal states, something about these people she was quite unaware of."

      For variations on Jackson's thought experiment that make this point clearer, consider Nagel's 'What is it Like to be a Bat?' and Nida-Rumelin's 'Marianna' thought experiment. http://plato.stanford.edu/entries/qualia-knowledge/#3.3

      Delete
    2. It's also false to say that any of Chalmers' or Jackson's arguments depend on the 'thinghood' of consciousness. Jackson would simply say that Mary intuitively lacks knowledge of a certain process (consciousness) even if she knows all about the microphysical things and processes and has arbitrarily great inferential capabilities. And Chalmers would simply say that all our neurological structures and processes are conceivable in the absence of any process of consciousness. No progress has been made.

      Delete
    3. I'm not talking about imagination. I'm talking about direct electrical stimulation of the precise neural pathways responsible for the experience itself. If Mary lacks these pathways, she will be unable to experience the color even if she sees it.

      The process of "seeing red" is a physical process that begins with the introduction of red light entering the eye. The redness of the light is converted into bio-electrical signals which stimulate the brain in the prescribed manner to produce the experience of seeing red. If one were to artificially stimulate the brain in an identical manner, then the subject should experience an identical sensation - unless sight is *not* a bioelectrical process, in which case we've got more hard problems than we think.

      Also, I didn't say that anyone's arguments are false because consciousness isn't a thing. =P I merely meant to highlight where I think most of the confusion comes from. 'Consciousness' means different things to different people, and so I think the word should be discarded in favor of more exact terms, like awareness or behavior - something that isn't so vague and ill defined.

      Delete
    4. Imagination and transcranial magnetic stimulation come to the same thing. The only difference is what kind of mechanism produces the spontaneous non-perceptual experience of redness. We could imagine that Mary has one of these superpowers, or that she just spontaneously experiences redness one day in a dream. But none of these seems to change the fact that Mary wouldn't know what phenomenal experience people are undergoing when they see tomatoes. Again, cf. the 'Marianna' variant.

      Delete
  7. It seems pretty clear that philosophers with religious or religious-like convictions are too often using the field of philosophy as a platform for something like apologetics.

    It's also clear that the kind of philosophy which is very closely tied to the practice of the various sciences and other disciplines is much less subject to these sorts of problems. Also, such scientifically-informed approaches help to keep the focus on real and important questions.

    ReplyDelete
    Replies
    1. Is this supposed to be relevant to the mind-body problem? I know for a fact that David Chalmers is an atheist. Frank Jackson is almost certainly an atheist as well.

      Delete
    2. Atheism is compatible with having a religious view of the world (e.g. Buddhism). It is also compatible with a belief in many religious or religious-like ideas (like the idea of some sort of cosmic moral order).

      But I was not and am not making claims about Jackson's or Chalmers's beliefs. I was alluding specifically to Gutting.

      Take religiously-oriented philosophers (like him) out of the picture and I think you would find the focus of discussion shifting away from these tedious thought-experiments.

      Chalmers's basic argument seemed to me so weak when I first read it (along the lines that Massimo explains) that I personally didn't feel the need to engage with it.

      Delete
    3. 'Your argument is dumb' and 'You sound like a religious person' are not objections. If you aren't interested in talking about the actual arguments, that's fine. But if you want others to share your views, then you'll have explain with premise of the argument is false (since it is a valid argument) and why. In the case of the zombies argument against physicalism, the premises are:

      1. If physicalism is true, then for every actually instantiated property, removing an instance of that property while leaving the purely physical facts unchanged would result in a logical contradiction.
      2. Some phenomenal properties are instantiated.
      3. Removing an instance of a phenomenal property, while holding the purely physical facts constant, would not result in a logical contradiction.

      Delete
    4. With regard to your (or Chalmers's) argument, I have doubts about the terms in which it is framed (e.g. it seems to draw a dubious parallel between phenomenal properties and physical properties). But even if one goes along with this, it is only about logical possibility. Isn't this Massimo's point also? It doesn't prove that there could actually be brains like ours that are not conscious.

      Delete
    5. Could you be more specific about why the parallelism is 'dubious'?

      "It doesn't prove that there could actually be brains like ours that are not conscious." - Again, read Chalmers. He talks about this at great length. He defines physicalism as the proposition that it is logically impossible for our world's facts to vary while its physical facts are held constant. In that sense, showing that no such logical necessity holds in the case of qualia is entirely relevant to the debate about physicalism, and Massimo is simply changing the topic when he says he doesn't care about logic. Arguing that consciousness is a biological phenomenon in no way proves that it isn't also an unphysical, dualistic phenomenon.

      Chalmers agrees with you already that there are laws of nature that make zombies and inverts impossible. His dualism doesn't claim that ghosts or zombies are possible in the real world; it only claims that there are irreducible psychophysical laws in our world that are needed to explain why zombies and ghosts aren't actual.

      Delete
  8. Massimo,
    The arguments try to show that there are certain phenomena outside the domain of physics. Hard to understand for naturalistic morons I guess.

    ReplyDelete
    Replies
    1. As opposed to supernaturalist morons?

      Delete
    2. Supernaturalistic morons don't feel they must try to misunderstand.

      Delete
  9. Massimo, you say:

    "But when we switch to the first-person feeling of perception we consider something that is, by its very nature, not a matter of scientific explanation at all. In other words, to charge science with failure of “explaining” first person experience, when one really means the feeling resulting from, not the mechanism that allows, such experiences, it so make a category mistake."

    If an experience - for instance a feeling - results from a mechanism that produces it, seems to me it's a legitimate philo-scientific question to ask for an explanation of how the feeling gets produced. That there's no canonical explanation thus far is called, as you know, the "explanatory gap." If as you say later on the problem of consciousness is a biological one, then one way to formulate the problem is: how do feelings get produced by a biological mechanism?

    But of course we can question the idea that experiences are produced in any standard causal sense by their neural correlates. Since we don't observe experience as a public object sitting along side the brain, it might be a sort of category mistake to suppose it is produced at all. Instead we could look for *non-causal* entailments between the physical and the phenomenal, for instance entailments from being a certain sort of representational system to being a subject of experience. So although there may not be an ontological reduction in the offing (to advert to Peter's good post above), there might be an empirically based, conceptually satisfying explanation that links a public, objective, physicalist ontology to private, subjective and phenomenal ontology: a naturalistic phenomenal-physical parallelism that involves facts about representation. I explore this possibility in "The appearance of reality" at http://www.naturalism.org/appearance.htm Consciousness, categorically private as it is, might well be outside the domain of science as an observable, but perhaps not beyond philo-scientific explanation as a natural phenomenon.

    ReplyDelete
    Replies
    1. I think you're on the right track.

      My approach to the explanatory gap is that there is none, really.

      If you had a complete understanding of how all causally efficacious behaviour (and beliefs) arise from physics, then you have an account of why people believe they perceive qualia and behave as if they do (e.g. making vocal claims regarding qualia).

      Once you have explained why people believe they can perceive qualia, it is not obvious to me that there is anything left to explain. If we can see how even a philosophical zombie must believe itself to be conscious and experiencing qualia, then how can we claim with any credibility that we are any more than that zombie?

      In short, consciousness and qualia may just be what we believe we are experiencing as predicted by the physics of neurons. Whether we "actually" experience them strikes me as a nonsensical question. In my view there is no difference between believing that you are experiencing something and actually experiencing something.

      For more, see various works by Dan Dennett and my own views here:
      http://disagreeableme.blogspot.co.uk/2012/12/strong-ai-illusion-of-real-consciousness.html

      Delete
  10. Massimo,

    I entirely agree with you that Jackson's and Chalmers' arguments are dreadful, and they have been eviscerated time and again in the literature (that Jackson's argument is formally invalid since it rests upon an equivocation of the term "knows" was made by several authors more than three decades ago, for example, and it amazes me that the likes of Gutting apparently still fail to appreciate this rather obvious point).

    Where I'm not sure I follow you, however, is when you claim that "Jackson’s thought experiment shows [...] that there are certain phenomena that are simply outside the domain of science [...] such as first person experiences".

    I'd be grateful if you could say a little more about in what sense, and how, you think that Jackson's argument establishes this.

    ReplyDelete
  11. Is anyone else totally baffled by the fact that Mary's hair is blue in the accompanying image? Of course, this leads to some well-known reflections on the coherence of the thought-experiment...

    ReplyDelete
  12. Hi Massimo,

    These thought experiments touch on topics dear to my heart. I have previously written blog posts giving my reactions to them.

    http://disagreeableme.blogspot.co.uk/2012/11/in-defence-of-strong-ai-qualia-and.html
    http://disagreeableme.blogspot.co.uk/2012/11/in-defence-of-strong-ai-philosophical.html

    I largely agree with your article, and in particular your reasons for rejecting the Mary's Room argument. It is precisely because of the equivocation about the meaning of the word "fact". My own formulation of the refutation regards Mary as having gained not facts but the ability to access brain states representing the new qualia.

    I differ from you slightly regarding philosophical zombies.

    >at best Chalmers has shown that there is no logical contradiction between talking and interacting like a human while at the same time not having conscious experiences. But logical constraints are too weak to tell us much about physical constraints,<

    I do regard logical possibility as being equivalent to physical possibility (at least in some possible world). I reject the philosophical zombie argument because on analysis it turns out to be circular and fails to establish that philosophical zombies are even logically possible.

    Conceivability as a guide to logical possibility is a heuristic that only works reliably in simple, well understood cases where we are confident we can correctly think through all the implications.

    We are confident that a square circle is logically impossible because those are pretty simple concepts and we would expect to be able to fully "simulate" such a state of affairs in our imaginations if possible. Since we can't do this, we say it's inconceivable and conclude that it is logically impossible.

    For more complicated states of affairs, conceivability works quite differently and is no longer helpful as a guide to logical possibility. An even number greater than two which is not the sum of two primes is certainly conceivable in the abstract, however nobody knows if such a number is in fact logically possible.

    A counter-argument is that unless we can identify such a number, then we have not really conceived of it, because the identity of that number must be part of a complete mental conception. Instead, we are making a rather more loose mental association between concepts of "prime" and "even" without mentally simulating a precise state of affairs.

    Well, fair enough. But when Chalmers conceives of a philosophical zombie, he means this latter, more abstract form of conception.

    Intelligence and consciousness are certainly complex phenomena. When we conceive of philosophical zombies, we are not actually conducting detailed "imaginative simulations" but just loosely associating the concepts of intelligence and unconsciousness.

    If the physicalist worldview is correct, and in particular the computational theory of mind, then consciousness is just the subjective aspect of the computations occurring in the brain. You couldn't have intelligent behaviour without those computations, and you couldn't have those computations without their subjective aspect. To assert that a philosophical zombie is conceivable or logically possible is just to beg the question and assert that consciousness is not simply the subjective aspect of computations.

    As such, I reject the philosophical zombie argument not because logical possibility is distinct from physical possibility, but because it is fundamentally circular.

    ReplyDelete
  13. The earliest example of a zombie-type thought experiment I know if is not in Chalmers, but in G.F. Stout, where he uses it to argue against Huxley's epiphenomenalism. I believe the term 'zombie' itself, in the philosophical sense, comes from Robert Kirk's work in the 1970s.

    "But when we switch to the first-person feeling of perception we consider something that is, by its very nature, not a matter of scientific explanation at all."

    Why not? If first-person experience is just as much a part of the real world as is any other phenomenon — if, in other words, we are not inclined towards eliminativism — then why shouldn't a perfected neuroscience allow us to understand just as much about 'what it's like' in the first person to undergo some experience, as about the more traditional third-person facts? Why, for example, will we never be able to gain a complete neuroscientific understanding of what it's like to be a bat? (Or, for that matter, of what it's like to go swimming?)

    Perhaps you have in mind that science is restricted to third-person empiricism, and that the first-person facts are (by definition) not third-person facts. But even if we grant this as an irreducible distinction (and I worry that doing so puts us only a hair's breadth away from dualism!), it doesn't follow that we couldn't infer all the first-person facts from the third-person ones. Why would such an inference be blocked, if the first-person facts are in some sense 'fixed' by the third-person ones? (And we must admit that they are fixed, if we deny the metaphysical possibility of zombies and inverts.)

    "to charge science with failure of 'explaining' first person experience, when one really means the feeling resulting from, not the mechanism that allows, such experiences, [is to] make a category mistake."

    Ordinary scientific practice isn't only concerned with positing mechanisms that account for the data, though. It's also concerned with describing and predicting data. The concern raised by dualists is that although a perfected neuroscience could posit mechanisms that are causally sufficient for experience, it could not redescribe experience in physical (or non-mentalistic biological) terms or come up with a purely physical theory (one devoid of, for instance, brute psychophysical laws) that predicts experience. And this seems to be a serious problem for physicalism, even if it isn't a problem for ordinary scientific practice.

    ReplyDelete
    Replies
    1. "What Chalmers (and Gutting, and several others) don’t seem to understand is the simple proposition that the problem of consciousness is a biological, not a logical one."

      I think Chalmers would respond: 'Sure, I'll happily grant that consciousness is a biological problem. But it's a unique and extraordinary biological problem, one requiring radically new (in fact, metaphysical!) resources.

      'Consider, for example, the property of being a parent. Intuitively, it seems that a logical contradiction would ensue if we held all the physical facts constant, while varying the distribution of biological properties like "parenthood" in our universe. Parenthood logically supervenes on the physical. In fact, all biological properties other than consciousness seem to logically supervene on the physical; there is no space, even logically or conceptually, for them to be subtracted without modifying their physical substrate. So consciousness, though biological inasmuch as it is a feature of living systems, is sui generis. There is an important sense in which it violates physicalism, and no other biological process violates physicalism in the same sense.

      'When I say that dualism is true, all I mean is that consciousness is unique among biological phenomena in that there is a permanent "open question" as to its nature and existence when we restrict our attention to its physical/chemical substrate. (This is analogous to Moore's "open question" in meta-ethics, though I think eliminativism about moral facts is a lot more reasonable than eliminativism about phenomenal facts.) I'm happy to grant that zombies aren't nomically possible, i.e., that there are psychophysical bridge laws that rule zombies out. I don't think there's anything scientifically illegitimate or unnatural about that; I just happen not to think that irreducible mind-matter laws should be called "physical". They're a bit too weird and unprecedented for that.'

      Delete
    2. >Why not? If first-person experience is just as much a part of the real world as is any other phenomenon — if, in other words, we are not inclined towards eliminativism <

      Maybe you've answered your own question? Perhaps eliminativism is the answer. Perhaps first-person experience is not a part of the real world.

      >first-person facts are (by definition) not third-person facts.<
      I don't think there are any first-person facts. Qualia, as Massimo explained, are not facts. But if you replace "facts" with "phenomena", then fair enough.

      >Why would such an inference be blocked, if the first-person facts are in some sense 'fixed' by the third-person ones?<
      It seems to me we can only infer facts from facts, and there are no first-person facts (only phenomena, or experiences). We could of course infer third-person facts such as that a person will believe and behave as if they are experiencing first-person phenomena.

      For those who hold that these phenomena really exist, we therefore lack an explanation of how they arise from physical interactions. For the eliminativist on the other hand, explaining how the belief in qualia arises is a full explanation.

      And as I've said elsewhere on this thread, once we've explained why we believe we perceive qualia, haven't we explained everything we have to? Assuming that qualia have some real ontological existence beyond physical brain state seems to me to be making the same mistake as those who argue that objective morality truly exists simply because we all intuit that some acts are wrong.

      Delete
    3. Well, I do happen to be an eliminativist about phenomenal consciousness. :) But I don't think eliminativism is •obviously• or •trivially• right, and I take seriously both Chalmers' arguments and his dualistic view. The world •seems• dualistic, and it will take some very sophisticated reasoning to explain way the epistemic obviousness of our own first-person consciousness without landing ourselves in absurdity or incoherence.

      "Qualia, as Massimo explained, are not facts."

      This is a little confusing. Chalmers would agree that qualia aren't facts; instead, they're properties. However, if Chalmers is right, then it's a fact that people possess the properties in question. Could you point me to where Massimo denies any of this?

      "there are no first-person facts (only phenomena, or experiences"

      I don't know what you mean by a 'phenomena/fact' distinction. Isn't experience, or at least phenomenal experience, precisely what the eliminativist is denying? Perhaps by 'phenomenon' you mean an •apparent• fact, but it's not a trivial task to explain what 'apparent' even means in an eliminativist theory!

      "For the eliminativist on the other hand, explaining how the belief in qualia arises is a full explanation."

      This is a common view, but I disagree with it. I think that in addition to explaining why people believe as they do, the eliminativist must also explain why our ordinary perception seems to •justify• such beliefs. Introspectively, we don't seem to be phenomenal zombies plagued by arbitrary, free-floating beliefs; rather, our phenomenal beliefs seem to be accurately mapping how it feels to be in, for instance, perceptual states. As an eliminativist, I owe people like Chalmers a plausible account of why perception seems to instantiate phenomenal properties. Since perception is not itself a kind of belief, this is a distinct explanatory burden, and, I would say, the more fundamental, interesting, and important one.

      "as I've said elsewhere on this thread, once we've explained why we believe we perceive qualia, haven't we explained everything we have to?"

      Not necessarily. In addition to the issue of perception I raised above, we also have to establish that eliminativism is plausible in the first place. Explaining why people believe in X is not itself sufficient for eliminating X! Moreover, Chalmers has written at length about this very topic. See the quotation here: http://lesswrong.com/lw/5ot/nature_red_in_truth_and_qualia/8etq

      Delete
    4. >But I don't think eliminativism is •obviously• or •trivially• right, and I take seriously both Chalmers' arguments and his dualistic view. <

      I think we're agreed on this. I don't take the philosophical zombie thought experiment to prove anything, however. In order to allow the premise you have to be a dualist in the first place. It's good as an intuition pump to explain what dualism entails, however.

      > Could you point me to where Massimo denies any of this?<

      I was referring to this bit: "The (apparent) force of Jackson’s hypothetical situation relies on an equivocation about the word “fact.” Mary had studied all the scientific facts about how the brain works, but those facts do not equate with having the first person experience of actually seeing color"

      I take this to mean that a fact is some bit of knowledge that can be expressed in words. Qualia are not facts in this sense. They are purely experiential.

      I want to respond to more but I don't have time at present. Will respond further tomorrow!

      Delete
    5. Just wanted to add to this: "Perhaps you have in mind that science is restricted to third-person empiricism, and that the first-person facts are (by definition) not third-person facts."

      It's a common mistake to suppose that physics can't deal with "subjectivity" and "points of view." But on most readings of these terms, this is false.

      Physics has no problem explaining why I can see the top of my desk but not the bottom. It has no problem explaining why I feel my computer keyboard but you don't.

      Indeed, 90% of physics is figuring out how different "points of view" fit together. Relativity theory is all about how to transform between different coordinate systems; quantum theory is about fitting together different bases of Hilbert space; gauge theory is all about making proper gauge transformations; etc.

      Delete
    6. I think the following gets at the heart of the "qualia" argument against physicalism:

      "The concern raised by dualists is that although a perfected neuroscience could posit mechanisms that are causally sufficient for experience, it could not redescribe experience in physical (or non-mentalistic biological) terms or come up with a purely physical theory (one devoid of, for instance, brute psychophysical laws) that predicts experience. And this seems to be a serious problem for physicalism, even if it isn't a problem for ordinary scientific practice."

      Unfortunately, I also don't understand this argument. It's not for lack of trying. But no one has ever been able to explain to me why this might be a "serious problem for physicalism." You say that "Science is concerned with describing and predicting data," and perhaps you would respond that if no physical theory can predict experience, then that's a sign that our physical theories are inadequate, and that therefore we need a non-physical theory of consciousness.

      But I would respond that no physical theory that has been taken seriously over the last half-century claims to be able to predict the state into which a quantum wave function will collapse upon measurement. It's true that this has been seen as a problem for quantum physics by some people. Nonetheless, I suspect that anyone who walked into a room full of mainstream physicists claiming that we therefore need a nonphysical theory of quantum physics would be laughed back out.

      Why are we holding our best science to higher predictive standards for consciousness than for wave function collapse? It would seem that we should be doing the exact opposite! It's no surprise that we can't explain something as complex as consciousness; we can't even fully explain the behavior of the fundamental building blocks of matter.

      Delete
    7. "In order to allow the premise you have to be a dualist in the first place."

      No. Type-A materialists generally grant the 'if zombies were conceivable, phenomenal properties would be non-physical,' and type-B materialists generally grant the 'zombies are conceivable' premise. Materialists disagree among themselves about which of these two premises to reject, so it's not as simple as saying that there's a specific premise that all dualists like and all physicalists hate.

      Actually diagnosing where the zombie argument goes wrong is extraordinarily difficult. As an eliminativist, for example, I can happily grant both of the aforementioned premises, and acknowledge that zombies are conceivable and that their conceivability proves that phenomenal properties are non-physical. I merely deny that this is a problem, since I deny that these conceivable, non-physical phenomenal properties are •instantiated•. In other words, I think zombies are not only conceivable and possible but •actual•.

      "I take this to mean that a fact is some bit of knowledge that can be expressed in words. Qualia are not facts in this sense."

      I think Massimo's point wasn't that qualia are ineffable, but that, even if effable, they wouldn't be •scientific•. Thus facts concerning qualia, whether linguistically expressible or not, are not •scientific• facts. He can correct me if this is wrong.

      Incidentally, I think Dan Dennett would agree with Massimo here, especially if Massimo's concern with phenomenological data is its first-person character: http://ase.tufts.edu/cogstud/papers/chalmersdeb3dft.htm . Personally, I think the question of whether this debate is 'scientific' is irrelevant. What matters is whether physicalism is true, not whether our favorite definition of science cares about physicalism.

      Delete
    8. Scott: You raise the objection that it's OK if not all mental facts are entailed by the physical facts, because quantum physics itself is indeterministic and yet still counts as a physical theory.

      I don't think the analogy between the two cases is very strong, though perhaps my presentation was unduly vague and invited interpretations of this sort. In the case of qualia, the problem is not that qualia seem to follow from physical laws that are only probabilistic; the problem is that qualia seem to follow from laws (which might be partly probabilistic, but in practice are surely almost entirely deterministic) that are not strictly physical, or that would require us to enrich the language of physics with brute appeals to experience. A probabilistic physics is still recognizably physicslike, whereas a physics expanded with fundamental terms meaning things like 'how it feels to experience redness' would be much less recognizably physicslike.

      In the quantum case, we have apparent nomic indeterminism, but not physical inexpressibility or a dependence on moral or mental or otherwise bizarre/unphysical brute facts/laws. In the qualia case, we have apparent nomic determinism (as Chalmers would grant), but apparent physical inexpressibility and an apparent dependence on weird fundamental laws.

      The quantum case might refute some particular kinds of physicalism (specifically, deterministic ones), but it doesn't refute physicalism simpliciter, because, plausibly, determinism is an inessential component of what we mean by 'physical world,' whereas the lack of fundamental, irreducible moral and mental and magical processes is an essential component of what we mean by 'physical world.'

      (Incidentally, I don't actually think the above interpretation of QM is right; I accept determinism. But that's a separate issue, and it's true that QM at least seems indeterministic, at first glance.)

      Delete
    9. Robby, you say:

      "Well, I do happen to be an eliminativist about phenomenal consciousness."

      I'm wondering if this means you don't think experiences (e.g., pain) exist. Or is it that experiences exist but they aren't non-physical (thus avoiding dualism)?

      If pains don't really exist, but it only *seems* like they do, it's still just as morally consequential that we not torture babies, in which case the seemings to have experiences play the same role in our psychological and moral economies as (non-existent) experiences. But then these seemings seem a lot like the experiences you've eliminated. Is there a difference that makes a difference, one might ask? To seem to be something, or to seem to have something (like an apparent experience), is for something to appear to be the case, to be like something, which of course is Nagel's characterization of experience itself. So it could be that all the explanatory problems simply resurface when attempting to quine conscious experiences as mere seemings (Dennett's approach and perhaps yours).

      If, on the other hand, you're just denying that experiences and phenomenal properties are *non-physical*, then you're not an eliminativist with respect to phenomenal consciousness, but simply a physicalist.

      Hope you'll correct any misunderstandings on my part, thanks!

      Delete
    10. The word "seems", I think, is ambiguous. It can be epistemic in meaning ('X seems to me to be the case iff I have evidence that on reflection decisively disposes me to believe in X'), or it can be phenomenological in meaning ('X seems to me to be the case iff I have an experience as of X occurring').

      To see that these two ideas can come apart, consider the Twin Tables optical illusion: http://www.psychworld.com/optical-illusion-of-the-week-the-twin-tables-2011-01 . To typical observers, it initially seems as though the vertical table is longer than the horizontal one. But what happens when the observer doxastically realizes that the tables are the same length, while continuing to visually misperceive them as different? There's a sense in which the tables cease to 'seem' to be different lengths to him; he has seen through the illusion, and things seem as they really are. But there is another sense in which the tables still 'seem' to be different lengths -- the sense in which his perception still indicates as much, even though he is no longer at all inclined to accept this indicator in his beliefs. Even the word 'phenomenal' itself attests to this duality, since philosophers generally use it in the phenomenological sense, but etymologically its original meaning of 'apparent' indicates that it had a strongly epistemic meaning.

      My view is that phenomenal seeming itself is an illusion -- that it epistemically seems to us as though things phenomenally seem some way, but in reality there is no such thing as phenomenal seeming. So I deny that there are pains, or experiences in general, in the sense that I deny there are first-person, or subjective, properties and facts. But (like most eliminativists) I do not deny that there are brain-states that fill a functional/behavioral role extremely similar to our folk concept of 'pain', and (unlike many eliminativists) I do not even deny that our perception, and reflection thereupon, gives good prima facie reason to believe in phenomenal properties. The epistemic seeming isn't itself an illusion; we really do epistemically seem not to be zombies. But this seeming is nonveridical; it stems from a pervasive second-order illusion, an illusion about our experience itself, analogous to lower-order optical illusions. Any adequate theory of consciousness, in my view, should be able to explain how and why this illusion manifests. If it persistently fails to do so, then dualism will be vindicated.

      Delete
    11. Notice that although I define epistemic seeming in terms of whether something disposes us to have beliefs, I don't assume that beliefs themselves are the only explanandum relevant to phenomenal consciousness. Unlike Dennett, I think that there are non-doxastic (access-)conscious mental states that are the source and basis of our reflective phenomenal beliefs. When I put myself in Mary's shoes, or think about zombies or inverts, my inclination to believe in dualism doesn't just spontaneously arise from nowhere; it seems to be eminently grounded and warranted by the character of my moment-to-moment introspection. This apparent character need not be exhausted in beliefs about that character; perception and quasi-perceptual states, in other words, need not be reduced to belief-like states. All the eliminativist need insist on is that to the extent any non-belief-like states (e.g., color perception) represent themselves as bearing phenomenal properties, these states are nonveridical.

      This is a very radical move, but it's less radical than Dennett's approach in some respects, more radical in others. Overall, I think it's more interesting, plausible, and dialectically successful than Dennett's pseudo-eliminativism. It appeals directly to physicalism and a belief in the fallibility of introspection for its plausibility, whereas Dennett's account is mainly grounded in methodological behaviorism and some rather dubious phenomenological assertions:

      "I am left defending the view that such judgments exhaust our immediate consciousness, that our individual stream of consciousness consists of nothing but such propositoinal episodes, or better: that such streams of consciousness, composed exclusively of such propositional episodes, are the reality that inspires the variety of misdescriptions that pass for theories of consciousness, both homegrown and academic.... My view, put bluntly, is that there is no phenomenological manifold in any such relation to our reports. There are the public reports we issue, and then there are the episodes of our propositional awareness, our judgments, and then there is -- so far as introspection is concerned -- darkness."

      My response is that if we take introspection to be constitutively phenomenal, then I am much more of an eliminativist than Dennett, since I think that 'darkness' swallows up even our episodes of propositional awareness. But if we instead take introspection to be a real process by which we receive flawed and noisy, but not completely unreliable, information about our own present brain states, then I think this internal monitoring reveals a much more diverse sequence of cognitive events than the mere succession of propositional judgments that Dennett asserts. One might say that my view is phenomenally impoverished, but cognitively-perceptually enriched, compared to Dennett's view.

      Delete
    12. The analogy between the two cases is indeed poor, and only useful insofar as both cases involve a failure of prediction. I'm happy to jettison it in favor of your more precise statement of the problem you see qualia as posing for the physicalist:

      "the problem is that qualia seem to follow from laws... that are not strictly physical, or that would require us to enrich the language of physics with brute appeals to experience."

      So we have two possibilities: 1) qualia follow from nonphysical laws or 2) qualia require us to make brute appeals to experience. I have two corresponding questions:

      First, what's the argument supporting the first possibility? I can't think of one that doesn't rely on equivocation between ontological and epistemological senses of the term "physical."

      Second, what exactly is the problem with the second possibility for physicalism? Aren't we -- every day -- making brute appeals to experience in science? And isn't all of our scientific knowledge of the physical world precisely the result of the compilation and systematization of the raw experience of individuals? Isn't that empiricism (in the Humean sense) in a nutshell?

      You say that "A probabilistic physics is still recognizably physicslike, whereas a physics expanded with fundamental terms meaning things like 'how it feels to experience redness' would be much less recognizably physicslike." But "how it feels to experience redness" is a precondition for "physics" in the sense you seem to be using it here. ("A set of theories about the world produced by a scientific discipline as exercised by embodied humans.") We have experience before we have physics, and if physics can't systematize every part of our experience perfectly, then perhaps there's a fundamental constraint on what can possibly be known through the discipline of physics. I don't see why we must therefore conclude that those aspects of our experience that cannot be systematized must not be physical in the ontological sense.

      Finally, you say that "lack of fundamental, irreducible moral and mental and magical processes is an essential component of what we mean by 'physical world.'" But here again, I find myself thinking that if "mental" processes are the basis for all our knowledge (our "justified true beliefs"), then how can we possibly place them in the same category as "magical" processes?

      So overall, it seems quite possible to me to posit a physics that is subject to certain "no-go" theorems showing that the quiddity of personal experience cannot be fully absorbed into our physical knowledge. We can still make lots of reasonable claims about what another being's personal experience could or could not be like, but we can't directly have the experience that the other being has, or determine the exact character of that experience. It seems to me that this is likely to be true for the physical reason that the fine structures of our bodies and brains are different. None of this seems to me to contradict physicalism.

      Delete
    13. Robby:

      “The epistemic seeming isn't itself an illusion; we really do epistemically seem not to be zombies. But this seeming is nonveridical; it stems from a pervasive second-order illusion, an illusion about our experience itself, analogous to lower-order optical illusions. Any adequate theory of consciousness, in my view, should be able to explain how and why this illusion manifests. If it persistently fails to do so, then dualism will be vindicated.”

      Thanks Robby, most interesting. I’m not sure how, if experience doesn’t exist (as you claim it does not), we can have “an illusion about our experience itself.” Rather, I think you mean we’re under the illusion we have experience.

      An epistemic seeming as you explain it is a judgment that such and such seems to be the case, hence propositional, e.g., one line is longer than another. But what you claim is the real but non-veridical epistemic seeming that there’s experience (the illusion of having experience) is given as a non-discursive, non-propositional presence of qualia, about which we can make judgments (I was in pain a second ago, and now I’m not). This seeming, this illusion, is *itself* undeniably qualitative and undeniably unavailable to anyone else, hence subjective as opposed to objective; it’s an example of what we mean by a phenomenal property. As you put it, it’s “the character of my moment-to-moment introspection,” not just a matter of doxastic states and propositional judgments. So again it seems to me that your eliminativist account leads us back to the existence of experience and phenomenal properties as normally construed: subjective qualia. But I have to think more about this in light of your most interesting proposals, thanks!

      Delete
    14. @Robby
      >No. Type-A materialists generally grant the 'if zombies were conceivable, phenomenal properties would be non-physical,' and type-B materialists generally grant the 'zombies are conceivable' premise.<

      Fair enough, however I regard any formulation of Type-B materialism which allows the existence of philosophical zombies to be incoherent -- for much the same reasons as Chalmers does.

      This leaves us with Type-A, zombie-denying Type-B or dualism, and of those, the philosophical zombie argument only works if you are a dualist. I suppose I could modify my criticism to state that it is only circular with respect to denying certain types of materialism, and perhaps does successfully refute other types of materialism.

      In my view, the position of a zombie-denying Type-B materialist is not that dissimilar from that of a Type-A materialist. If they differ only on whether qualia and consciousness actually exist, then to me it simply depends on how you look at it, or what you mean by "exist". As with mathematical Platonism, the whole debate is mired in the vagueness of the concept of existence and as such I don't think there is a correct resolution to it.

      Similarly, your own position seems to be pretty much the same as mine, even though I deny the possibility of zombies and you insist they do exist. This is because when you suggest that *we* are zombies, you are implicitly denying the possibility of a zombie which is qualitatively different from us with regard to (pseudo-) consciousness, and it's that latter kind of zombie I take Chalmers to be discussing and of which I deny conceivability.

      Delete
    15. @Robby
      >I think Massimo's point wasn't that qualia are ineffable, but that, even if effable, they wouldn't be •scientific•.<

      You could be right. I'm not going to second-guess Massimo on this.

      But I still have my own views to fall back on. In response to your original concern over where the block might be in inferring first-person "facts" from third-person facts, I suggest that the answer is that you can't derive something ineffable from something effable.

      But I do feel it's wrong to call first-person experiences "facts" or "data". I don't think they are informational in nature at all, and these words imply otherwise.

      Delete
    16. @Robby
      > I think that in addition to explaining why people believe as they do, the eliminativist must also explain why our ordinary perception seems to •justify• such beliefs.<

      I think this is one of the few points we disagree on.

      From an eliminativist perspective, we can account for all behaviour and beliefs in terms of physical interactions. We can therefore explain why people believe they experience qualia, and indeed why it seems to people that their perceptions justify this belief. In fact, the latter is pretty much the same as the former.

      >we also have to establish that eliminativism is plausible in the first place. Explaining why people believe in X is not itself sufficient for eliminating X!<

      Well, it may be good reason to doubt X if the explanation does not depend on X actually existing, especially if X is incoherent and/or ineffable.

      As for why eliminativism is plausible, it is because it is the most parsimonious account which is consistent with the facts.

      Before we go further, I want to say what I mean when I refer to qualia.

      Unlike Dan Dennett, I actually think it is useful to talk of qualia, even if only to refer to the false intuitions we all share. The discussion might then be cast in terms of how this term should be interpreted. I prefer this to debating whether they actually exist (as they clearly do, for certain interpretations of existence).

      A "physicalist quale" can be thought of as an abstraction, a mental label which corresponds to a certain pattern of neuronal activity. Its function is to aid the brain in distinguishing different percepts from each other, and it has no content other than that needed for this function.

      In this view, there are are two possibilities, as I see it:

      (1) Our intuition that qualia are more than labels is correct and so our belief in ineffable, mysterious contentful qualia is justified
      (2) Qualia are no more than contentless mental labels and our intuitive beliefs otherwise are misguided.

      Now, (1) implies dualism to me, and casts doubt on basic foundational assumptions of science, so it needs some pretty extraordinary evidence to back it up.

      But no available evidence can discriminate between (1) and (2). In this sense, there is nothing left to explain. There is no observation that is not explained by (2), so there is no reason not to adopt (2) as the more parsimonious position.

      > Since perception is not itself a kind of belief<

      I suspect this is only true if dualism is correct.

      Of course there is a difference between the perception of colour the belief that Paris is the capital of France. However, the difference is arguably only in the domain of the belief. In this view, perceptions are beliefs in the domain of "current sensory input". It feels different because our brains are wired to process these beliefs differently -- current sensory input is of more immediate importance than facts about foreign capitols.

      In other words, if I could make you believe you were experiencing a familiar quale such as the colour blue, then you would genuinely be experiencing that quale.

      I don't know enough about it to be confident in this, but I suggest hypnosis as one example where this might actually happen.

      Dreams are a more familiar case. It's debatable whether you are actually perceiving anything, but you believe you are, and so you experience those qualia.

      Delete
    17. @Tom Clark

      >If pains don't really exist, but it only *seems* like they do, it's still just as morally consequential that we not torture babies, in which case the seemings to have experiences play the same role in our psychological and moral economies as (non-existent) experiences. But then these seemings seem a lot like the experiences you've eliminated. Is there a difference that makes a difference, one might ask?<

      This is a good point, and I'd like to give my answer as a physicalist who believes in qualia if they are identified with your "seemings".

      The difference between true experiential phenomena and seemings is that seemings can be accounted for physically. Seemings in this sense can be regarded as beliefs about sensory input, and these can be explained in terms of neuronal firing, etc.

      Qualia "seem" to be contentful, ineffable, mysterious sensations that are more than their physical correlates. If physicalism is true, then it may be that they are no more than discriminatory labels for sensory data with no actual content of their own. As such, qualia do exist in this limited sense, and it only "seems" like they are more than that.

      Delete
    18. Scott:

      "First, what's the argument supporting the first possibility [that qualia follow from nonphysical laws]?" - Here's one:

      1. If a law can only be expressed by making explicit reference to mental things qua mental, then the law is not physical.
      2. Phenomenal properties are instantiated in accord with our universe's laws.
      3. Phenomenal properties are mental.
      4. Phenomenal properties cannot be redescribed in non-mental terms.
      5. Therefore phenomenal properties are instantiated in accord with non-physical laws.

      "Second, what exactly is the problem with the second possibility for physicalism? Aren't we -- every day -- making brute appeals to experience in science?"

      It's not a 'brute' appeal if it's only provisionally non-reduced, but admits of being reducible someday. 'Brute' as I intend it here means that it's forever irreducible. It's not spooky if science can't currently reduce experience to a blind fundament, but it is spooky if science can't EVER reduce experience to a blind fundament.

      "isn't all of our scientific knowledge of the physical world precisely the result of the compilation and systematization of the raw experience of individuals? Isn't that empiricism (in the Humean sense) in a nutshell?"

      Sounds more like phenomenalism to me. (Phenomenalism is an idealism (or sometimes a neutralism), an anti-physicalism, that was mainstream in the early 20th century. It accepts the irreducibility of consciousness and argues that we should reduce the physical to the relationship between conscious events, rather than vice versa.)

      "But "how it feels to experience redness" is a precondition for "physics" in the sense you seem to be using it here."

      1. No, it isn't. Color-blind physicists can do good physics. Even a robot could, I'd argue, do good physics.

      2. Even if color vision were a prerequisite for physics, it wouldn't follow that color vision is irreducible (or that it's unproblematic or unsurprising for it to be irreducible!). Electron microscopes are also a prerequisite for physics as we understand it today, but electron microscopes aren't physically inexplicable. Even the practice of science itself admits of scientific explanation.

      Delete
    19. "perhaps there's a fundamental constraint on what can possibly be known through the discipline of physics. I don't see why we must therefore conclude that those aspects of our experience that cannot be systematized must not be physical in the ontological sense."

      Well, dualists agree that there's a fundamental constraint on what can be known through physics -- because they think there are irreducible mental facts in addition to physical ones. Your challenge is to explain why your view isn't just a form of dualism, and why it's a more reasonable explanation for the apparent irreducibility of mind than is dualism. In particular, this will require explaining how and why a 'fundamental constraint' could block our physical knowledge of the nature of experience in a world in which experience nevertheless is in some hidden way reducible to the physical. (In other words, you need to explain why we should accept mysterianism.)

      "if 'mental' processes are the basis for all our knowledge (our 'justified true beliefs'), then how can we possibly place them in the same category as 'magical' processes?"

      If we discovered that tiny electron microscopes are floating around in the sea of electrons and quarks, that would fundamentally upend our concept of the universe even though we rely on electron microscopes to learn about microphysics in the first place. In the same way (though to a greater extreme), discovering that experiential states are fundamental to nature would completely upend our notion of what nature is, even though we relied on those experiential states to come to understand nature in the first place. Learning, for instance, that quarks have inner lives or experiences, would completely shatter anything reasonably similar to our current concept of 'the physical' whether or not those quarks were also sorcerers or otherwise magical.

      "we can't[...] determine the exact character of that experience. It seems to me that this is likely to be true for the physical reason that the fine structures of our bodies and brains are different."

      Our brains are very different from most of the things we think about. Certainly our brains are more like those of other humans, or like those of bats, than they are like storm systems. Yet it doesn't seem that there is in principle anything we could never come to know about storm systems, except in virtue of our computational limitations. So why should there be anything we can never in principle come to scientifically learn about other people's (or animals') experiences, for reasons other than boring old computational limitations? Why is there a hard problem of consciousness but not a hard problem of El Niño?

      Tom:

      Your argument seems to assume that if something is not itself linguiform, then it cannot be fully described linguistically. I don't accept this. Acts of perception, and cacti, and bookshelves, are not structured propositionally; but everything about them can still be captured in third-person scientific language (at least, if eliminativism is true). Some of our mental events are sentence-like, while others are organized more like melodies or short films or static pictures. In claiming that our mental events exhibit this diversity, and allowing that the apparent diversity need not be reducible to a cognitive process that is completely propositional, I part ways with Dennett. But in allowing that our brain can organize information in non-linguiform ways, I need not also allow that our brain can organize information in *ineffable* ways, i.e., in such a way that it is impossible for a third-person observer to fully model what's going on linguistically. That would be tantamount to claiming that sentences can only describe other sentences.

      Delete
    20. Part of the problem is that I don't consider epistemic seemings "judgments". Rather, I consider them a more general class of mental events that dispose one to judge in a certain way. An act of seeing a red ball can involve it 'seeming as though there's a red ball before you' inasmuch as it disposes you to make the latter judgment, even though that seeming precedes any actual propositional thought about balls or about your vision. Seeming is causal/dispositional.

      Thinking more about the red ball example, my claim is that our perceptual faculties really do represent there being a red ball before us. And this representation might well be veridical. But we also have a meta-representation of undergoing a visual episode involving a representation of a red ball. This meta-representation, like the original representation, might be veridical in some respects and nonveridical in others. My view is that there is a pervasive and constant element of nonveridicality in such meta-representations inasmuch as they perceptually impute qualia to many of our first-order representational states. But in asserting that the apparent qualitative/phenomenal aspect of our perception is not really there, I need not also assert that non-phenomenal aspects of our perception are absent as well. For instance, I grant that there really is information in my brain that's distinguishing the wavelength of light emitted from the ball from that emitted by the ocean, by dandelions, etc. in a systematic way. I just deny that this perceptual distinction is as rich as it epistemically seems to be. I deny that an agent who hasn't experienced red is missing out on anything, other than missing out on undergoing a certain illusion.

      Disagreeable Me:

      "I regard any formulation of Type-B materialism which allows the existence of philosophical zombies to be incoherent -- for much the same reasons as Chalmers does."

      I don't understand what you mean. Which Chalmers argument are you alluding to? Type-B materialists by definition deny that zombies are possible. They just allow that they're conceivable, and try to give arguments for why conceivability and possibility come unstuck in this case even though they match up surprisingly well in other cases.

      "the philosophical zombie argument only works if you are a dualist"

      I don't know what you mean by the argument 'working'. Again, our task is not simply to deny the conclusion; it's to explain what's wrong with the argument. If you're suggesting that it's impossible for someone to find the zombie argument initially plausible without already being a dualist, then you are simply mistaken. Lots of people, myself included, have found the argument's premises quite acceptable and had a correspondingly hard time figuring out where the argument goes wrong. If the premises were presented in isolation, without any effort to combine them all into an argument against physicalism, I expect that most physicalists would find them completely unobjectionable, albeit perhaps a bit eccentric.

      "I suggest that the answer is that you can't derive something ineffable from something effable."

      But how would a completely effable microphysics yield an ineffable higher-level phenomenon? Can you give an example of how that could come about?

      Delete
    21. "We can therefore explain why people believe they experience qualia, and indeed why it seems to people that their perceptions justify this belief. In fact, the latter is pretty much the same as the former."

      I think you're relying on an equivocal use of 'seems'. Seeming-that-X and judging-that-X are not the same thing. Consider optical illusions. If you provided a complicated theory that explained people's belief formation about the Two Tables as a spontaneous but consistent error in judgment that has nothing to do with a misrepresentation in their visual field itself, then your theory would be false, or at best incomplete.

      Similarly, I claim, if you explain the Grand Illusion of Qualia merely in terms of beliefs/judgments, without taking into account the fact that these nonveridical judgments causally stem from nonveridical pre-doxastic states (either perceptual or quasiperceptual), then your theory will either be incomplete or simply false.

      We could imagine a world with agents like the ones Dennett describes, who have no perceptual states but spontaneously form judgments serving similar functional roles. It would basically be a world where everyone has extremely complex and skillful blindsight with respect to all of their senses. So I'd know about my environment via a string of descriptive propositions, but I'd have no visual, auditory, olfactory, etc. sensations; I'd have no access-conscious mental events structured like pictures and songs (for examples), merely ones structured like lots and lots and lots of sentences. But although we could imagine building an agent like that, we don't seem to be an agent of that sort. We have multiple sense-modalities, and the distinction between sense-modalities isn't simply a difference in the grammar or vocabulary of certain sentences.

      I won't go so far as to say that Dennett's view is impossible; I already think systematic delusions about our introspection are possible, so perhaps this is another one of those. But I see no phenomenological or neurological reasons to think that we actually do suffer from such a delusion; there is no 'hard problem' of olfaction or perception distinct from the hard problem of phenomenal consciousness, so once we dispose of the latter there is no reason to demand that non-phenomenal sensation/perception be eliminated as well.

      "I suspect this ['perception is not itself a kind of belief'] is only true if dualism is correct."

      Why? I neither see how one could reduce perception to belief or abstract thought in practice, nor do I see any reason for someone who accepts physicalism and neuroscience to even take this hypothesis seriously. Treating taste and touch as kinds of belief seems to me as silly and unmotivated as picking some other sense-modality and claiming, say, that belief and taste and touch are all really just especially attenuated kinds of visual experience. That's not how we actually process information.

      The fact that we process information in multiple different ways -- some linguiform, some not -- doesn't itself endanger physicalism in any way. Digestion and visual perception are two different biological processes, neither one reducible to belief or thought, but both nonetheless completely describable *via* beliefs and thoughts. Again, beliefs and thoughts can describe things that aren't themselves either a belief or a thought; being non-propositional doesn't make you ineffable or mysterious or unphysical. A cactus, like a visual experience, is not in reality a sequence of sentences; yet we can still exhaustively describe cacti and visual experiences (again, assuming that vision is devoid of actual qualia) in the language of science.

      Delete
    22. "if I could make you believe you were experiencing a familiar quale such as the colour blue, then you would genuinely be experiencing that quale."

      This suggests that I can't be mistaken about the character of my experience. But I'm mistaken about it all the time -- mistaken before I even realize that I'm mistaken, not just mistaken-in-memory. This is why things like optical illusions are possible; optical illusions aren't just cases where certain arrangements of objects reliably lead me to false beliefs, they're cases where certain arrangements of objects reliably lead me to pictorially misrepresent my environment in my visual field. Such illusions can persist whether or not this visual misrepresentation results in any false *beliefs* about anything; if I've studied the illusion enough my beliefs about what's going on in my head and in my environment will be completely accurate, and yet clearly there's still some sort of persistent misrepresentation going on. That's possible because our brain has representational systems that aren't themselves beliefs or judgments.

      "I suggest hypnosis as one example where this might actually happen." - It's hard to interpret data from hypnosis. Suppose I suggest to you that you'll taste strawberries when I ring a bell. On your theory, my belief that I'll taste strawberries is so strong that it results in a vivid 'taste-style' belief. But on my theory, I simply say that vivid gustatory hallucinations that aren't themselves beliefs can be triggered by top-down linguistic suggestions like commands. Moreover, it's also open to both of us to say that the hypnotic subject isn't really tasting strawberries at all; maybe they just have a persistent false belief to the effect that they're having a certain gustatory hallucination (or, on your theory, a persistent false non-taste-style belief to the effect that they're having a certain taste-style belief).

      Incidentally, to follow up on Tom's suggestion that 'seemings' are what's of moral consequence: As an eliminativist, I'm tempted to say that it is certain physical brain-states, which may not even be epistemic or 'seeming'-involving, that are the proper objects of our concern. I agree that the moral and prudential implications of eliminativism are some of the most important and troubling things to fall out of the view.

      Delete
    23. @Robby
      >I don't understand what you mean. Which Chalmers argument are you alluding to?<

      I found the following quote here (http://consc.net/papers/moving.html#2.3):

      "I was attracted to type-B materialism for many years myself, until I came to the conclusion that it simply cannot work. The basic reason for this is simple. Physical theories are ultimately specified in terms of structure and dynamics: they are cast in terms of basic physical structures, and principles specifying how these structures change over time. Structure and dynamics at a low level can combine in all sort of interesting ways to explain the structure and function of high-level systems; but still, structure and function only ever adds up to more structure and function. In most domains, this is quite enough, as we have seen, as structure and function are all that need to be explained. But when it comes to consciousness, something other than structure and function needs to be accounted for. To get there, an explanation needs a further ingredient."

      >Type-B materialists by definition deny that zombies are possible. They just allow that they're conceivable<

      Fair enough. I was being unhelpfully loose with language. When I said "any formulation of Type-B materialism which allows the existence of philosophical zombies" I meant to say "any formulation of Type-B materialism which allows the conceivability of the existence of philosophical zombies".

      >and try to give arguments for why conceivability and possibility come unstuck<

      This clears it up a bit for me. If they deny that zombies are logically possible, then I'm on board with them, although I think there is some equivocation about the meaning of "conceivability" going on. This view is more fully explained in my earlier comment on this page:
      http://rationallyspeaking.blogspot.co.uk/2013/03/i-just-dont-get-it-one-more-on-mary-and.html?showComment=1364324206217#c7530832974414652937

      Delete
    24. @Robby

      > If you're suggesting that it's impossible for someone to find the zombie argument initially plausible without already being a dualist, then you are simply mistaken. <

      Sorry, that's not at all what I was trying to say. I do regard it as an interesting thought experiment, and it would not be so if it were so obviously circular. Instead, I'm arguing that it is subtly circular.

      In light of Chalmers' argument, and given sufficient consideration, philosophical zombies are only really conceivable if dualism is true. Therefore you can't conclude that dualism is true simply because philosophical zombies are conceivable. This is what I meant to say when I asserted that only a dualist would allow this premise.

      Delete
    25. @Robby

      >"I suggest that the answer is that you can't derive something ineffable from something effable."

      But how would a completely effable microphysics yield an ineffable higher-level phenomenon? Can you give an example of how that could come about?<

      I think you've lost track of the conversation here a little bit. I'm denying precisely what you're asking me to demonstrate.

      Delete
    26. @Robby
      >If you provided a complicated theory that explained people's belief formation about the Two Tables as a spontaneous but consistent error in judgment that has nothing to do with a misrepresentation in their visual field itself, then your theory would be false, or at best incomplete.<

      An interesting point, however I think it doesn't work if we remind ourselves that there is no homunculus sitting in a Cartesian theatre examining the visual field. I agree with Dan Dennett's view in Consciousness Explained, that there is no "visual field" which is represented in consciousness. Sure, the whole image is represented in various unconscious parts of the mind, but its conscious representation is not an image but a set of beliefs about what we are seeing. This view is borne out by studies of people with certain types of brain damage or users of hallucinogens.

      For example, ff you have a hallucinatory experience that you are talking to a giant magic frog, I don't believe your brain renders an exquisitely detailed Industrial Light & Magic special-effects creature. Rather, you simply have a belief that you see a frog. If you need more detail, for example, you look to see what colour its eyes are, then your brain simply makes up an answer, and now you have an additional belief.

      Your example of the optical illusion can be understood in this way too. While the "correct" image of the visual field is somewhere represented in your brain, your conscious self has no direct access to it. Instead, it perceives it indirectly through a large set of beliefs. One of those beliefs is that one table is longer than the other.

      I think this resolves the problem you identified with my approach.

      Delete
    27. @Robby

      Reading on to some of your later comments, it may be more helpful if I replaced "perceptions are beliefs" with "perceptions are judgements", as you seem to be interpreting "belief" in terms of linguiform data only (and perhaps justifiably so).

      You can take it that my view is essentially the same as Dennett's, so if it seems otherwise then I am probably not expressing myself clearly.

      I do, for example, acknowledge that there is a vast difference between believing a ball is red and perceiving that it is red. However I might be inclined to characterise the latter as having a belief that I am perceiving the ball as red, since literally perceiving redness might be construed as admitting that ineffable qualia really exist.

      And again, if we adopt Dennett's view that the visual representation directly available to the conscious mind is not a literal field of vision made of brain pixels but a series of sensory propositions such as "this is a circular shape", "this bit is red", "this shape is longer than that shape", then you can regard those sensory propositions as the beliefs I was talking about.

      Delete
    28. [Deleted and re-posted to fix a typo]

      I fear we will have a hard time coming to an understanding. I couldn't read past the first premise of your argument without becoming concerned.

      "1. If a law can only be expressed by making explicit reference to mental things qua mental, then the law is not physical."

      I simply don't feel compelled to accept this premise. The expression of the law is not the law itself. Why should our inability to express a law in physically reduced terms imply that it is non-physical?

      "Color-blind physicists can do good physics. Even a robot could, I'd argue, do good physics."

      I was being imprecise; what I meant is that the problems of qualia seem to accompany the fact of having experience, and the fact of having experience is a precondition for doing physics, as I understand it.

      I think this is true even for a robot. I agree that, hypothetically, a robot could do physics. But to do physics -- to ask a question about the physical world, design a good experiment to answer it, to execute that experiment, and to interpret the results within a theoretical paradigm -- the robot would have to be very, very sophisticated -- so sophisticated that I think we could make a persuasive case that it also experiences qualia. (Or at least as persuasive a case as I can make that anyone other than myself experiences qualia.)

      I see how you might read this argument as having a phenomenological character but I'm not really coming at it from that direction. I'm thinking Hume, not Husserl. I simply have a very concrete sense that knowledge is generated by locally embodied beings with specialized perceptual equipment through which they experience the world. This seems to me to be a straightforward fact about all the science that has ever been done, as far as any of us knows. I don't know what "science" would mean if we took the human (or human-like) beings out of it. We may simply have to agree to disagree on that basis.

      Delete
    29. @Robby
      >a world where everyone has extremely complex and skillful blindsight with respect to all of their senses.<

      That's pretty interesting. Not having experienced blindsight, I find it hard to answer. Blindsight is relatively inconceivable to me as it is - nevertheless I recognise that it is a real phenomenon.

      I might surmise that perhaps having an extremely accurate sense of blindsight is essentially the same as normal sight. Perhaps blindsight is necessarily impoverished compared to normal sight.

      But that's just one way of approaching this interesting thought experiment. There may be others. Perhaps skillful and accurate blindsighted people have had their percepts disrupted such that they no longer recognise them as familiar qualia. Perhaps if they had been born in that condition they would refer to their "hunch" percepts as qualia and ask much the same questions about qualia as we do.

      Another approach might be that blindsight is damage to the part of the brain that makes sensory "beliefs" more immediate and emotive than regular beliefs. They still have access to the all the same information, but they simply don't recognise it as sensory, in the same way that brain damage sufferers may refuse to believe that their left arm is their own. In this view, what makes something a quale rather than a dry fact is a mental marker or circuit that leads the brain to treat it with greater significance.

      Delete
    30. Robby, I see that I missed the second half of your response to me! (I failed to read past the break.) I think most of your later argument is summed up in the assertion that I "need to explain why we should accept mysterianism." And indeed, I can't give a positive justification for "mysterianism" (a term that I hate but will acknowledge for our purposes), but I can briefly say that given the tripartite choice between dualism, eliminativism, and mysterianism, I think the third accords best with the sum total of our experience. Dualism is problematic for me because it posits the existence of a second kind of thing without offering any evidence that is subject to third-person verification. Eliminativism is problematic for me because it entails affirming that some parts of our experience are veridical, while denying that other parts are veridical in a way that seems very ad-hoc to me. Should we deny the veridicality of some parts of our experience simply because they don't suit our (hubristic!) assumptions about what can be known?

      Mysterianism, on the other hand, doesn't ask us to deny the veridicality of parts of our experience. It makes a more modest claim than dualism. The claim that it does make has a chance of receiving a priori support, insofar as we have had success reasoning about the limits of knowledge in other areas of study--most notably in mathematics. And if those limits are anything like the limits of mathematical formalism, they aren't likely to be of any great practical concern. To me, this really does seem like the most parsimonious response to the questions raised by qualia.

      One last note: you object that "discovering that experiential states are fundamental to nature would completely upend our notion of what nature is, even though we relied on those experiential states to come to understand nature in the first place." But as far as I understand the ramifications of my own claims, I don't believe that I am suggesting that experiential states are fundamental to nature. I am only suggesting that they are fundamental to knowledge, and in a way that renders some aspects of them impervious to knowledge.

      Delete
  14. Excellent post, Massimo, especially on the category mistake issue. It's amazing how many philosophers still make this mistake on thought experiments of this type.

    ReplyDelete
  15. @Massimo,
    "That is because at best Chalmers has shown that there is no logical contradiction between talking and interacting like a human while at the same time not having conscious experiences."
    He hasn't even done that since you can't interact at all without being at least aware of what it is you're reactively interacting to. Logic doesn't exist or need to in a vacuum.

    ReplyDelete
    Replies
    1. >He hasn't even done that since you can't interact at all without being at least aware of what it is you're reactively interacting to.<
      While I agree with your conclusion that Chalmers has not shown that it is logically possible, I disagree with your reasoning. Simple automated systems can interact with their environments, yet few people regard them as being conscious.

      Delete
    2. Simple automated systems are not "interacting like a human" unless they are at least aware of what the human that built them has programmed them to be able to observe and consequently react to. In the case of the conceptualized zombies they haven't been built by humans, they're imaginarily a form of human that has somehow had its awareness unbuilt.
      Humans in short are able to be conscious of the state of their awareness. Machines aren't.

      Delete
    3. >Simple automated systems are not "interacting like a human"<

      Ah yes, but you said "can't interact at all". If you stayed with Massimo's wording and instead said "you can't interact like a human without being aware" then I might have agreed with you.

      Except that it's still more of a flat assertion than an argument. How do you know that a machine couldn't interact like a human without being aware? I agree with you, but do you have an argument to back this up?

      Delete
    4. DM, I actually said, "you can't interact at all without being at least aware of what it is you're reactively interacting to." What's my argument to back that up? My hypothetical assumption that all reactive systems are anticipatorily aware. I'm aware of no effective counter argument that awareness is not a factor in determining reactive strategies - human or otherwise.

      Delete
    5. But simple mechanical systems are reactive, and not aware. As I said before.

      If by reactive you specifically mean conscious intelligence, then of course you need to have awareness to be reactive, based on how you define it.

      Delete
    6. Simple reactive systems are aware in an anticipatory sense but not consciously aware. You should look into the study of anticipatory systems before you decide that mechanical operations have no strategic components that determine their appropriate responses.

      Delete
    7. This discussion is about awareness in the sense of consciousness, however, so whatever limited behavioural awareness exhibited by simple reactive systems is a different kettle of fish.

      Delete
    8. DM,
      You're the one that said, "Simple automated systems can interact with their environments, yet few people regard them as being conscious." And then you said, "But simple mechanical systems are reactive, and not aware. As I said before."
      So now you want to move that mechanical fish to a different kettle because it turns out to have limited awareness after all.
      Do you recall asking me, "How do you know that a machine couldn't interact like a human without being aware?"
      Rather than attempt to move the fish, you might at least try to tell me why you didn't like the answer.

      Delete
    9. @Baron P
      You're equivocating about the meaning of "aware". Nobody, not even Chalmers, would deny that reactive systems have trivial non-conscious awareness in the same sense that a mousetrap is "aware" of a mouse or a thermostat is "aware" of the temperature.

      The point you initially made only establishes the presence of this trivial sense of awareness. This is irrelevant as nobody doubts it. What is under discussion is consciousness, and your point does nothing to address this.

      >Do you recall asking me, "How do you know that a machine couldn't interact like a human without being aware?"<

      Yes. And you haven't answered this, since I'm talking about consciousness not the "trivial awareness" you have established.

      Delete
    10. DM, I'm afraid the equivocating is all on your side (or your shifting sides) of the argument. Trivial awareness? Not only did you deny it and are now "equivocating" that you did, you don't even understand what anticipatory forms of awareness (as well as anticipatory forms of consciousness) consist of. A mousetrap is not aware of a mouse, now is a thermostat aware of temperature.
      What they are aware of is the necessity to react according to their strategically constructed purposes. The mousetrap anticipates the forces that will make it snap. (I'll let you figure out, if you can, what the thermometer was made to anticipate.)
      So your argumentative tactic as usual has been to "trivialize" the question.
      The point of my asking if you remembered asking about being aware is that you were asking about an awareness that you now claim was too trivial to be asked about. And you are still doing your trivial best to avoid dealing with my answer.
      Not just laughable but pitiful.

      Delete
    11. This conversation is impossible because you introduced the use of the term "aware" to refer to unconscious responses to mechanical stimuli.

      If we're going to continue, then let's call the two ideas "responsiveness" and "consciousness", because I can't make sense of what you're saying right now.

      With this terminology, David Chalmers has made an argument about consciousness, and your initial statement becomes:

      "since you can't interact at all without being at least being responsive to what it is you're reactively interacting to."

      My question to you becomes "how do you know that a machine couldn't act like a human without being conscious?"

      I hope that clarifies what I'm saying to you.

      Delete
    12. No, you haven't cleared the question up, you've simply changed it. And responsiveness is not a lesser aspect of consciousness, it's a blanket term for both consciousness and the awareness that can be less than conscious.
      I know that sticking with the dictionary definitions of a word makes argumentation more rigorous, but isn't rigor what we're striving for when we attempt to call this stuff philosophy? (This entire thread offering one of the worst examples of philosophizing ever, but I digress as usual.)
      But just for fun, let's look at the newest version of your question. To which my answer would be, I know that a machine couldn't act like a human without being conscious because a human couldn't act like a conscious human without being conscious. But a machine could conceivably act like an unconscious human, although I'm hard pressed to prove that by example.

      Delete
    13. @Baron P
      I'm not trying to be difficult. I just think it's unhelpful and confusing to call basic response to stimulus "awareness", because that's not what Chalmers was talking about. In fact what you quoted from Massimo was explicitly about consciousness and not this type of awareness.

      So it seems we agree that consciousness is not required for interaction but I also agree with you that consciousness is required for human-like behaviour. However, I don't think you have really addressed Chalmers' argument. He has proposed the idea of a human that behaves normally but unconsciously. Your counterargument is simply that this is impossible. My only problem with you then is that that's not really an argument, it's an assertion (which I happen to agree with).

      Delete
    14. DM, You misquoted me again to make a rather pointless point.
      I wrote, "He hasn't even done that since you can't interact at all without being at least aware of what it is you're reactively interacting to. Logic doesn't exist or need to in a vacuum."
      The last part is the statement of an obvious fact. And the further fact that you agreed with the initial assertion shows that asserting the obvious is an effective way to argue. Especially when it's doubly self evident.

      Delete
    15. I didn't quote you at all that time so I honestly don't know what you're talking about.

      I have no idea which part of that you regard as an obvious fact which is relevant to Chalmers' argument. Depending on which part of your quote you're referring to and how you interpret it, either it's trivial and irrelevant or relevant and not self evident at all.

      Delete
    16. DM, look here:
      Disagreeable MeMarch 28, 2013 1:50 PM
      "since you can't interact at all without being at least being responsive to what it is you're reactively interacting to."

      That was a quote, no?

      And then this is the latest "quote" of myself re Chalmers:
      "He hasn't even done that since you can't interact at all without being at least aware of what it is you're reactively interacting to. Logic doesn't exist or need to in a vacuum."

      You've now asserted that it's either trivial and irrelevant or relevant and not self evident.
      And yet you also had said earlier that you agreed with it (although I expect you to reply that you can't find where you did). So for you it has been both a trivial assertion and an argument you agreed with. Pretty much par for your course.
      I'd suggest that you need help except someone might call that an ad hominem.

      Delete
    17. Ok, that quote was a couple of comments back so I wasn't sure which one you were talking about. Also, it wasn't a misquote, because I clearly explained that I was replacing terms in order to clarify my understanding of your meaning.

      What I agree with is the consciousness interpretation, which is relevant but not self-evident. Something doesn't have to be self-evident to be correct, but if it isn't self-evident it needs an argument to back it up.

      We seem to be simply incapable of understanding each other. You can have the last word if you wish.

      Delete
    18. If I write two sentences, one of which explains the other, and you leave the explanation out so that you can quote the first sentence as only an assertion, that's a form of equivocation, to put it kindly. And it's easy to claim the self-evident is not evidence if it doesn't fit your prior understanding and you don't want it to. And if you weren't capable of at least thinking that you understood what I wrote, there'd be no need to tinker with it before you make that "misunderstanding" claim, now would there? And I can assure you that I understand everything you wrote, as inconsistent and logically simplistic as it always is.

      Delete
  16. Oh goodness. If zombies exist, that would mean that I (who is not a zombie. I guarantee you this)am writing this not because this is my opinion, but because there's a part of me that talks nonsense, and yet by some mysterious mechanism, this nonsense is exactly the same as the opinion I hold. When I say I'm conscious, I'm not saying that because I am conscious and know that I am conscious, but because of this crazy part of me that nonsensically talks about a consciousness about which it has no way to know.

    ReplyDelete
  17. Connecting philosophy and science

    Philosophy is still searching for truth and science for the absolute. Someday they both will be united it.

    If I may suggest: head down to the river and study nature, it worked for Michelangelo.
    And simplify, it worked well for Einstein and me.

    The problem with science or measure is the uncertainty, the solution as is nature is measureless, is absolute, is true.

    Truth unites us All,
    Truth is

    =

    ReplyDelete
  18. Robby:

    "I just deny that this perceptual distinction is as rich as it epistemically seems to be. I deny that an agent who hasn't experienced red is missing out on anything, other than missing out on undergoing a certain illusion."

    The richness of what you say is the illusion of experience is equivalent to the informational discriminations made by the neural correlates of that illusion. It matters tremendously to us that we not miss out on the variety and subtlety of the illusion of sensory and emotional experience. So again, the role played by the illusion of experience seems exactly that played by experience. So I'm tempted to say that the illusion of qualitative richness *is* experience; after all it's qualitative in character. If so, then we don't have to be eliminativitists about phenomenal properties, but can naturalize them as a kind of representational judo - perhaps.

    ReplyDelete
    Replies
    1. I think by 'rich' you mean vibrant, detailed, etc. Such richness can be fully explicated in functional terms, hence poses no problem for physicalism. But when I said 'rich' I had in mind the phenomenal's apparent autonomy, its arbitrariness, its overflowing of any particular functional role. (As in, for instance, the Inverted Qualia intuition and the Zombie hunch.) I accept dualists' claims that we do have persuasive intuitions of these sort, and I even accept that these intuitions stem from genuine meta-representations we possess of our own experiences. But given that our experience is represented as 'rich' in the sense of fully autonomous, this representation must either be veridical (in which case we actually instantiate irreducible qualia) or nonveridical (in which case we don't instantiate such qualia).

      As a physicalist, one's only options are to either deny that such representations are veridical (which means endorsing my brand of eliminativism), or to deny that our experience motivates Inverted Qualia or Zombie hunches in the first place (i.e., to deny that experience even seems to represent Chalmers-style qualia). When I myself introspect I find the latter an inaccurate description of my phenomenology, and a great many people, both physicalist and anti-physicalist, seem to feel similarly; they don't think the arguments for dualism are popping out of thin air. If others do feel that their phenomenology inclines them more toward 'we don't self-represent as having qualia' or 'we don't epistemically seem to have qualia', should they conclude that I and all of these dualists and physicalists are just stuck in a flight of fancy with no experiential basis, or should they conclude that they perhaps just haven't attended carefully enough to the arguments and/or to their own phenomenology?

      Delete

Note: Only a member of this blog may post a comment.