John Perry
August 21, 1996
Words stand for things of various kinds and for various kinds of things. Because words do this, the sentences made up of words mean what they do, and are capable of expressing our thoughts, our beliefs and conjectures, desires and wishes. This simple idea seems right to me, but it flies in the face of formidable authority. In a famous passage in ``Reality without Reference,'' Donald Davidson criticizes what he calls the ``building-block theory:''
[T]he essential question is whether [reference] is the, or at least one, place where there is direct contact between linguistic theory and events, actions, or objects described in nonlinguistic terms. If we could give the desired analysis or reduction of the concept of reference then all would, I suppose, be clear sailing. Having explained directly the semantic features of proper names and simple predicates, we could go on to explain the reference of complex singular terms and complex predicates, we could characterize satisfaction (as a derivative concept), and finally truth. This picture of how to do semantics is (aside from the details) and old and natural one. It is often called the building-block theory. It has often been tried. And it is hopeless.
The picture I find attractive and Davidson finds confused and hopeless seems to be the very one Wittgenstein saw in Augustine. At the beginning of the Philosophical Investigations he quotes a passage from Augustine, and finds in Augustine's words ``a particular picture of the essence of human language,'' that goes like this:
[T]he individual words in language name objects--sentences are combinations of such names.--In this picture of language we find the roots of the following idea: Every word has a meaning. This meaning is correlated with the word. It is the object for which the word stands.
Davidson and Wittgenstein seem to be in considerable
agreement about this picture, for similar reasons. Both
Wittgenstein and Davidson insist that the interpretation of
language derives from its connection with human action;
Wittgenstein introduces his concept of a language game, and
Davidson emphasizes the idea that semantic terms are to be
explained by being connected with ``human ends and
activities.'' Wittgenstein quotes
with apparent approval Frege's dictum that words only have
meaning as parts of sentences, one of Davidson's favorite
passages.
Nevertheless, there are important differences between what Wittgenstein and Davidson reject about the building block picture. Wittgenstein's main point is that we need to understand the phenomenon of words having meaning by reference to their role in the fabric of human action, as opposed, say, to their being associated with internal images. It may be, for all the considerations he raises show, that reference is the point or one of the points where linguistic theory makes contact with human intentions and actions. In fact, in this essay I'll use Wittgenstein's language game involving a builder and his assistant to argue that this is so.