Thursday, May 11, 2006

Global Supervenience and Physicalism

One class of facts supervene on another if the former couldn't differ without a change in the latter. I've previously talked about this a lot in connection with morality, for instance: you couldn't have a world that was identical to our own in all respects except that Hitler was a good person. Because the moral facts supervene on the non-moral, the only way to change the former would be also to change the latter. (If Hitler had stayed out of politics and instead become a harmless hermit, then perhaps he would not have been so evil.)

Now, it seems to me that mostly all facts supervene on the physical. (The main exceptions being facts about phenomenal consciousness, and anything else which in turn supervenes upon those.) Indeed, this seems kind of obvious. But sometimes people object to such claims by adopting an overly restrictive view of the physical. For example, they might consider two objects which share all their intrinsic physical properties and yet differ in the supposedly supervening property X. But that merely demonstrates a failure of "local supervenience". X doesn't supervene on this restricted cluster of "local" physical facts. But that does nothing to refute the claim that X supervenes on the totality of physical facts. You can't just ignore all the relational physical properties and pretend you have a counterexample. Not that some people don't try.

A recent example is provided by the Maverick Philosopher, who writes:
But the reference is indeterminate if we go by the physical properties of the representations alone. Suppose I hand you two drawings of your mother, one an exact copy of the other, but you do not know which is the orig[i]nal and which is the copy. You cannot, by inspection of these drawings, tell which is which. Thus you cannot determine the reference from the physical properties.

But you cannot, by inspection alone, tell all of the physical properties of the drawings. You cannot tell, for instance, at what time each was drawn. But that's clearly a physical fact about the drawings. And clearly if someone knew all the physical facts, in this appropriately broad sense, they would no longer have any difficulty determining which drawing was the copy.

So BV has done nothing at all to show "that materialism doesn't work" (where "materialism" is synonymous with "physicalism", i.e. the claim that all facts supervene on the physical facts). He's shown that intentionality doesn't locally supervene on some restricted class of physical facts. But that shouldn't be surprising. Of course if we overlook some physical facts, then we won't be able to derive some other truths (say, about what a drawing refers to). That's because you're ignoring some crucial facts! The appropriate response is to open your eyes, not reject physicalism. (Actually, I think we should reject physicalism for other reasons. But this sorry argument surely isn't one of them.)

This confusion can also arise in discussions of scientific reductionism, and whether the biological supervenes on the physical. You couldn't have a world physically identical to our own in all respects, but somehow biologically different. There couldn't be any differences in species, for example, or in which organisms are living or dead, without some physical difference in the world. One might suggest that biological kinds aren't determined by the laws of physics alone; they also depend on physical history. But of course physical history is included in the class of physical facts, so we never should have excluded that in the first place. The thing to consider here is the entire 4-dimensional space-time "loaf" (encompassing past, present, and future), not just little bits of it.

Incidentally, this is a nice chance to talk about misusing the "vitalism" analogy. Recall our past discussion of how Kripkean examples of conceivable impossibities (e.g. Hesperus' being non-identical to Phosphorus) do nothing to support the impossibility of zombies. Another bad analogy often invoked in defence of materialism is that of vitalism: "People used to think that being alive was something beyond the physical, but science later proved them wrong. Why think consciousness is any different?"

But 'life' can clearly be analysed in functional and structural terms. There is no sense to be given to the notion of something that is functionally and structurally indiscernible from a duck, having all the same kinds of relations to other objects as another duck does, and yet somehow fails to really be a living duck. To be a living duck just is to have the right kinds of functional relations and so forth. There's nothing more to it than that. Not even vitalists disagreed, as Chalmers explains:
Vitalists typically accepted, implicitly or explicitly, that the biological functions in question were what needed explaining. Their vitalism arose because they thought that the functions (adaptation, growth, reproduction, and so on) would not be physically explained. So this is quite different from the case of consciousness. The disanalogy is very clear in the case of Broad. Broad was a vitalist about life, holding that the functions would require a non-mechanical explanation. But at the same time, he held that in the case of life, unlike the case of consciousness, the only evidence we have for the phenomenon is behavioral, and that "being alive" means exhibiting certain sorts of behavior. Other vitalists were less explicit, but very few of them held that something more than the functions needed explaining (except consciousness itself, in some cases). If a vitalist had held this, the obvious reply would have been that there is no reason to believe in such an explanandum. So there is no analogy here.

Now, we do indeed have every reason to believe that physical science can provide adequate functional explanations. But consciousness is different in that it cannot be analysed in functional terms. That's what zombies show us, for example: we can imagine creatures functionally indiscernable from ourselves, that nevertheless lacks conscious experience. Whatever consciousness is, it isn't just the having of certain functional relations. That's not what we mean by the term - our phenomenal concepts are something quite different.

One might more plausibly hold that our brain states "give rise to" our conscious states. But this is already to recognize that they are two different things, even if a (natural) lawful connection holds between them. The natural laws might have been different, after all. And in that case there might have been the same physical stuff but without giving rise to any phenomenal stuff. So that's why the phenomenal isn't reducible to the physical. (At most it supervenes nomologically, not metaphysically.)


  1. > You cannot tell, for instance, at what time each was drawn. But that's clearly a physical fact about the drawings.

    how does this relate to the previous post about endurance? Is when it was made really a physical property of the drawing or is it a physical property of the drawing IN THE PAST.
    does a drawing carry with it all the things that ever happened to it in some sense?

  2. I've been hounded by zombies for a few months now, and I'm coming to the conclusion that Dennett is correct about them.

    All we know as individuals is our experience of things, i.e., the phenomenal. Reasoning beyond this seems futile. If we are to say that the "experience of X" is not synonymous with "X," then we are led to useless absurdities like solipsism. For there is no conscious experience that can tell me whether or not "X=>Y" follows from my "experience of X" implying my "experience of Y." If "X" is not my "experience of X," then neither is "implication" necessarily my "experience of implication."

    Humans are used to referring to "X" and "Y" instead of
    "experience of X" and "experience of Y," but this is ultimately a notational shorthand. We can safely use this shorthand as long as we don't use it to analyze the nature of experience itself. If we're considering such questions, we can't neglect the fact that our inputs are always experiences of things and not the things themselves.

    Dennett sees that, for example, "experience of understanding" is all that we have access to, and that to speak of "understanding" as a distinct other is a sort of notational confusion.

    More broadly, we can't say that we "know X" or that "X exists" independently of experience because these expressions rely on "experience of knowledge" and "experience of existence" respectively. For example, existence of X isn't just referring to an alien X beyond experience of X, but to an alien kind of knowledge, too.

  3. Doctor logic, what you are describing seems to me a lot like Berkeleyian idealism. I'm pretty sure that materialist Dennett would not endorse your description of his position. His position, to the best of my understanding, is that talk of a real "X" is always of a physical object, and that talk of "the experience of X" can be paraphrased as talking of how X seems to me, where "seems" is cashed out in a set of beliefs and judgements analyzable in functionalistic terms. There is no such real, existing object such as a "phenomenal experience" (with qualia as its properties), so the question of identifying "X" and "my experience of X" does not arise.

    Richard, I wonder how you would reply to the argument I gave that the difference between nomological and metaphysical supervenience, and arguments form conceptual analysis, are irrelevant to scientific questions. You say that the zombie argument shows that "That's not what we mean by the term - our phenomenal concepts are something quite different." But that reaction to the zombie argument is not universal; I seem to remember Chalmers reports that one third of people he questioned in informal polls rejected it. Perhaps "what we mean by the term" is in state of flux and as more people are convinced by materialism what we understand by "phenomenal concept" will shift in meaning to something definable functionally along Dennettian or other lines.

  4. Doctor Logic - that's some extreme empiricist assumptions you've got going on there. Seems pretty misguided to me. We can sensibly speak of a mind-independent external world. If some of your assumptions lead you to reject this fact, so much the worse for those assumptions, I say.

    Alejandro - shifts in language won't alter the nature of reality itself. Sure, we could simply refuse to speak of (what I call) phenomenal consciousness. We might redefine the term to mean something purely functional. But that wouldn't change the fact that (what I call) phenomenal consciousness both (i) exists, and (ii) fails to metaphysically supervene on the physical, and hence that (what I call) zombies are metaphysically possible.

    I'm in no position to speak to the scientific relevance of all this, but you might have a look at Chalmers' papers on the science of consciousness. (Intuitively, it would seem important for scientists to recognize that there is a non-functional, phenomenal component to reality. Otherwise they aren't explaining all that stands in need of explanation.)

  5. G. - I don't think the earlier post has any relevance to this one. Eternalists can treat the drawings as temporally extended, "perduring" objects, and their earlier temporal parts will (timelessly) have the properties in question.

  6. OK a different tack,

    If I "Scottie" beamed you "Dr Spock" up to the deck would the "Dr Spock" that appears to be the "dr spock" that was decomposed on the planet not really be him?

    It seems to me that since you share almost no atoms in common with yourself 10 years ago you face the same question.

    My sugestion is that it is generally speaking the continuality of pattern that matters.

    If there were two identical objects that continuality would be ambiguous. I.e. the new spock remembers being the old one just like you remember being at intermediate school (I guess) and your identical painting has all the evidence in it to suggest it was made identically to the other painting (it must have some history evident in it even if it is in this example the "wrong" one).

    * and Ill bite the bullet if that means there can theoretically be two me existing at the same time - although I propose any duplication would destroy the first you via quantum mechanics.

  7. I think you are missing my point, Richard. Of course reality doesn't change when our concepts do. But notions like "whether this bit of reality is conceptually different and only nomologically related to this other one, or whether they are really the same thing" can be dependent on our conceptualization. As I discussed in my post, science does work by identifying fundamental and conceptually distinct aspects of reality and relating them with contigent laws. The necessary/contingent, conceptual/nomological distinction is irrelevant for science (if seen as fundamental and used to draw metaphysical conclusions; it may work as an approximation for the language of science at a given time). Mass and energy were conceptually different, and now are not; likewise with gravity and geometry. Consciousness and functional information processing may follow the same way. A purely conceptual a priori argument cannot establish that they will not.

    I have read several of Chalmers' papers, and also his book. I like them, but I don't find them convincing.

  8. Ooops! When I said "science does work by identifying..." I meant "does not work..."

  9. Oh, it's certainly true that distinct concepts may turn out to coincide ('water' and 'H2O' is an obvious example -- distinct concepts, same stuff). Presumably this is because the concepts in question do not fully specify their object, but merely some or other functional role, and so some work is left for "the world" to do. Science can discover, for example, that the same physical property plays the mass role as does the energy role (or some such, I don't know the scientific details).

    But there's no analogy to phenomenal consciousness, because this rare concept is not underspecified. No part of its identity conditions are left open with some placeholding functional account of "the consciousness role" to be filled in by empirical science. Our grasp of the concept is sufficient to grasp the identity of that to which it refers. So there's not the same conceptual room for empirical revision here.

  10. alejandro,

    You are right. I should not have cast Dennett's conclusion as one stemming from assumptions similar to my own. (Not that I'm fully persuaded that my agreement with him is purely coincidental either... :) )


    We can sensibly speak of a mind-independent external world. If some of your assumptions lead you to reject this fact, so much the worse for those assumptions, I say.

    I think you dismiss my claim too quickly.

    Can we sensibly speak about a mind-independent world? What are the properties of such a world, and how are any of those properties violated by my assumptions? Surely, the concept of a mind-independent world derives from observations that things appear to persist even when we are unaware of them, or that things appear to exist when the minds of others are unaware of them. Nothing I have said rejects or denigrates any of these phenomena. What seems unjustified to me is the assumption that the symbol "mind-independent" now means something more than the phenomena it was invented to represent.

    Again, I'm not rejecting the validity of the sciences or anything commonly associated with phenomena of mind-independence. I'm just trying to be careful with my definitions. And, as I say, this is unimportant except when considering the nature of experience itself, e.g., when thinking about zombies.

  11. "No part of its identity conditions are left open with some placeholding functional account of "the consciousness role" to be filled in by empirical science. Our grasp of the concept is sufficient to grasp the identity of that to which it refers. So there's not the same conceptual room for empirical revision here."

    Well, I for myself don't see my concept of consciousness as unrevisable in this way. I think that to assume it is effectively begs the case against materialism. No other concept used in science has this kind of strange "transparency".

    I think there may be a shift from the indubutability of some first-person facts (such as "I am now feeling pain/seeing green") to the "transparency" of the concepts used to describe them. I agree that in some cases, the words "I am seeing green" express a fact as certain as something can be. But it seems to me that you are assuming that these words are more or less equivalent to "I am having an phenomenal experience which has greeness as an intrinsic, nonpysical, non-functional property", and then try to move from the indubitability of the fact to the unrevisability of the language in which you have expressed it. I would say that the very same fact we started with can probably be reexpressed in functional terms, and that insofar our ordinary language does make your paraphrase more or less synonymous with "I am seeing green", then that language will be open to scientific revision if we decide that intrinsic nonphysical properties are a myth.

    By the way, my examples from physics are not exactly equivalent to the water/H2O one. The concepts of water and H2O are clearly different (at least in their "primary intension") and there is an a posteriori identity between their referents. My examples attempted to undermine the distinction between conceptual/a priori and nomological/a posteriori dependence altogether. Modern physics does not state clearly whether mass is identical to energy, or whether they are different things with a strict nomological relation betweeen them. And it doesn't need to draw such a clear-cut distinction. I don't see the need for it in the case of consciousness either.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)