The first section of this chapter takes advantage of the opportunity offered by a few propositions of Wittgenstein’s Tractatus Logico-Philosophicus to pose some of the questions to be pursued throughout the succeeding sections in the context of what has been said by a selection of earlier and later philosophers about representation in language or, as we shall eventually find reason to write, “representation” in language.
How can we represent how language represents? How language represents does not have to be represented in language, if by language we mean words used in sentences, as I am using them here now. It could be done, apparently, by a diagram, thus: W→T. Or does this fail because we have to add that “W” represents a word, “T” represents a thing, and “→” represents the relation of representation, so that the original question recurs? Should we conclude that an attempted pictorial representation of linguistic representation demands a linguistic representation of pictorial representation? Or should we conclude that linguistic representation defies representation? It might be said, and it has been said, that we should not oppose pictorial to linguistic representation, and that we should distinguish representation as saying from representation as showing. A famous advocate of this double proposal argues that linguistic representation is ultimately pictorial representation, provided the latter be understood as a projective relation between the concatenations of logically proper names that make up elementary propositional facts and the objects or things arranged in other facts or possible states of affairs—in a Sachverhalt, a moglicher Sachverhalt, a Sachlage, or a mogliche Sachlage, as Wittgenstein variously says, leaving open a perplexingly wide range of options. I shall resist the temptation to embark upon the detailed analysis of the Tractatus without which we cannot begin to determine more precisely what these options are. I shall do no more than note that evidence can be produced