Labyrinths of Reason (34 page)

The cave dweller might be given a more active role with robotic technology. His arms and legs could be fitted with sensors. The movements of the real limbs would be relayed to a robot body outside, situated near the TV camera. As the cave dweller lifts his finger, the robot finger moves, much as in a radioisotope lab. With sophisticated robotics, this could mimic the brains-in-vats situation. The cave dweller would think he inhabited the robot body in the outside world. He would have no way of knowing he was actually in a cave, and would be skeptical of any such claim.

One question raised by these fantasies is how much “information” is necessary to create a mental image of a world. Here’s a more extreme predicament. Let the prisoner’s TV screen be a video-text terminal. Rather than pictures, a running commentary (in written English) of what is happening outside the cave fills the screen at all times. Never does the prisoner see a bird or even a TV image of a bird. Instead a message like “
A BLUEBIRD JUST PERCHED IN THE TREE OUTSIDE THE CAVE OPENING
” scrolls across his screen.

Now the situation resembles that of the decipherers of a lost language. Again, the cave dweller has no experience of the outside world at all. He has been strapped to the wall of the cave since birth. He cannot know how to read
unless
he can puzzle out the meaning of the strange symbols filling his visual field. Is that possible?

The prisoner would probably learn a lot about the structure of written English. He would know the shapes of every letter and punctuation mark by heart. He would recognize short common words. Nothing else demands his time in the cave; much of his imagination would be directed to the screen hieroglyphs. Maybe he
would come to know all the common words without making an effort, just as a rancher may recognize, though not know the names for, dozens of kinds of wildflowers.

Learning the meaning of words is something else again. Language is based on shared experiences. There is no way of describing the color red to a person blind from birth. The poverty of the cave dweller’s life would allow few points of reference.

Conceivably, the prisoner might notice that the sentence “
THE SUN IS RISING
” appears at regular intervals. With enough patience, he might notice that some words or phrases (like “
MORNING,” “OWL,” “FULL MOON
,” and “
SNOW
“) appear on the screen at certain times and not at others. These temporal clues might provide a starting point for deducing the meaning of the writing (though it seems unlikely the prisoner would get very far).

The absolute minimum amount of information that can be conveyed about anything is a simple yes or no—the “bit” or binary digit of computer science. Consider this ultimate version of the parable: Outside the cave is a TV camera as before. It converts images of the landscape into electrical impulses. Information about brightness, hue, saturation—all the pellucid effects of a rainbow—are coded as a sequence of 1
’S
and 0’s that have meaning to digital video equipment: 01011010011001010110111001101111 … This information travels by cable into the cave. In this case there is no TV screen in the cave. Instead the stream of incoming bits is fed into a much simpler display device. When the device receives a 1 in the input, it projects a dot of white light on the wall of the cave in front of the prisoner. When it receives a o, the wall is dark. The result is a flickering dot of light on the cave wall. The prisoner spends his entire life watching that flickering dot and nothing else.

In a strong sense, we are
all
like this prisoner of the cave. Our entire experience of the world is a sequence of nerve impulses that could be expressed as a sequence of 1’s and 0’s. What is amazing is that we can draw any conclusions whatsoever from such an impoverished input. Everything, from the fact that space has three dimensions to World Series predictions, is derived from that same abstract input. That is the puzzle of knowledge: how we derive meaning from symbols that seem incapable of expressing anything of the variety of the world.

Next to this denizen of the cave is another. His situation is apparently
the same. His entire life is spent watching a flickering dot encoding a sequence of 1
’S
and 0’s. By the same miracle of imagination, he has built a rich mental picture of the outside world: its geometry, its eons of past history, its distant future. But because of an equipment malfunction that has caused his dot of light to go on and off at random, this picture is
completely wrong
.

Now presumably this is impossible. To put it in a slightly different context, suppose there are two brains in vats. Brain A receives a carefully modulated stream of impulses to create the illusion of a world, and brain B receives a random stream of impulses because of a hardware error. Surely there is something “worldlike” in A’s input, and not in B’s random input. Brain B would not be able to make sense of its input at all. The meaning is inherent in A’s input—so we would guess.

Several philosophers have used brains-in-vats thought experiments to explore the issue of meaning. In
Reason, Truth and History
(1981), Hilary Putnam controversially argues that we are
not
brains in vats and can know it. This brought howls of protest from the philosophic community. His reasoning is clever and does not quite assert what it seems.

Assume, as the premise of a reductio ad absurdum, that we are brains in vats. Then when we say “bowling ball” (and of course we don’t really
say
anything, having no lips) we refer not to physical round objects with three holes in them—of which there may be none. (There may be no bowling alleys in the “real” world outside the brains-in-vats laboratory.) Even so, “bowling ball” refers to
something
. It refers to a certain pattern of electrical stimulation by which the laboratory’s mad scientists create the illusion of bowling balls. This is the physical counterpart of the thought, the object of the reference.

You might say there are two languages, vat-speak and lab-speak. “Bowling ball” in lab-speak refers to the round thing with three holes. “Bowling ball” in vat-speak refers to an electrical impulse that creates the image of a round thing with three holes.

If “bowling ball” refers, in vat-speak, to electrical impulses, what does “brain” refer to? Not to a gray lump of neurons, but to another set of electrical impulses, the impulses by which the illusion of physical brains is created. “Vat” refers to an electrical stimulation too. So if we are brains in vats, the words “brains in vats” refer not to
physical brains in physical vats but to one type of electrical stimulation “in” another type of electrical stimulation. To say “Yup, I’m a brain in a vat” is wrong because we are not electrical impulses, we are
real
brains in
real
vats!

Putnam therefore sees the statement “I am a brain in a vat” as necessarily false. It is perhaps comparable to the statement “The universe rests on the back of a big turtle.” You can maintain that this is necessarily false too, because “universe” means everything, including the turtle if it exists. The universe can’t rest
on
anything else because there
is
nothing else—by definition.

Defensible as this is, it neglects the flexibility of language. The average person would understand “The universe rests on the back of a big turtle” to mean that the known universe of stars and galaxies rests on an unknown turtle. We automatically redefine “universe” to fit the context of the sentence—as we would with a statement like “We are brains in vats.”

A brain in a vat
might
be able to express the “real” state of affairs and still satisfy semantic purists. It would have to recognize the difference between vat-speak and lab-speak and assert something like “I am what ‘a brain in a vat’ means in lab-speak.” Arguably (?) this avoids the same problem of misplaced reference because “lab-speak” is a metaphysical term of vat-speak with no physical correlative.

Putnam’s best-known thought experiment challenges the idea that meanings are “in the head,” a matter of mental states. Suppose, Putnam says, there is another planet in our galaxy called Twin Earth. Twin Earth is exactly like Earth in almost every way. There are regular-looking people on Twin Earth, and they even speak English (as in a lot of science-fiction movies). Twin Earth is so incredibly similar to Earth that they call their planet “Earth.” (It would be silly of them to call it “Twin Earth”!)

One difference between Earth and Twin Earth is this: Oceans, rivers, lakes, raindrops, and tears on Twin Earth are made out of a transparent liquid that looks exactly like water but isn’t water. That is to say, it isn’t
chemically
water. Instead of H
₂
O, it has a different formula we’ll write as XYZ. But so amazing is the “parallel evolution” of Twin Earth that they still call this liquid “water.” When an inhabitant of Twin Earth talks of watering the lawn, he means putting XYZ on the lawn. Putting H
₂
O on the lawn might kill it.

There is H
₂
O on Twin Earth, a few tightly stoppered bottles of it in chemistry labs, but they call it something other than “water.” There is XYZ on Earth. Chemists on both planets can distinguish the two compounds with a simple test.

Now think what would happen if, a few centuries from now, we sent a spaceship to Twin Earth. Our astronauts would get out of the ship, take off their space helmets, and introduce themselves in English to the natives. After a while, one astronaut would get thirsty and ask for some water. A Twin Earth host would go to the faucet and draw out a nice, tall glass of “water.” The astronaut would put the glass to her lips, take a sip, and spit it out! The Earthlings would run chemical tests and find that the “water” on Twin Earth is toxic, undrinkable XYZ.

Now turn back the calendar to the year 1750—the year called “1750 A.D.” on both Earth and Twin Earth, for they have identical calendars. There is no such thing as space travel. Our astronomers’ crude telescopes have not yet picked out the star that Twin Earth orbits, and vice versa. Chemistry is in its infancy as well. Our chemists have not yet discovered that water is made of hydrogen and oxygen atoms. Twin Earth’s chemists have not yet discovered that their “water” is made of X, Y, and Z.

In the Earth of 1750 there is an individual called Oscar. There is an extremely similar individual, also called Oscar, on Twin Earth. The two Oscars are so alike that they even have the same thoughts at all moments of their lives. When Earth’s Oscar uses the word “water,” it has precisely the same memories and mental associations as it does when Twin Earth’s Oscar uses the word. Both think back to a certain water fountain on a school playground; to the first time they saw the Atlantic Ocean (a geographic feature common to both planets); to the water that leaks through the roof of their house in a heavy rain. If you were to ask the Earth Oscar to explain what water is, he would say it’s such and such, and if you were to ask the Twin Earth Oscar, he would say exactly the same thing. There is nothing in the one Oscar’s consciousness that distinguishes his understanding of water from the other Oscar’s. Yet the two waters are not at all the same. Putnam concludes: “Cut the pie any way you like, ‘meanings’ just ain’t in the
head!”

But if meanings aren’t in the head, where are they?

Many philosophers
do
believe that meaning is largely “in the head.” A word like “water” could mean anything at all. It means what it happens to mean, what we are thinking of when we say “water,” and not something that is dictated by the shapes of the letters. If we found a very short scroll in a lost language and all it said was “water,” there would be no hope of translating it.

There was a time for all of us when we were unsure what “water” meant. Parents and other adults said “water” in certain contexts, and it was up to each of us to decide what was common to each context. As adults we think our experience is broad enough to eliminate that ambiguity. At this late date, could “water” possibly be anything other than what we think it is?

Two objections are often raised to Putnam’s thought experiment. It is true, though not necessarily germane, that Twin Earth is biochemically implausible. Lest that concern get in the way, a word about the putative chemistry of Twin Earth may be in order.

In Putnam’s 1975 article, he describes the hypothetical XYZ as a liquid “whose chemical formula is very long and complicated.” It is said to be liquid at the same range of temperatures and pressures as H
₂
O, and of course it quenches Twin Earth thirsts and otherwise fulfills the ecological and biochemical role that our water does.

There is no known substance that is so nearly similar to water, without
being
water, as that. It is doubtful that any substance with a complex formula would fit the bill. Water’s predominant role in Earthly biochemistry depends on the small size of its molecules. Those liquids with long, complicated formulas are generally oily, viscous, and otherwise unwaterlike.

Hydrogen peroxide (H
₂
O
₂
), the only other compound of hydrogen and oxygen, is wildly unstable and could not exist as oceans. (The “hydrogen peroxide” sold in drugstores is a very weak water solution of this compound.) Hydrogen sulfide (H
₂
S) is chemically analogous to water but is a gas. Looser analogues are gaseous ammonia (NH
₃
) and hydrogen fluoride (HF), a virulent acid whose boiling point is just below (Earthly) room temperature.

Science-fiction writers sometimes speculate about life on planets where ammonia takes the place of water in the biochemistry. Such planets would have to be much colder than Earth, for ammonia is a liquid only below temperatures of 36 degrees below 0 Fahrenheit
(—33 degrees C.). (The liquid “ammonia” sold for cleaning windows is again a water solution of gaseous ammonia.)