Private Language Argument

The setup

Descartes thought he could doubt the existence of the world, his body and other minds, but not the existence of his own current thoughts. From Descartes through Locke, Hume, and into 20th-century empiricism, philosophers assumed the inner life was the most certain thing — and therefore the most natural starting point for an account of meaning. Words got their content by labelling inner ideas; communication worked because we each labelled similar inner ideas similarly.

Wittgenstein attacks this whole picture in Philosophical Investigations §§243–315. The target is not just Descartes but a generic Cartesian: anyone who thinks the meaning of a sensation-word is fixed by an inward act of attention to a private object. He defines a private language carefully at §243 — not the language of someone alone, but one whose words refer to "what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language."

The argument that follows is reconstructed (Wittgenstein does not lay it out in numbered steps) but the core moves are clear: the diary case at §258, the misleading-by-an-image diagnosis at §270, and the beetle-in-the-box at §293.

The diary case (§258)

Imagine I want to keep a diary about a recurring sensation. I have no public name for it, so I invent one: every time the sensation occurs I will write "S" in my diary. Today I write S. Tomorrow I have a feeling that seems similar — should I write S again?

Wittgenstein presses: what is the difference between actually having the same sensation and merely seeming to? In the public case, criteria settle it — a doctor checks symptoms; we both look at the apple again. In the strictly private case, there are no such criteria. There is only my impression that the new feeling matches the old. Whatever is going to seem right to me is right.

And that, Wittgenstein concludes, "only means that here we can't talk about 'right.'" Without a distinction between seeming right and being right, there is no rule, no correctness, no meaning. The diary entry is empty: a sound made into the void.

The beetle in the box (§293)

The vivid image arrives fifty paragraphs later. Suppose everyone had a box. Inside is something we each call a "beetle." No one can look into anyone else's box. Each person says she knows what a beetle is only by looking inside her own.

It is, Wittgenstein says, possible that the boxes contain different things — one a stone, another nothing at all, a third a real beetle — and the language-game would be unaffected. The thing in the box plays no role in the public use of "beetle." If the inner object cannot make a public difference, then it is not what fixes the meaning of the word. The object "drops out of consideration as irrelevant."

The image generalises immediately to pain, anger, after-images, qualia. The Cartesian had thought private sensations were the bedrock of meaning. The beetle shows they are doing no work.

Theories of mental-term meaning compared

Theory	What fixes meaning of "pain"	Champion(s)	Strength	Weakness
Cartesian / inner ostension	Speaker pointing inward at a private quale	Descartes; Locke; sense-data theorists	Matches first-person feel	Falls to private-language argument
Behaviourism	Behaviour patterns of pain-bearer	Watson; Skinner; Ryle (modified)	Public, observable	"Super-Spartan" can suppress all behaviour
Identity theory	Brain state (C-fibres firing)	Place 1956; Smart 1959; Armstrong	Naturalistic, scientific	Multiple realisability objection (Putnam)
Functionalism	Causal role between inputs, other states, outputs	Putnam 1967; Lewis 1972	Allows multiple realisations	Inverted-spectrum and absent-qualia worries
Language-game / criterial	Public criteria + form of life	Wittgenstein; Hacker; Kenny	Handles both first- and third-person uses	Critics say "criterion" is under-defined
Kripkean sceptical solution	Communal agreement in dispositions	Kripke 1982	Confronts rule-following paradox head-on	Threatens to make truth a popularity contest
Higher-order thought (HOT)	Mental state + thought about that state	Rosenthal 1986; Carruthers	Naturalistic story for consciousness	Doesn't avoid the regress in private-ostension

A worked example: the asymmetric pain

Take a clear case. I burn my hand on a stove and say "ouch — that hurts." A child watching learns the word. The Cartesian says: I associate the word with the inner sensation; the child guesses (because she has had similar experiences) that I am referring to her kind of inner thing.

Wittgenstein points out the actual learning route. Children do not learn "pain" by introspection. They are trained: they cry, an adult says "poor thing — does it hurt?", they get comforted. The word is taught as a partial replacement for, and elaboration of, the natural pain-behaviour. By the time the child can articulate "my finger hurts," she has been embedded in a public practice that makes the utterance correct or incorrect by external standards — adults can check the wound, watch the wincing, ask questions.

The first-person use is not founded on an introspective baptism. It is the trained next move in the public game. This explains why we can be wrong about our own sensations (we sometimes say we're fine when our face is grey with pain) without making sensation-talk meaningless: the criteria are public, even when the avowals are first-person.

Kripke's reframing — the rule-following problem

Saul Kripke's 1982 book Wittgenstein on Rules and Private Language repositioned the argument. For Kripke the centre is §§138–242 (rule-following) rather than §§243–315 (sensation). The sceptic asks: what fact about your past use of "+" makes it the case that, by "plus," you meant addition rather than "quaddition" — a deviant function defined to agree with addition on all numbers below some bound and yield 5 thereafter? Any past use is consistent with both functions.

No mental image, no felt understanding, no dispositional fact (you have only ever computed finitely many sums) determines that you meant the infinite function rather than the deviant one. Kripke calls this "the most radical and original sceptical problem that philosophy has seen to date." His "sceptical solution": meaning is fixed by communal agreement — your community would correct you if you said 5 + 7 = 5 — not by any fact about you.

Critics (most influentially McDowell and Wright) object that the communal-dispositions story collapses truth into majority opinion. The literature on "Kripkenstein" is now larger than the original passages.

Counterarguments

• Ayer's Robinson Crusoe (1954). A.J. Ayer argued that a person stranded on an island from birth could in principle invent a private vocabulary by re-using the same sign for the same kind of sensation. Critics reply that Ayer's Crusoe still relies on memory acting as a public-like criterion — and the diary case shows memory cannot play that role.
• Strawson's sympathy + pressure (1954). P.F. Strawson's long review of the Investigations agreed with the conclusion but pressed Wittgenstein on whether the notion of "criterion" was clear enough to do the work. The criterion debate is still active.
• Functionalist pre-emption. If pain just is a functional role, the meaning of "pain" is fixed by that role — no inner ostension needed. Wittgenstein could be read as offering the criterial version of this story before functionalism had a name.
• Phenomenal-concept defenders. Block, Chalmers and others argue that we have direct phenomenal concepts of qualia that don't reduce to functional or behavioural roles. These "phenomenal concepts" are exactly what Wittgenstein denies are coherent.
• Conservative reading. Some commentators (notably Cora Diamond) read Wittgenstein as therapeutic rather than thesis-asserting: the argument doesn't prove a private language is impossible; it dissolves the urge to think we needed one.

Variants

• The community view (Malcolm, Kripke). A real language requires a community of speakers; even Crusoe couldn't have one alone.
• The individualist view (Baker & Hacker). A solitary speaker could have a language if she has stable, checkable criteria — what she can't have is a strictly inner one with no criteria at all.
• The therapeutic view (Diamond, Conant). Wittgenstein is not arguing that a private language is impossible; he is showing that the very idea is a confused one we can be talked out of.
• Kripkean view. The argument is fundamentally about rule-following, not sensation; the sensation case is a particularly vivid instance.

Common confusions

• Wittgenstein denies that pain exists. No. §304 explicitly: "It is not a Something, but not a Nothing either." He denies that pain's reality is what fixes the meaning of "pain."
• The argument is anti-introspection. No. We can introspect; we just can't bootstrap a language out of pure introspection.
• It's about secret codes. No. A code can be cracked and translated; a Wittgensteinian private language could not in principle be translated, because there's nothing publicly available to translate against.
• Crusoe alone refutes the argument. Crusoe inherited a public language from his society; he is not the relevant counterexample. The relevant case is a creature raised entirely in isolation with only inner objects to label.
• It's identical to behaviourism. Wittgenstein rejects behaviourism (§307 mocks the suggestion). Behaviour is a criterion, not a constituent, of inner states.
• The argument is conclusive. Seventy years of literature say it is contested. Even sympathetic readers (Hacker, Kenny) reconstruct it differently.