ABSTRACT
The theory that sentences are models seems to offer an account of language which solves a large number of the problems which Wittgenstein inherited from Frege and Russell. It also seems to vindicate a distinctive metaphysics, and so to contribute decisively to what has seemed to be a central task of philosophy, throughout its history. For all that, the theory as it stands is incomplete, even as an account of language, in two very obvious ways. First of all, the theory that sentences are models seems to
apply so far only to basic non-compound sentences. It seems not to apply straightforwardly to any compound sentences — sentences which have whole sentences as parts — or to any of the sentences of our everyday languages, which Wittgenstein seems to regard as general statements whose instances would be elementary sentences. So it seems that the theory that sentences are models applies straightforwardly
only to sentences which Wittgenstein thinks must be possible, in principle — not to the sentences we actually have. And, second, no account has yet been given of a large class
of subsentential expressions. We know that sentences are taken to be facts, and describe facts (correctly or incorrectly); they are not objects themselves, and are not meaningful in virtue of being correlated with objects. We know that genuine singular terms — the names of which elementary sentences are formed — are held to be meaningful in virtue of being correlated with objects. It is plausible to suggest that this correlation between names and objects is simply a matter of there being a rule which correlates whole sentences in which the names appear with certain types of fact. And this allows at least some expressions which we might previously have regarded as predicates to be classed as names. What we do not know is how Wittgenstein thinks that the so-called ‘logical constants’ should be treated. We do not know how he thinks the negation sign is meaningful, or how the connectives which are at the basis of modern logical systems — connectives which translate ordinary-language terms like ‘if’, ‘and’, and ‘or’ — are supposed to work. And we do not know yet what his view is of signs of generality — in particular, ‘all’ and ‘some’. Since almost all sentences of everyday language are to be understood in terms of generality, on Wittgenstein’s view, this is a very significant lacuna in the theory so far. Furthermore, in addition to these two ways in which the
account of language is incomplete, the Tractatus has not yet addressed another major concern which Wittgenstein inherited from Frege and Russell, a concern which might indeed be thought to underlie his whole attitude to philosophy: to show how logic (and with it mathematics) was independent of everything to do with intuition or acquaintance, and thereby reject the Kantian commitment to synthetic a priori truths.
