However, using these would lead to trivialities like The meaning of 'Every farmer' is the meaning of 'every farmer' 'Every farmer owns a donkey' means that every farmer owns a donkey We may, of course, also say something less trivial, e.g. 'Every farmer owns a donkey' means that for every x, if x is a farmer, then x owns a donkey; however, what is nontrivial with this is the purported synonymy of 'Every farmer owns a donkey' and 'For every x, if x is a farmer, then x owns a donkey'; i.e. the fact that the latter is - in a certain sense - a faithful translation of the former.
Thus it seems that there is no interesting direct describing of the meaning of an expression which would not rest on finding an interesting translation of the expression into another language (or an interesting paraphrase of the expression in the same language); and providing (i) seems to be in this sense parasitic upon providing (iv)9.
The Myth of the Structure One of the common way to avoid this ‘intractability of meaning’ is to move the concept of meaning to the periphery of one’s teaching and to concentrate on the word struture.
Thus, for many theoreticians of language, meaning has come to coincide with something like the semantic structure; and semantic analysis with pinpointing this structure.
This might be understood as accepting the structural approach to language urged above - but usually it is not. The point is whereas what we have urged is an approach which sees meaning of an expression as the possitionof the expression within the network of language, the common way of engaging the concept of structure is based on the picture that an expression is like, say, a mineral: that it can be analyzed and examined with tools akin to microscopes up to the point where we see its structure.
It is, of course, also only another expression of the fact spelled out earlier in the paper: namely that meaning is not a thing, but rather a value.
Hans Kamp has suggested to me that one of the ways to express this is the following: „A theory of the implementation of memory presupposes a theory of understanding of meaning“.
Note that this would not be the case if (3) were the picture of the meaning of the analyzed sentence.
In any case, talk about meaning is in the clear sense talk about types, not about tokens; and semantics is - in this sense - inevitably realistic.
On the other hand, even the most diehard realist has to assume that there are some contingent facts that elicit which meaning an individual expression has. We do not discover meaning by an 'intellectual trip' into a realm of abstracta where we would see them attached to expressions; but rather by observing and recording certain concreta.