Language Games (Wittgenstein)

From picture theory to language games

The young Wittgenstein had given philosophy a final answer. The Tractatus Logico-Philosophicus (1921) declared that a meaningful proposition was a logical picture of a fact: names stood for objects, the structure of the sentence mirrored the structure of the world, and anything else was nonsense to be passed over in silence. The book closed with the line, "Whereof one cannot speak, thereof one must be silent." Wittgenstein, satisfied that he had solved the problems of philosophy, retreated to teach village schoolchildren in Lower Austria.

By 1929 he had returned to Cambridge and concluded he was wrong. The work he produced over the next two decades — published posthumously in 1953 as the Philosophical Investigations — opens by quoting Augustine's Confessions on how he learned language: adults pointed at things, said the names, and Augustine connected the sound to the object. Wittgenstein treats this as the cleanest expression of the picture theory and spends the next 600 paragraphs taking it apart.

The replacement metaphor arrives in §7. Imagine a primitive language — only four words. A builder calls "Slab!" and his assistant brings a slab. "Block!" brings a block. The whole language consists of orders and the actions they trigger. Wittgenstein notes that we cannot understand "slab" by staring at the word or even at slabs: we have to watch the routine on the building site. The word, the call, the reaching motion, the act of carrying — together these form a language game.

By §23 Wittgenstein lists what some of these games are: giving orders, describing an object, reporting an event, speculating, telling a joke, asking, thanking, cursing, greeting, praying. "There are countless kinds" of sentence, "countless different kinds of use of what we call 'symbols', 'words', 'sentences.'" The picture theory had assumed all sentences shared one job — picturing facts. Investigations argues there is no single job; there is a busy multiplicity.

Meaning is use

The most-quoted line of the book sits at §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." The qualification "for a large class" is important and frequently dropped. Wittgenstein is not reducing all meaning to use; he is dislodging the assumption that meaning is a private mental image or a Platonic referent waiting behind the word.

The diagnostic move is to ask, when philosophers feel puzzled by a word, whether the word is doing its everyday job or whether it has been wrenched out of context. "What time is it on the sun?" uses the everyday word "time" in a setting where its everyday game (clocks, schedules, sunrise) does not apply. The question feels deep but it is, on Wittgenstein's diagnosis, a sentence on holiday — language gone idle, the wheels spinning without engaging anything.

Family resemblance

If meaning is use, what holds a word together across all its uses? §§65–67 raises the obvious objection: surely all the uses of "game" must share something — the essence of game-hood — or we couldn't call them all games. Wittgenstein answers: don't assume; look.

Look at board games, card games, ball games, Olympic games, ring-around-the-rosy. Some involve winning and losing (chess, tennis); some don't (catch, ring-around-the-rosy). Some are competitive (boxing); some cooperative (a puzzle). Some involve skill (chess); some chance (snap). What is common? "Don't say there must be something common, or they would not be called 'games' — but look and see whether there is anything common to all." What you find, Wittgenstein writes, is "a complicated network of similarities, overlapping and crisscrossing." He calls this family resemblance (Familienähnlichkeit).

The implications are large. Many philosophical concepts — knowledge, art, religion, justice — have been hunted for centuries on the assumption they must have one essence. If concepts can be held together by overlapping similarities instead, the hunt was misguided.

Forms of life

Language games sit inside forms of life (Lebensformen). The phrase appears only five times in the book but does heavy work. To imagine a language is to imagine a form of life (§19): the shared practices, biological needs, social customs, and reactions that make particular language games make sense. Money-talk requires a form of life with property and exchange. Pain-talk requires creatures that wince and grimace.

The famous remark at p.223 — "If a lion could speak, we could not understand him" — is a corollary. Even if we could parse the lion's words, his form of life is too different from ours for the words to latch onto recognisable practices.

Wittgenstein's later view vs other theories of meaning

Theory	Meaning is...	Champion(s)	Headline strength	Headline weakness
Augustinian / referential	The object the word names	Augustine; early modern	Simple, intuitive for nouns	Fails for "and," "not," "hello," pain
Picture theory (early Wittgenstein)	A logical picture of a fact	Wittgenstein 1921; Russell	Crisp account of truth conditions	Forces all sentences into one mould
Verificationism	The method of verification	Vienna Circle (Carnap, Schlick)	Sharp tool against pseudo-science	Can't verify itself; falls to Quine
Truth-conditional (Davidson, Tarski)	The conditions under which it's true	Tarski 1933; Davidson 1967	Compositional, formal-friendly	Can't handle non-declarative sentences
Use theory (later Wittgenstein)	The role it plays in a language game	Wittgenstein 1953; Ryle	Handles the full multiplicity of speech	Use of what? Critics say "use" is too vague
Causal-historical	A name's reference is fixed by a baptism + chain	Kripke 1980; Putnam	Explains rigid designators, natural kinds	Limited to proper names and natural kinds
Inferential role	The inferences it licenses	Brandom 1994; Sellars	Connects meaning to reasoning	Still owes an account of empirical content

A worked example: the builders

Investigations §2 sketches the simplest possible language. A builder, A, and his assistant, B. Stones lie around: blocks, pillars, slabs, beams. A calls out the words; B brings the stones. There are exactly four words. There is no grammar, no syntax, no inner mental act of meaning.

The Augustinian theorist will say: each word names a kind of stone, and B understands by associating the sound with a mental image. Wittgenstein asks what we would say if B's associations differed — if when A said "Slab" B pictured a block but still brought the slab. We'd say B understands "slab." The mental image is idle; what matters is that B brings the right stone. Now run the case the other way: B pictures a slab but reliably brings a beam. We'd say B does not understand. The mental image, again, is irrelevant; the use is everything.

Notice what falls out for free. There is no "hidden" private fact about whether B has the same concept as A. There is only the public game and B's competence at it. This is the seed of the private-language argument that follows in §§243–315.

Objections and replies

• Doesn't this make meaning behavioural? Wittgenstein resists the label. Behaviourism reduces meaning to bare overt behaviour; language games are richer — they include rules, customs, training, the surrounding form of life. He calls it "a grammatical investigation," not a psychological one.
• If meaning is use, anything goes. No: not every use is correct. Rules constrain. But the rules are themselves embedded in practice, not in a private inner act of grasping.
• Family resemblance is too loose. Critics (e.g., Bambrough 1960; Fodor) have argued the criterion permits any pair of objects to count as members of any concept. Defenders reply that the network of similarities is not unrestricted — it is anchored in a tradition of use.
• Kripke's sceptical reading. In Wittgenstein on Rules and Private Language (1982) Saul Kripke read §§138–242 as posing a paradox: nothing in your past use of "plus" determines that you meant addition rather than "quaddition" (a deviant function that agrees with plus on small numbers but answers 5 for sums above some threshold). The community-dispositions reply Kripke offers is widely debated.

Variants and developments

• Ordinary-language philosophy (Oxford, 1950s). Gilbert Ryle's Concept of Mind (1949) attacks Cartesian dualism using a Wittgensteinian method. J.L. Austin's How to Do Things With Words (1962) develops speech-act theory: utterances do things (promise, name, baptise) as well as describe.
• Therapeutic Wittgenstein (the "new Wittgensteinians"). Cora Diamond, James Conant and Alice Crary read Wittgenstein as offering not a positive theory but a therapy that dissolves philosophical confusion.
• Resolute Wittgenstein vs. interpretive. The two camps differ on whether the Tractatus itself was meant as nonsense to climb past or contained substantive metaphysics.
• Continental reception. Stanley Cavell's The Claim of Reason (1979) takes Wittgenstein into ethics, scepticism and tragedy.

Common confusions

• "Meaning is use" reduces meaning to behaviour. No — the qualification at §43 and the surrounding paragraphs locate use within rule-governed practices, not bare stimulus-response.
• Family resemblance means anything-goes relativism. The web of overlapping similarities is anchored in a tradition; arbitrary pairs do not become a concept just by stipulation.
• The Tractatus and the Investigations are unrelated. Wittgenstein himself wrote in the Investigations preface that the new ideas were best seen against the background of the old.
• Wittgenstein denies that words refer. He denies that reference is the universal explanation of meaning. Reference is one game among many.
• Forms of life means cultures. They include cultural practices but go deeper — biology, custom, the agreement in reactions that lets training take hold (§242).
• The book is a positive theory. Wittgenstein insists he is not advancing theses but dissolving puzzles. Whether he succeeds in avoiding theory is itself a contested reading.