Darwin's Dangerous Idea (78 page)

rather than a falsehood about horses? (When does the aboutness or inten-Unless there were "meaningless" or "indeterminate" variation in the trig-tionality leap to the new position?—a saltation demanded by these theorists.) gering conditions of the various frogs' eyes, there could be no raw material Would you ever make the transition? Would you somehow make it without (blind variation) for selection for a
new
purpose to act upon. The indeter-knowing it? The two-bitser was forever oblivious of the change of meaning minacy that Fodor (and others) see as a flaw in Darwinian accounts of the of its internal state, after all.

evolution of meaning is actually a precondition for any such evolution. The I suppose you may be inclined to think that you are quite radically unlike idea that there must be
something determinate
that the frog's eye really the frog and the two-bitser. You, it may seem, have intrinsic or original means—some possibly unknowable proposition in froggish that expresses intentionality, and this marvelous property has a certain amount of inertia:
exactly
what the frog's eye is telling the frog's brain—is just essentialism your brain can't just turn on a dime and suddenly mean something entirely applied to meaning (or function). Meaning, like function, on which it so new by its old state. In contrast, frogs don't have much of a memory, and directly depends, is not something determinate at its birth. It arises not by two-bitsers none at all. What
you
mean by the word "horse" (your private saltation or special creation, but by a (typically gradual) shift of circum-mental concept of a horse) is something like
one of those equinish beasts that
stances.

we Earthlings like to ride,
an epithet anchored in your mind by all your Now we are ready for the only case that really matters to these philoso-memories of horse shows and cowboy movies. Let us agree that this memory phers: what happens when we move a
person
from one environment to matrix
fixes
the kind of thing to which your concept of horse applies.
Ex
410 THE EVOLUTION OF MEANINGS

The Quest for Real Meaning
411

hypothesi,
schmorses are not beasts of
that
kind; they are not of the same error, something we might never be able to discover, but a fact nonetheless?

species at all, but just conveniently indistinguishable by you from horses. So, The same thing could happen to you, after all: imagine that a biologist told according to this line of thought, you make unwitting errors every time you vou one day, that coyotes are in fact dogs—members of the same species.

(mis)classify a schmorse perceptually or in reflection ("Wasn't that a fine You might find yourself wondering whether the biologist and you had the horse I saw gallop by my window yesterday!").

same concept of dog. How strong was your allegiance to the view that "dog"

But there is another way of thinking of the same example. Nothing forces was a species term and not the name for a large subspecies of
domestic
dogs?

us to suppose that your concept of a horse wasn't more relaxed in the first Does your heart of hearts tell you loud and clear that you had already ruled place, rather like your concept of a table. (Try telling the story of Twin Earth out "by definition" the hypothesis that coyotes are dogs, or does it silently with the suggestion that the tables there aren't
really
tables, but just look like allow that your concept has all along had the openness to admit this tables and are used for tables. It doesn't work, does it?) Horses and schmorses purported discovery? Or would you find that now that the issue has been may not be the same biological species, but what if you, like most Earthlings, raised, you would have to settle, one way or the other, something that had have no clear concept of species, and classify by appearance:
living thing that
simply not been fixed before because it had never come up before?

looks like Man-o-War.
Horses and schmorses both fall into that kind, so, Such a threat of indeterminacy undermines the argument in Putnam's when you call a Twin Earth beast a horse, you're
right
after all. Given what thought experiment. To preserve its point, Putnam tries to plug this gap by you mean by "horse," schmorses
are
horses—a non-Earthly kind of horse, declaring that our concepts—whether we know it or not—refer to
natural
but a horse just the same. Non-Earthly tables are tables, too. It is clear that
kinds.
But which kinds are natural? Varieties are just as natural as species, you
could
have such a relaxed concept of horses, and that you could have a which are as natural as genera and higher classifications. Essentialism with tighter concept, according to which schmorses are not horses, not being of the regard to the meaning of the frog's mental state and the two-bitser's inner same Earthly species. Both cases are possible. Now, must it be
determinate
state
Q
(or
QB)
was seen to evaporate; it must evaporate just as surely for us.

whether your horse concept (prior to your move ) meant the species or the The frog
would have
zapped just as readily at pellets in the wild, had any wider class? It might be, if you are well read in biology, for instance, but come its way, so it certainly was not equipped with anything that discrim-suppose you are not. Then your concept— what "horse"
actually
means to inated
against
pellets. In one sense,
fly-or-pellet
is a natural kind for frogs; you—would suffer the same indeterminacy as the frog's concept
of fly
(or they
naturally
fail to discriminate between the two. In another sense,
fly-orwas it all along the concept
small airborne food item?
).

pellet
is not a natural kind for frogs; their natural environment has never It might help to have a more realistic example, something that could made that classification relevant before. Exactly the same is true for you.

happen right here on Earth. I have been told that it once was the case that the Had schmorses been secretly brought to Earth, you would have just as Siamese had a word for, well, "cat" but had never seen or imagined any other readily called them "horses." You would have been wrong if it was somehow cats than Siamese cats. Let's suppose their word was "kat"—it doesn't matter already fixed that what your term meant was the species, not the lookalikes, what the actual details were, or even whether this particular tale is true. It but if not, there would be no grounds at all for calling your classification an could be. When they discovered that other varieties existed, they had a error, since the distinction had never before come up. Like the frog and the problem: did their word mean "cat" or "Siamese cat"? Had they just two-bitser, you have internal states that get their meanings from their discovered that there were other, rather different-looking, sorts of kats, or functional roles, and where function fails to yield an answer, there is nothing that kats and those other creatures belonged to a supergroup? Was their more to inquire about.

traditional term the name of a species or a variety? If they lacked the The tale of Twin Earth, if we read it through Darwinian spectacles, proves biological theory that made this distinction, how could there possibly be a that human meanings are just as derived as the meanings of two-bitsers and fact of the matter? (Well, they might discover that peculiarities of appearance frogs. This is not what it was intended to show, but any attempt to block this were really very important to them—"That just doesn't look like a kat, so it interpretation is forced to postulate mysterious and unmotivated doctrines of isn't one!" And you might discover that you similarly resisted the suggestion essentialism, and to insist, point-blank, that there are facts about meaning that Shetland ponies were horses.)

that are utterly inert and undiscoverable, but facts all the same. Since some When a Siamese person saw a ( non-Siamese ) cat walking by and thought philosophers are ready to swallow these bitter pills, I need to provide a few

"Lo, a kat!" would this be an error or the simple truth? Perhaps the Siamese farther persuasions.

person wouldn't have any opinion about how to answer this question, but The idea that our meanings are just as dependent on function as the could there nevertheless be a determinate fact about whether this was an meanings of the states of artifacts, and hence just as derivative and potentially indeterminate, strikes some philosophers as intolerable because it fails 412 THE EVOLUTION OF MEANINGS

Two Black Boxes
413

to give meaning a proper
causal role.
This is an idea we have seen in an earlier incarnation, as the worry about minds' being mere
effects,
not originating
causes.
If meaning gets determined by the selective forces that endorse certain functional roles, then all meaning may seem, in a sense, to be only
retrospectively
attributed: what something means is not an
intrinsic
property it has, capable of making a difference in the world at the moment of its birth, but at best a retrospective coronation secured only by an analysis of the subsequent effects engendered. That is not quite right: an engineering analysis of the two-bitser, newly arrived in Panama, would permit us to say what roles the device, so configured, would be
good for,
even if it had not yet been chosen for any role. We could reach this verdict: its acceptance state FIGURE 14.1

could
mean "quarter-balboa here now" if we put it in the right environment.

But of course it could also mean lots of other things, if placed in other pushed the a button on box A, the red light flashed briefly on box B, and environments, so it won't mean any one of them until a particular functional whenever you pushed the (ß button on box A, the green light flashed briefly.

role for it gets established—and there is no threshold for how long it takes a The amber light never seemed to flash. They performed a few billion trials, functional role to become established.

under a very wide variety of conditions, and found no exceptions. There That is not enough for some philosophers, who think that meaning, so seemed to them to be a causal regularity, which they conveniently summa-construed, doesn't pull its weight. The clearest expression of this idea is Fred rized thus:

Dretske's (1986) insistence that meaning must itself play a causal role in our mental lives in a way that meaning never plays a causal role in the career of All a's cause reds.

an artifact. Put this way, the attempt to distinguish real meaning from All ß's cause greens.

artificial meaning betrays a yearning for skyhooks, a yearning for something

"principled" that could block the gradual emergence of meaning from some The causation passed through the copper wire somehow, they determined, cascade of mere purposeless, mechanical causes, but this is (you must since severing it turned off all effects in box B, and shielding the two boxes suspect) an optional and tendentious way of putting it. As usual, the issues from each other without severing the wire never disrupted the regularity. So are more complex than I am showing,4 but we can force the key points into naturally they were curious to know just how the causal regularity they had the open with the help of a little fable I recently devised precisely to give discovered was effected through the wire. Perhaps, they thought, pressing these philosophers fits. It works. First the fable, and then the fits.

button a caused a low-voltage pulse to be emitted down the wire, triggering the red light, and pressing button (3 caused a high-voltage pulse, which triggered the green. Or perhaps pressing a caused a single pulse, which triggered the red light, and pressing ß caused a double pulse. Clearly, there 2. Two B

5

LACK BOXES

was something that always happened in the wire when you pressed button
a,
and something different that always happened in the wire when you pressed Once upon a time, there were two large black boxes, A and B, connected by ß. Discovering just what this was would explain the causal regularity they a long insulated copper wire. On box A there were two buttons, marked a had discovered.

and ß, and on box B there were three lights, red, green, and amber. Scientists A wiretap of sorts on the wire soon revealed that things were more studying the behavior of the boxes had observed that whenever you complicated. Whenever
either
button was pushed on box A, a long stream of pulses and gaps (ons and offs, or bits) was sent swiftly down the wire to box B—ten thousand bits, to be exact. But it was a different pattern each time!

Clearly, there had to be a feature or property of the strings of bits that 4.
Dretske and His Critics
(McLaughlin, ed, 1991) is devoted largely to this issue.

triggered the red light in one case and the green light in the other. What 5. This began as an impromptu response to Jaegwon Kim's talk at Harvard, November 29, could it be? They decided to open up box B and see what happened to the 1990: "Emergence, Non-Reductive Materialism, and 'Downward Causation,' " and evolved under the persistent rebuttals of Kim and many other philosophers, to whom I am grateful.

414 THE EVOLUTION OF MEANINGS

Two Black Boxes
415

strings of bits when they arrived. Inside B they found a supercomputer— just
suggest
to the scientists that there were really three varieties of strings: red an ordinary digital, serial supercomputer, with a large memory, containing a strings, green strings, and amber strings—and amber strings outnumbered huge program and a huge data base, written, of course, in more bit strings.

red and green strings by many, many orders of magnitude. Almost all strings And when they traced the effects of the incoming bit strings on this computer were amber strings. That made the regularity they had discovered all the program, they found nothing out of the ordinary: the input string would more exciting and puzzling.