Philosophy, you understand, is a very pharmacopoeia of cures that are worse
than the corresponding diseases. This started a long while ago; perhaps with
Plato’s suggestion that, although there is a problem about how so many
different things can all be chairs, philosophy can fix it: there is only one
chair that is really a chair, the Chair on which no one can sit; the One
Chair that is in Heaven. This kind of philosophical overkill, having once
got started, has never stopped. Thus Descartes: the way to explain how your
mind causes your body to move is to say that the pineal gland performs a
miracle each time it does. Or Berkeley: the way to avoid scepticism about
perceptual beliefs is to say that chairs, tables and everything else are
made of ideas. Or Wittgenstein and Ryle: the solution of the epistemological
problem about how anybody can know whether anybody else is in pain is that
(other people’s) pains reduce to their pain behaviours, there being, by
assumption, no epistemological problem about recognizing them. Or take
Carnap and Ayer: the way to understand the semantics of “electron” and other
such “theoretical terms” is to hold that electrons are “logical
constructions” out of the pointer readings of experimental instruments. Or
take Frege: given that Venus and the Morning Star are the very same thing,
there’s this worry about how John can believe that he sees Venus while not
believing that he sees the Morning Star. One avoids the worry by saying
that, though the two expressions refer to the same thing in sentences like
“John saw Venus” (the Morning Star), they do not refer to the same thing in
sentences like “John believes (thinks/knows) that he saw Venus” and “John
believes (thinks/knows) that he saw the Morning Star”.
Now there is externalism.
Externalism was invented by Hilary Putnam to explain certain semantic
intuitions to which he called philosophers’ attention. Here is a version of
the classic example: Imagine two glasses, each of which is filled with what
is, as far as we can see, the same sort of stuff, but only one of which is
filled with H2O. (The stuff in the other glass has some different chemical
composition, called “XYZ” by convention.) Question: which of the glasses
contains water? Answer: Only the H2O glass does. XYZ isn’t water, however
much it may appear to be. That is so even if only a properly equipped
chemist could tell them apart. And it was so even in Chaucer’s time, when
nobody could have told them apart.
There is, in fact, some uncertainty about how many people share these Putnam
intuitions; but, for present purposes, let’s suppose they are the norm and
that second thoughts don’t make them go away. Then it is arguable – indeed,
it has now been vociferously argued for several decades — that they throw
doubt on a number of venerable philosophical theses. I’ll mention two before
I turn to Michael Tye, who is very much into the externalist line of work.
Consider, for one example, philosophical theories of meaning. Traditional
semantics often rested heavily on two assumptions: (1) the meaning of a word
is (part of) what you must learn if you are to learn the word; (2) the
meaning of a word is what determines its reference (its “extension”, to use
the usual jargon). Roughly, words apply to things that satisfy their
definitions. It's because the word “cat” means (approximately) domestic
feline, that “cat” applies to, and only to, domestic felines. If that’s not
the semantic theory that they taught you in grade school, you must have gone
to a better grade school than I did.
But now it seems, in light of the Putnam intuitions, that (1) or (2), or both,
must be false. For, on the one hand, the intuitions suggest that nothing
except H2O is or could be water, hence that being H2O is what determines
whether stuff is in the extension of “water”. But H2O isn’t any part of what
the word “water” means. Surely, you learnt the word “water” long before you
learnt any chemistry; and, as previously remarked, though Chaucer never did
learn about water’s being H2O, he was quite clear that water is what the
“shoures soote” of Aprille are shoures soote of. In short, it can’t both be
that the meaning of the word “water” determines what it refers to and that
the meaning of the word “water” is what you learnt when you learnt the word.
About seventy-four reasons for rejecting this line of argument will no doubt
have occurred to you (maybe “water” doesn’t mean to us what it meant to
Chaucer?). Never mind; Tye thinks that the argument is sound and, as we’ll
see, his thinking that is essential to the position that he develops in
Consciousness Revisited.
We’ll return to that presently. Meanwhile, let’s consider briefly some of the
troubles that the Putnam intuitions are supposed to raise for traditional
views in the philosophy of mind.
There’s a very natural way to psychologize the thesis that the meaning of a
word is what determines its extension: words express concepts; so if words
determine extensions, concepts do too. But concepts are what you think with;
the concept DOG is what you use to mentally represent dogs when you think,
for example, that dogs bite. On this view of concepts, they are
interestingly Janus-faced; “Fido is a dog” and “Fido falls under the concept
DOG” are equivalent in that each is true if and only if the other is. Still,
they look in different directions. The former looks “outward”; it attributes
dogness to something in the world, viz to Fido. The latter, by contrast,
looks “inward”; it says something about what goes on in a mind when the
thought that Fido is a dog occurs to it. A word expresses a corresponding
concept; a concept represents what the corresponding word refers to.
This representationalist view of concepts is manifestly in need of sharpening
and refinement; but I confess that I love it very much. It has kept me
company through many a night’s insomnia. It is, however, in jeopardy from
Putnam’s intuitions. For, on the one hand, it seems clear that though the
concepts WATER and H2O have the same extensions, they can’t be the same
concepts; in particular, they differ in (what philosophers call) their
“possession conditions”: you can’t have the concept H2O if you lack the
concept HYDROGEN (or the concept OYYGEN; or, come to think of it, the
concept 2). But you needn’t have any of those in order to have the concept
WATER. It would seem to follow that, just as the Putnam intuitions militate
against the semantical view that word meanings determine extensions, so too
they militate against the psychologistic view that concepts are mental
representations of extensions. Or, to put it in the way that Putnam did, in
the paper that started all this: if meanings, concepts and the like are
distinguished by their extensions, then meanings, concepts and the like
“ain’t in the head”; they ain’t what cognitive scientists call
“psychologically real”. What concepts you have and what the words you use
mean, depend not (or not only) on how things are in your head, but also on
how things are in the world external to your head. Hence the usage according
to which views of meaning and concepts that are grounded in the Putnam
intuitions are said to be “externalist”. It is perhaps less shocking to
claim, with Putnam, that your concepts aren’t in your head than to claim,
with Berkeley, that chairs are made of ideas; but it’s quite sufficiently
shocking to be getting on with.
The question thus arises whether externalism is true. That, however, isn’t the
present question; I’ve brushed against it only by way of providing some
background for Tye’s book. For, if I read Tye right (which I wouldn’t swear
to; I found him hard going) he thinks that it’s not just issues about
meaning and mental representation to which the Putnam intuitions are germane,
but also issues about consciousness and the nature of perception. This is
either a great step forward or a great step backwards, depending (of course)
on whether Tye is going in the right direction.
Here’s a way of thinking about perception that has been persistent in both
empiricist and rationalist philosophies: What happens in perception is that
one makes inferences from the phenomenal content of one’s perceptual
experiences to how things are in one’s locality. One has, for example,
auditory sensations as of woofs or of miaows, and one infers, in the first
case, that there is an ambient dog and, in the second, that there is an
ambient cat. In effect, perceptual processes infer from the content of one’s
experiences to perceptual beliefs about things-in-the world. First blush,
what could be more plausible? But the second blush is less sanguine. For one
thing, perception doesn’t feel very inferential. As Tye remarks, “it seems
natural to suppose that vision involves direct contact [my emphasis] with
external things in standard veridical cases”, and “direct” means, at a
minimum, not inferential. “When I perceive a tomato, for example, there is
no tomato-like sense impression that stands as an intermediary between the
tomato and me. Nor am I related to the tomato as I am to a deer when I see
its footprint in the snow. I do not experience the tomato by experiencing
something else [sic] over and above the tomato and its facing surface. I see
the facing surface of the tomato directly [sic].” (Tye has a lot of trouble
deciding what, exactly, it is that you see “directly” when you see a tomato;
is it the tomato? Or the surface of the tomato? Or the colour and shape of
the tomato? Or the colour and shape of the surface of the tomato? Or what?
If Tye means to be serious about perception being direct, he must hold that
“what is it that is directly perceived?” is a serious question. But he’s
unforthcoming about the answer.)
There is a lot going on here. For one thing Tye is surely wrong to suggest
that an indirect-perception theorist (one who says, in effect, that
experiences provide premisses for perceptual inferences) must also say that
one “experience[s] the tomato by experiencing something else [sic] over and
above the tomato and its facing surface”. In particular, it’s not part of
the indirect-perception story that, in order to infer from the experience to
the tomato, one must first infer to the experience from something else.
Experiences are themselves modes of awareness; one doesn’t infer them, one
just has them. Accordingly, the object of an experience is not an experience
but (as it might be) an ambient cat. How do I become aware that the phone is
ringing? Why, from the ringing. But how do I become aware of the ringing? I
don’t; I hear it (as one might say) “directly”. The mind is active in
perception; but it is passive in experience; which is to say that perception
is inferential but experience isn’t. There may be problems with inferential
theories of perception, but they don’t — at least, they needn’t — get into
the kind of loop that Tye claims to detect.
And, anyhow, what’s the alternative? Tye thinks, rightly I imagine, that
making sense of things-in-the-world being “directly” perceived requires
supposing that such things are “part of” the experiences from which
perceptual beliefs arise. So, as Tye reads externalism, it’s not just
semantic stuff like the contents of beliefs and concepts that is (partly)
constituted by the external surround; so too is the phenomenal content of
experience. When your perceptual belief that there is a passing cat is
veridical, the passing cat is part of the experience that gives rise to the
belief. So I take Tye’s suggestion to go. Well, I guess I understand, sort
of, how a passing cat could be what an experience is of. But I’m a monkey’s
uncle if I understand how a cat – not just a mental representation of a cat,
or an experience of a cat mind you, but the very cat itself – could be part
of an experience (or, mutatis mutandis, of the perceptual belief that the
experience engenders). I just don’t get it. I give up. I’m, like, off the
bus.
However, a confession: It struck me as I was writing this that Tye simply
couldn’t be saying what I was taking him to say, viz that his blending of
externalism with direct realism vindicates the thesis that
things-in-the-world (like, tables and chairs?) are constituents of (not just
causes of, but constituents of) one’s phenomenology. It struck me that
nobody could believe that. So I went and tried it out on a couple of
philosophy friends (who hadn’t, however, read Tye’s book) and they agreed
that nobody could believe what I was writing that Tye believes. Fair enough,
but then, what is one to make of such a passage as this: "An object’s
looking F . . . [isn’t] a matter of an object’s causing an experience which
represents simply that something is F [sic]. The experience one has of the
seen object is one into whose content the seen object itself enters"
(my emphasis). Notice that it is the phenomenal content of the experience,
not just what the experience is an experience of, into which the seen object
is said to enter.
Now, I’m kind of a Tarskian about meaning. I don’t do “radical
interpretation”. So, when someone writes “the experience one has of the seen
object is one into whose content the seen object itself enters” I suppose
that he is probably saying that the experience one has of the seen object is
one into whose content the seen object itself enters. Perhaps someone of a
more hermeneutical temperament than mine will correct this reading in next
week’s Letters page in the TLS, and I will then feel a perfect goose. For
now, however, I shall proceed on the assumption that I have got Tye more or
less right.
There is, to be sure a way of construing “the content of an experience” that
would make Tye’s sort of externalism untendentious; cf “Make sure that
seeing the Chrysler Building is part of your experience of New York”. That’s
harmless, to be sure; it just means: when you’re in New York make sure that
you see the Chrysler Building. But it’s no help to Tye, who is in pursuit of
a theory not of the object of experience but of the phenomenal content of
experience. Now, prima facie, the phenomenal content of an experience of the
Chrysler Building is not the building itself but the look of the Chrysler
Building from the point of view of the owner of the experience. God only
knows where the phenomenal content of an experience of the Chrysler Building
is; that is, in effect, the question to which Tye’s externalism purports to
be the answer. But surely there is no corresponding question where the
building itself is; it’s on 42nd at the intersection with Lexington; at
least it sure was the last time I looked.
Here’s another way to make much the same point. Inferentialist theories of
perception in psychology, like scepticism about perception in epistemology,
generally take for granted that the phenomenal content of an hallucinatory
experience of the Chrysler Building might be arbitrarily similar to the
phenomenal content of a veridical experience of the Chrysler Building. (In
effect, Descartes’s formulation of scepticism starts from the assumption
that the phenomenal content of a veridical experience of the Chrysler
Building might be identical to the phenomenal content of a dream of the
Chrysler Building.) But, of course, that couldn’t be true if the Chrysler
Building is a part of the former but not a part of the latter. So there’s a
smooth transition from externalism to the thesis (often called
“disjunctivism” these days; philosophers find such pretty names for things)
that a veridical perception and a misperception can’t be identical in their
experiential content.
Disjunctivism thus sets its face not only against sceptical epistemology
(which is no doubt its main stalking horse) but also against the theories
and practices of research psychologists, who routinely take it for granted
that explanations of veridical perception ought also to explain perceptual
constancies, perceptual illusions and the like. In effect perceptual
psychologists (except Gibsonians) hold that veridical perception and
misperception both require inferences from the content of experiences to the
content of beliefs, and that, in dreams, hallucinations and the like, the
content of an experience may be arbitrarily similar to that of an experience
that’s veridical; it’s not the content but the inferences that somehow go
awry. But how could that be right if, as Tye has it, the latter contain
things-in-the-world where the former contain only “gaps”? Of course, it
could be that the psychologists have just got perception all wrong. But the
arguments they give for their inferential story are, after all, backed by
quite a lot of predictive and explanatory success. Externalists who propose
to dismiss inferentialism on a priori grounds (or on the common-sense ground
that perception doesn’t feel inferential) should bear it in mind: first that
the track record of a priorism in philosophical psychology has not been
good; and second that hubris is a sin.
Tye tries very hard to make sense of the claim that things-in-the-world are
parts of veridical experiences of things in the world. He suggests, for
example, that it might be construed on analogy to the doctrine, familiar at
least since Russell, that things that a proposition is about are part of the
proposition about them; on this way of putting things, John is part of the
proposition that John sneezed. But I think the analogy between Russell’s way
of talking about parts of propositions and Tye’s way of talking about parts
of experiences doesn’t really bear much weight. Propositions, unlike
experiences, are abstracta, so the question whether their constituents are
“inside” or “outside” of them doesn’t seriously arise. Indeed, as far as I
can see, to say that John is part of the proposition that John sneezed is
just to say that whether it’s true that John sneezed depends on how things
were with John. Sometimes it’s useful to talk Russell’s way; arguably, doing
so helps explain why whether John sneezed doesn’t depend on how John is
described. But Russell’s way of talking about propositions doesn’t license
claiming that John is part of an experience of his sneezing (or, come to
think of it, even that experiences have parts).
Now, Tye’s book has a by-no-means-hidden agenda. For its purposes, the
“problem of consciousness” is, par excellence, to reconcile an exhaustively
physical metaphysics with Realism about conscious experience. There really
are not just tables and chairs but also the phenomenal contents of
experiences of tables and chairs. Tye thinks that doing justice to both
requires the externalist assumption that the phenomenal content of a
veridical experience includes its object.
This is, as I’ve been suggesting, rather a lump to swallow; but never mind.
Let’s suppose, from here on, that something of the sort is true. So then,
Tye’s idea is that, given externalism we get direct perception, and given
direct perception we get to save physicalism from familiar problems about
phenomenal content. But is that all we’d need to make phenomenal content
consort with physicalism? And, if not, how much else is required? I’m pretty
sure there’s a lot.
Much of the problem that experiential content poses for materialists is, of
course, that the content of experience is, by assumption, conscious; and it
is, putting it mildly, puzzling how mere brains, or mere brain states, both
of which are assumed to be material through and through, could be conscious.
Direct perception, as Tye construes it, offers to reconcile materialism with
the consciousness of experience by making the object of an experience part
of the content of the experience and by assuming that the content of an
experience is ipso facto conscious content. But it seems, for all sorts of
reasons, to be just not true that the content of experience is ipso facto
conscious content. Psychologists will, at the drop of a hat, cite bushels of
prima facie examples to the contrary (“change blindness” and the “Sperling
effect”, both of which Tye discusses, would seem to be egregious examples).
Tye struggles bravely to accommodate such phenomena, but he does so at the
cost of needing extra wheels and gears to make his story run. He
acknowledges, for example, that “Sperling’s experiment provides strong
evidence that we can see and thus be conscious of things to which we are not
attending”. But is it true that we see and are conscious of things that we
don’t notice? Consider the worry about the brown moth to which Tye
frequently recurs: There is this brown moth sitting on a brown limb of a
brown tree, feeling fully disguised. One looks at the tree, attending to the
part of the branch that the moth is sitting on. Does it follow that one sees
and is conscious of the brown moth? Is the moth the object of (hence a part
of) one’s conscious experience when one looks at the tree? If Tye says “Yes,
you do see the moth” that’s bad for him because he needs the principle that
the constituents of experience are ipso facto conscious to connect his
externalism with his story about why phenomenal content is compatible with
physicalism. But if you see and are conscious of the moth, what more is
required if you’re to notice it? There’s the moth, mothing away, right
before your wide open eyes. And there’s no failure of attention since, by
assumption, you’re attending like mad to precisely the place on the tree
where the moth is. It looks as if the moth is in the attended part of your
visible field but you’re not conscious of there being a moth there. But if
Tye says “No, you don’t see the moth” that looks bad for him too: It raises
the question, “Why don’t you see it”. Or, to put it Tye’s way, why is it
that, in such cases, experience that contains a moth doesn’t contribute the
moth that it contains to the perceptual belief that it causes? It looks like
Tye can either have it that the parts of one’s experience are ipso facto
conscious, or he can have it that the object of an experience is ipso facto
one of its parts. But it doesn’t look like he can have both. (Tye has what
he takes to be an answer to this sort of question, but it requires a lot
more wheels than I imagine you want to hear about.)
I think, in fact, that the problem that the case of the moth raises for Tye is
much more serious than he supposes. Surely the natural way to deal with the
moth is to say that the perceiver sees it but doesn’t see it as a moth (he
sees it as part of the tree). Tye gets into trouble here precisely because
he can’t say the natural thing while also holding the externalist thesis
that it’s moths rather than moth-representations that experience of moths
delivers to the corresponding perceptual belief. The problem is perfectly
general given Tye’s externalism: He can’t explain how it is that the
perceptual beliefs that an experience causes represent things in the same
way that the experience does. How is it that, if the content of your
experience is a tree, the belief it causes will be that there is a tree
(rather than some other belief that is also true of the tree; e.g. that
there is a big tree, or that there is a leafy tree, etc). After all, the
assumption was that what the experience of the tree delivers to perception
is not a representation of the tree but the very tree itself; and trees
don’t come with descriptions attached. This is, of course, the moth all over
again; why doesn’t an experience of seeing a moth cause a belief that one
has seen a moth in the kind of case that Tye imagines?
What perceptual experience delivers to the perceptual belief it causes is not
the X but the X experienced-as-a-such-and-such; and it’s what the X is
experienced as that determines what belief is formed in consequence of the
seeing. The reliable correspondence between the content of beliefs and the
content of experiences that cause them can’t be an accident. To the
contrary, it’s what makes experiences worth having.
I think externalism has just about outlived its usefulness. Much of the
philosophy it has recently inspired looks like a reductio of its externalist
premisses; in addition to Tye on what is and isn’t a part of an experience,
see also Andy Clark’s new book Supersizing the Mind (reviewed in the TLS,
June 26, 2009) about how his iPhone has become (“literally”) part of his
mind. (If it has, I’m glad I don’t have one). Contrary to prevailing views,
I don’t think that it helps much with the mind–body problem to assume that
the object of your experience is part of your experience; or that it helps
much with epistemology to assume that your perception of things in the world
is “direct”; or that it helps much with cognitive psychology to deny that
perceptual processes are inferential. Nor, by the way, do I think that
externalism helps much with semantics. I have wept and prayed and changed my
mind a number of times; but I’ve now come to think that the moral of the
Putnam intuitions has been quite comprehensively misconstrued. What they
show isn’t that we need an externalist theory of meaning; what they show is
that we don’t need a theory of meaning at all. The only content that’s
required for the purposes either of theories of mind or of theories of
language, is referential content.
That, however, is another story.
Michael Tye
CONSCIOUSNESS REVISITED
Materialism without phenomenal concepts
229pp. MIT Press. $35; distributed in the UK by Wiley. £22.95.
978 0 262 01273 7
Jerry Fodor is Professor of Philosophy at Rutgers University. His books
include The Mind Doesn’t Work That Way: The scope and limits of
computational psychology, 2000, Hume Variations, 2003, and LOT 2: The
language of thought revisited, which was published last year.
Your Comments
Order By: