A Puzzle about Earth2 Beliefs,

COGNITIVE SCIENCE 103
A Puzzle about Earth2 Beliefs The upshot of the discussion so far is that
we need for purposes of theorizing about the intentional properties of behavior
the two coordinate constructs: de dicto specification of a propositional
attitude and content of a mental representation. If, then, we want to think
about the implications of the Earth2 examples for RTM, we need to ask
what they show about the de dicto propositional attitudes of people who talk
English2- We now turn to that question.
The first thing to say is that Putnam gives us very little help here. His
discussion is framed almost solely in terms of semantic and lexicographic
issues: e.g., in terms of such questions as "what does 'waterj mean!"1 Now,
one might suppose at first-blush that to settle questions about what "water2"
means is to settle the corresponding questions about the de dicto propositional
attitudes of speakers of English2. The reason one might suppose this is that
it is very natural to assume a "Gricean" account of the relation between the
meanings of linguistic forms and the de dicto propositional attitudes of
speaker/hearers of the language that contains the forms. The idea would be
that meanings are, as it were, logical constructs out of the de dicto propositional
attitudes of speaker/hearers (e.g., out of their de dicto havings and
recognizings of communicative intentions). There might be some uncertainty
about just which logical construction out of propositional attitudes meanings
are; but however that goes the view would be that to fix the ascription of
meanings to verbal forms is to presuppose, more or less uniquely, an ascription
of corresponding de dicto propositional attitudes to speaker/hearers.8
I am very much in sympathy with this sort of view, and it may be that
it can be reconciled with Putnam's intuitions about what "water2" means.
The present point, however, is that if one does accept Putnam's intuitions,
one cannot simply take the existence of a Gricean reduction of meanings
to propositional attitudes for granted.9 We will see more of this later (indeed,
it will become a major theme) but here is one relevant consideration: Putnam
wants to make the extension of a term one of the parameters of its meaning
so that, presumably, "water2" means XYZ together with some other stuff.
And Putnam also wants to argue that speaker/hearers need not know what the
terms they use refer to. It is, indeed, the conjunction of these two doctrines
that Putnam expresses by the slogan "meanings aren't in the head". However,
I suppose we can assume that people normally do know their own de dicto
intentions; that de dicto propositional attitudes are "in the head" even if
meanings are not. Surely this assumption is part and parcel of the sort of
reduction of meanings to propositional attitudes that Grice proposes. But if
this is right, it is hard to see how the view that "water2" means (something
like) XYZ could be squared with the idea that a word means what it does
because speaker/hearers have the de dicto propositional attitudes that they
have.
All this is pretty tentative. It would, for example, be possible to give
up the idea that people know their own de dicto propositional attitudes. You
might then manage a Pickwickean-Gricean reduction of meanings to communicative
intentions; Pickwickean because the communicative intentions to
which meanings are reduced would, in a relevant sense, not themselves be
psychological states. (And, of course, you would need to find some way of