Ludwig Wittgenstein,
Philosophical Investigations
Overview of §§1-120

Curtis Brown

Theories of Meaning: Review

We have considered three main theories of meaning so far:

  1. Referential view. Meaning is naming; every word stands for something, and the something it stands for is its meaning.
  2. Ideational view. Meanings are ideas. In the crudest version, ideas are understood as images.
  3. Propositional view. Meanings are abstract entities.

Of these three views, the third has not yet been fleshed out enough to really constitute a theory. But Wittgenstein explicitly argues against 1 and 2, and implicitly criticizes 3. He also offers a rival view, often called the view that meaning is use.

Against the Referential Theory

Much of the early sections of the Investigations is devoted to criticizing the idea that meaning is just naming.

In section 1, the referential theory is introduced with a quotation from Saint Augustine that seems to express it. W links the referential theory with an account of language learning: the referential theory seems to go along with the idea that all words are learned ostensively. Subsequent sections offer many criticisms, including these:

  1. ostensive definition only works relative to a substantial background: 28 on ostensively defining 'two'; 32 on Augustine (cf. Locke)
  2. Without such a background you can't tell what is being ostended: 33 color, shape; 73 leaf, schema; cf. Quine
  3. words that aren't names of things: 27
  4. nondenoting terms: Excalibur vs. 'Excalibur's meaning (we don't use language this way) - 39-40
  5. negative existentials: 79 on Moses. (Also rejects simple description theory in favor of a cluster-of-descriptions view)

Against the Ideational Theory

In §73, W shows that there is a natural transition from the referential view to the ideational view, the view that to understand what an expression means is to have the corresponding idea. (We need to remember that ideas are being thought of here as something like mental images. The term "idea" can also be used for abstract objects, but that would give us a version of the "propositional theory" rather than the "ideational theory.") Wittgenstein suggests that this too is a mistake.

  1. images are too specific: green vs. shade of green, specific shape vs. schematic shape
  2. images can't tell us how to use them. So the image couldn't be the whole story; we would also need rules for its usage. But in that case why not skip the image altogether?

Against the Propositional Theory

W doesn't really argue explicitly or in detail against the idea that meanings are abstract. But he does seem out of sympathy with it. See, for instance, §§93-95 about the idea that "propositions" are "queer".

Against Logical Atomism

W argues that, although "analysis" can be helpful, it is a mistake to think that it must lead to one single set of absolute primitives (as Russell thought, and as W had also held in the Tractatus). See §§90-92.

Wittgenstein's Positive View

Wittgenstein's positive views are expressed primarily by means of metaphors and questions; he never states a positive theory of meaning, and indeed seems to reject the very idea of a philosophical theory. Nevertheless, there are a number of positive suggestions.

  1. Meaning as use. §43: "For a large class of cases -- though not for all -- in which we employ the word "meaning" it can be defined thus: the meaning of a word is its use in the language."
  2. Language is like a game. This analogy is used from §3 on, as W calls his example languages "language games." He uses the game analogy to suggest a number of things about language:
    (a) no essence, nothing in common to all language games (§65)
    (b) there are rules of language use, as there are rules of games. But this doesn't really explain all that much: the rules don't cover everything, and the rules themselves stand in need of interpretation (do we need rules for how to apply the rules?). see §§81ff, especially §86
  3. Family resemblances. Often the things a word is used for are tied together, not by necessary and sufficient conditions, but by "family resemblances" -- see §67, among many other places. (Notice that W uses the idea of a game both as a metaphor for language as a whole, and as a specific example to illustrate his views about meaning.)

Problems for the View

Lycan mentions several problems for meaning-as-use views. The two most important seem to be:

  1. (Objection 3) How can one account for compositionality on this view?
  2. (Objections 4-5) Many rule-governed activities do not involve meaning. Similarly, we can understand use without knowing meaning. So it seems that there must be more to meaning than just rule-governed activity. What distinguishes language-games from non-linguistic games?

Last update: September 16, 2014
Curtis Brown | Philosophy of LanguagePhilosophy Department | Trinity University