indirectly related to this unsettling conclusion,

COGNITIVE SCIENCE 99
indirectly related to this unsettling conclusion. In fact, these arguments arose,
in the first instance, out of discussions of lexical meaning.2 Working out just
how the issues about lexical meaning connect with the issues about mental
representation is actually not easy much of this paper will be devoted to doing
so. For starters, though, here's the original argument without comment or
elaboration.
Hilary Putnam (see especially [9]) imagines a place that's just like here
except for certain peculiarities of microchemistry. Call this place "Twin Earth"
("Earth2" for short). On Earth2 they speak a language that is just like English
in respect of its phonological, morphological, and syntactic properties. They
call this language by a word which they pronounce /English/ and which they
write "English" (but which we will write "English2" in aid of notational
perspicuousness). Since English2 is phono-morpho-syntactically just like English,
it contains a word (which we will write as "water2") that is pronounced
/water/.
The microchemical difference between Earth and Earth2 is that, although
they, like us, have a transparent fluid that they drink, sail on, wash thencars
with, and refer to by the vocable /water/, and although that fluid passes
all the, as it were, phenomenological tests for water (it has specific gravity 1,
it freezes at zero C, and so forth), still the stuff that looks like water on Earth2
is, in point of chemical fact, made of XYZ {Φ H2O).
Putnam's intuitions about what's going on on Earth2 run like this:
a. What English2 speakers refer to by using the word "water2" is not
water.
b. English2 expressions like "water2 is wet" have different truth conditions
from the homophonic expressions of English. In particular,
unlike the English homophone, the truth of "water2 is wet" does not
depend upon the wetness of water. Hence,
c. "water2" is not the same word as "water"; the two words differ in
their semantic properties.
It follows trivially that English Φ English2 ("water" occurs in one language
but not in the other). Since, however, we have assumed that the difference
between XYZ and H2O is the only (relevant) difference between Earth and
Earth2, it also follows that people who are as similar as you like in their
physical constitution (people who are, as we shall say, molecularly identical*)
may nevertheless speak different languages. Notice that it has not been shown
to follow (yet) that molecularly identical people may mean different things
by what they say; the grounds for that latter inference remain to be explored.
So much for Earth2. Its philosophical chemistry is, I suppose, now sufficiently
well publicized that we needn't develop the example in further detail.
However, before we get to our main topic, which is what Putnam's case is
supposed to show about the notion content of a mental representation, it is
desirable to understand something of what cognitive scientists have wanted
that notion for; what role appeals to content are supposed to play in the
sorts of explanations that cognitive science seeks to provide. Hence the following
section.