Although the proper formulation and assessment of Ludwig Wittgenstein's argument (or arguments) against the possibility of a private language continues to be disputed, the issue has lost none of its urgency. At stake is a broadly Cartesian conception of experiences that is found today in much philosophy of mind.
In §243 of Philosophical Investigations (1967; see also §256) Wittgenstein introduces the idea of a language in which "a person could write down or give vocal expression to his inner experiences -- his feelings, moods, and the rest -- for his private use. . . . The individual words of this language are to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language." In subsequent sections (according to some commentators, up to as far as §315) Wittgenstein criticizes the possibility of such a "private language," and this is where "the private language argument" is usually supposed to be located.
Following the main essay, suppose that Privatus speaks Privatish, a private language. §243 suggests that Privatish has two features:
(1) Privatish contains a referring expression n that refers to one of Privatus's "immediate private sensations" s.
(2) s "can only be known to" Privatus.
According to Wittgenstein Privatish has a third feature, which he apparently thinks follows from (1) and (2):
(3) n can only be understood by Privatus (and, hence, Privatish can only be completely understood by Privatus).
At this point three questions arise. First, what does (2) mean? Second, what sort of referring expression is n? Third, why is (3) supposed to follow from (1) and (2)?
By (2) Wittgenstein seems to mean that only Privatus can know whether he is having s. "The essential thing about private experience is . . . that nobody knows whether other people also have this or something else" (§272, see also §246). Of course, this conception of sensations is held by Wittgenstein's opponent (a defender of the possibility of a private language), not Wittgenstein himself.
As to the referring expression n, it is a name, not a description (e.g., "the private sensation caused by pinpricks") (see §§256-257). Not even Wittgenstein's opponent would accept that to understand a description that in fact refers to sensation s one has to know that it refers to s.
Is n a proper name of a token sensation, or is it a common noun referring to a type of sensation? If n refers to a token sensation, something occurring only in the mind of Privatus, then Wittgenstein's opponent looks exactly like Bertrand Russell in "The Philosophy of Logical Atomism" (1918/1956; see also Candlish 2004). According to Russell, "[i]n order to understand a proposition in which the name of a particular occurs, you must already be acquainted with that particular" (1918/1956, p. 204; see also Russell 1912, chapter 5). Since, on Russell's view, the only particular things with which one is acquainted are private items he calls sense-data (and, in addition, perhaps one's self), no two people can be acquainted with the same particular, and so no two people can understand a genuine name (as Russell puts it, a name in the "logical sense"). Hence, Russell thinks, if Privatus's name n refers to a token sensation, no one else can understand it.
However, it is clear that Wittgenstein takes the sensations in question to be types, not tokens (see, in particular, §258); accordingly, the name n is a common noun. But then Russell's views about acquaintance and understanding play no role in the justification of (3), for Russell holds that two people can be acquainted with the same property (or type), including properties of private objects. Thus, if Privatus is acquainted with a certain type of sensation s, that is no barrier, on Russell's view, to others also being acquainted with s.
So why does Wittgenstein think that (3) follows from (1) and (2)? His argument may be this: one cannot know that Privatus's name n refers to s, so one cannot know what n means, and hence one cannot understand it. But it is not obvious that knowledge that n refers to s is necessary for understanding n, or for successful communication using it: perhaps all that is required is that one believes that Privatus's name n refers to s. The upshot is that Wittgenstein's double characterization of a private language as one "which describes my inner experiences and which only I myself can understand" (§256) is contentious. (This point is due to Edward Craig ; for further discussion see Craig .)
Given that the two characterizations of a private language should be separated, it is probably better to use the first, leaving the second as a disputed consequence. Thus, a private language may be explained as one containing names for types of inner experiences, with the further stipulation that, if there are any inner experiences, no one knows whether others have the same types of inner experiences as him- or herself.
§258 contains the famous example of keeping "a diary about the recurrence of a certain sensation," and most commentators identify it as the core of the private language argument. The key move in this section is to cast doubt on whether the diary keeper can "impress on [himself] the connexion between the sign ["S"] and the sensation," and so "remember the connexion right in the future." Anthony Kenny points out, against some commentators, that "remembering the connexion right" does not mean that one correctly applies "S" to one's sensation, but that one remembers the meaning of "S" (1973, pp. 191-193).
Supposedly, there is no fact about the meaning of "S" for the diarist to remember because there is "no criterion of correctness." Here, there is little consensus on what the missing criterion amounts to, or whether its absence does indeed show that the diarist fails to attach a meaning to "S". For some representative examples of exegesis, see Malcolm Budd (1989, chapter 3), Stewart Candlish (1980, 2004), John V. Canfield (1991, 2001), Robert J. Fogelin (1987, chapter 12), P. M. S. Hacker (1986, chapter 9; 1990, 61-67); Colin McGinn (1997, chapter 4), David Pears (1988, chapters 13, 14, 15), Scott Soames (2003, chapter 2), and Crispin Wright (1986).
In Wittgenstein on Rules and Private Language (1982), Saul A. Kripke suggests (without unreservedly endorsing) a novel and exciting interpretation of the private language argument (see also Fogelin 1987, p. 241, n. 10). On this interpretation the main argument appears in the earlier long discussion of following a rule starting around §138. As Kripke observes, the conclusion of the private language argument is stated in §202, well before the argument's traditional location, "Hence it is not possible to obey a rule 'privately': otherwise thinking one was obeying a rule would be the same thing as obeying it." Kripke takes the sections following §243 to be a discussion of a purported counterexample -- namely, sensation language -- to the conclusion argued for earlier.
The argument Kripke extracts from Wittgenstein is in two parts. The first part purports to establish that there are no facts that make it the case that an individual (Jones, say) means something by a word (addition by "+," say). This conclusion is reached by canvassing all the plausible candidates for such meaning-constituting facts and finding them all wanting. A skeptical paradox ("our paradox" of §201) looms, "[N]o facts, no truth conditions, correspond to such statements such as 'Jones means addition by "+"'" (1982, p. 77). The second part offers a skeptical solution: "skeptical" because the paradoxical conclusion is embraced; a "solution" because sentences such as "Jones means addition by '+'" remain assertible, despite the lack of any "corresponding fact." And the account of why such sentences are assertible essentially involves a linguistic community, so that if Jones is "considered in isolation," he cannot be said to mean anything by his words. This is the most general sense in which a "private language" is impossible: An individual considered in isolation from other speakers cannot be said to speak a language (see Kripke 1982, pp. 109-110).
Most commentators have not endorsed Kripke's interpretation (in particular, see Baker and Hacker 1984, chapter 1; McGinn 1984, chapter 2). However, Kripke's Wittgenstein has become a philosopher of considerable interest in his own right.
Kripke's book revived interest in the issue of whether the private language argument and related material on rule following is supposed to exclude a Robinson Crusoe isolated from birth from speaking a language (discussion of this topic goes back to Alfred J. Ayer  and Rush Rhees ; see also Kripke [1982, p. 110, n. 84]). While the characterization of a private language in §243 seems to leave room for such a Crusoe, other sections, notably §198, suggest the opposite. Norman Malcolm (1986, 1989) offers a defense of the "community view," and is countered by Baker and Hacker (1990). The community view is rejected by most commentators; for further discussion and references, see Canfield (1996).
See also Rule Following.
Ayer, Alfred J. "Can There Be a Private Language?" In The Philosophy of Wittgenstein, edited by John V. Canfield. New York: Garland, 1986. Ayer's essay originally published in the journal Proceedings of the Aristotelian Society in 1954.
Baker, G. P., and P. Hacker. "Malcolm on Language and Rules." Philosophy 65 (1990): 167-179.
Baker, G. P., and P. M. S. Hacker. Scepticism, Rules, and Language. Oxford, U.K.: Basil Blackwell, 1984.
Budd, Malcolm. Wittgenstein's Philosophy of Psychology. London: Routledge, 1989.
Candlish, Stewart. "Private Language." Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta. Stanford, CA: Metaphysics Research Lab, Center for the Study of Language and Information, Stanford University, 2003 <http://plato.stanford.edu/archives/spr2004/entries/private-language/>
Candlish, Stewart. "The Real Private Language Argument." Philosophy 55 (1980): 85-94.
Canfield, John V. "The Community View." Philosophical Review 105 (1996): 469-488.
Canfield, John V., ed. The Philosophy of Wittgenstein, Vol. 9: The Private Language Argument. New York: Garland, 1986.
Canfield, John V. "Private Language." In Wittgenstein's Philosophical Investigations: Text and Context, edited by Robert L. Arrington and Hans-Johann Glock. London: Routledge, 1991.
Canfield, John V. "Private Language: The Diary Case." Australasian Journal of Philosophy 79 (2001): 377-394.
Craig, Edward. "Meaning and Privacy." A Companion to the Philosophy of Language, edited by Robert Hale and Crispin Wright, 127-145. Oxford, U.K.: Blackwell, 1997.
Craig, Edward. "Meaning, Use, and Privacy." Mind 91 (1982): 541-564.
Fogelin, Robert J. Wittgenstein, 2nd ed. London: Routledge and Kegan Paul, 1987.
Hacker, P. M. S. Insight and Illusion: Themes in the Philosophy of Wittgenstein, Rev. ed. New York: Oxford University Press, 1986.
Hacker, P. M. S. Wittgenstein: Meaning and Mind. Oxford, U.K.: Basil Blackwell, 1990.
Kenny, Anthony. Wittgenstein. London: Penguin, 1973.
Kripke, Saul A. Wittgenstein on Rules and Private Language: An Elementary Exposition. Oxford, U.K.: Blackwell, 1982.
Malcolm, Norman. Nothing Is Hidden. Oxford, U.K.: Blackwell, 1986.
Malcolm, Norman. "Wittgenstein on Language and Rules." Philosophy 64 (1989): 5-28.
McGinn, Colin. Wittgenstein on Meaning. Oxford, U.K.: Basil Blackwell, 1984.
McGinn, Marie. Routledge Philosophy Guidebook to Wittgenstein and the Philosophical Investigations. London: Routledge, 1997.
Pears, David. The False Prison, Vol. 2. Oxford, U.K.: Clarendon Press, 1988.
Rhees, Rush. "Can There Be a Private Language?" In The Philosophy of Wittgenstein, edited by John V. Canfield. New York: Garland, 1986. Rhees's essay originally published in the journal Proceedings of the Aristotelian Society in 1954.
Russell, Bertrand. "The Philosophy of Logical Atomism" (1918). In Logic and Knowledge: Essays, 1901-1950, edited by Robert Charles Marsh. London: George Allen and Unwin, 1956.
Russell, Bertrand. The Problems of Philosophy. New York: Henry Holt, 1912.
Soames, Scott. Philosophical Analysis in the Twentieth Century, Vol. 2. Princeton, NJ: Princeton University Press, 2003.
Wittgenstein, Ludwig. Philosophical Investigations, 2nd ed. Translated by G. E. M. Anscombe. Oxford, U.K.: Blackwell, 1967.
Wright, Crispin. "Does Philosophical Investigations I.258-60 Suggest a Cogent Argument against Private Language?" In Subject, Thought, and Context, edited by Philip Pettit and John McDowell, 209-266. New York: Oxford University Press, 1986.