Although the Tractatus was intimately concerned with linguistic meaning, it had put semantics on the index. The things which cannot be ‘said’ include the meaning of signs and that two signs have the same meaning (see TLP 3.33, 3.332 and 6.23), what a given symbol signifies (4.126) and the sense of a proposition (2.221, 4.022). But these things can be ‘shown’: they reveal themselves in bipolar propositions with a sense, provided that the latter are properly analysed. On his return to philosophy from 1929 onwards, Wittgenstein abandoned the saying/showing distinction, though gradually and hesitantly. As a result, matters of meaning assumed a central role not just implicitly, but officially. Wittgenstein even declared the ‘transition from the question of truth to the question of meaning’ (MS 106, 46, my translation) to be central to his philosophical method. This clear and succinct statement of the linguistic turn makes the meaning of expressions central to philosophy. The later Wittgenstein also revolutionized analytic philosophy by following through another incipient idea of the Tractatus, namely by completing the latter's partial move away from a referential conception of meaning. According to this conception every meaningful expression stands for an object, the latter being its meaning. Wittgenstein's main objections match the simplicity, if not the simplemindedness, of the target.

Not all meaningful words refer to objects. The referential conception is modelled mainly on proper names, mass nouns and sortal nouns. It ignores verbs, adjectives, adverbs, connectives, prepositions, indexicals and exclamations (PI §§1–64). Moreover, even the meaning of a referring expression is not the object it stands for. If it were, referential failure would have to render a proposition like ‘Mr. N.N. died’ senseless (PI §40).

One should add Ryle's criticism (1971b, Ch. 27): identifying the meaning of a word with its referent is a ‘category mistake’, namely of confusing what a word stands for with its meaning. I can be 20 kilometres away from the referent of the definite description ‘the highest peak on earth’, but not from its meaning. One should further add an argument inspired by Austin (1946) (cited after the reprint in 1970, 96–97; see Rundle 1979, 380).