332. A. J. Ayer to Harrod , 2 December 1933 [a]

[The exchange continues at 336 ]

Christ Church, [Oxford]

2 December 1933

Dear Roy,

.1. [1] I share Russell's and your view that the propositions of physics are "ultimately about" sense-data. [2] I do not think that sense-data can be resolved into the fundamental entities of atomic physics. I think now that my use of the word physical in my article [3] was misleading. Perhaps "material" would have been better. My contention was that all propositions apparently about mental events in other minds are really about physical events, in a sense of physical in which "the fire is burning" or "your sister is shivering" would be descriptions of physical events. I maintained that unless we interpreted propositions about other minds in that way they would have to be regarded as meaningless.

.2. This follows from my assumption that the meaning of a proposition consists in its method of verification, [4] with the corollary that (prima facie) propositions which are in principle unverifiable are meaningless--not genuine propositions. I think the sense in which I am using the word meaning here is such that it is not an empirical postulate but a tautology. I should not say that this was even one of the most common senses in which the word meaning was used: but I think it is the same in which philosophers use it, who admit as I think you do that the sole business of philosophy is to analyse the meaning of expressions. [5] You may challenge me to <prove> this. I can only refer you to such a "paradigm of philosophy" as Russell's theory of descriptions and suggest that what Russell is there doing is considering how such propositions as "Scott is the author of Waverley" are in fact verified--only he does not push it far enough. [6] And if you reject my criterion of meaning--as the term is used in critical philosophy--I shall challenge you to produce another.

.3. You seem to suggest that even if my criterion be accepted, it is still not necessary to hold that all propositions about other minds [7] either are nonsensical or describe material events--unless the notion of verifiability be interpreted in an unjustifiably narrow fashion. Now I grant you that a proposition whose truth observations can render probable is <not> nonsensical. Throughout [b] my paper I was using the word verification in a wide sense, the sense in which a law is said to be verified by any fact which confirms [c] it. I do not think that any empirical proposition can be conclusively verified. No finite number of observations can establish the truth of an empirical proposition--even a singular proposition--with certainty. I regard a proposition as certain only when it is formally true.

.4. But if proposition about other minds mean what you wish them to, I deny that their truth is <even> in principle confirmable in experience. You might appeal to facts of co-consciousness--Sally Beauchamp & so on [8] --but in such cases there are not two minds but only one unusual mind. This by definition. And I am sure the argument from analogy will not serve you. An argument from analogy can only be used to render probable the existence of a state of affairs which could in principle be observed but in practice--through some accident of circumstance--can not. It cannot be used to render probable a proposition describing a state of affairs which is in principle unobservable. Because such a (prima facie) proposition would if you accept my criterion be nonsensical. [9] The argument from analogy only comes in as explaining why I use a similar form of words in referring to such very dissimilar "objects" as your headache and mine. I observe certain bodily motion of the body here when I feel a headache. I observe similar motions of the body over there and by association use the word headache. I do not infer the existence of a headache over there--because one can not be said to infer--even incorrectly--the truth <or> the probability of a nonsensical assertion. Psychical states are not here or there--but nowhere or everywhere--I therefore say, "there is a headache", "there is anger"--rather than "I have a headache" or "I am angry". [10]

.5. For me the problem is not to prove the existence of other minds. [11] I accept Moore's view that we are concerned [d] with the analysis of propositions we unhesitatingly accept as true, not with justifying them. [12] I believe the proposition that there are other minds to be as nearly certain as any empirical proposition can be. But I analyse it in a way that you find psychologically unsatisfying. I insist however that if you accept my premiss about verification, then your "other" minds are unobtainable. All [e] the propositions you attempt to make about them will be nonsensical.


Freddie Ayer.

  1. 1. The discussion regards A. J. Ayer, "The Case for Behaviourism", The New Oxford Outlook 1:2, November 1933, pp. 229-42. This letter seems to be a reply to a letter not found, although this is not explicitly cited.

    2. See for instance B. Russell, The Problems of Philosophy, London: Williams & Norgate, 1912, in particular chapters 1-3.

    3. Ayer's argument begins with the characterization of two different theories of behaviorism in terms of their interpretations of behavior in terms of physical language or as physical phenomena: the "philosophical or logical theory of Behaviourism" is

    • like all genuine philosophical theories, a theory about the meaning of a certain class of propositions. It asserts that all propositions ordinarily expressed in the language of psychology may equally well be expressed in the language of physics [...]. It asserts, in other words, that all the propositions of psychology are about certain physical events, namely certain occurrences in the human body (p. 230).

    The scientific theory of behaviourism, on the other hand,

    • is for its part a theory about the way in which a certain class of phenomena may be predicted or explained. It asserts that all human and animal behaviour, all the so-called phenomena of consciousness, can be explained in terms of physical stimuli and the physical reaction of the bodily organism to them (pp. 230-31).

    Ayer's case for logical behaviourism rests on the necessity, for psychology, of "ridding itself of metaphysical accretions" in order to deserve the title of science (p. 242).

    4. Ayer, "The Case for Behaviourism", p. 233.

    5. "Philosophy is not concerned with the truth or falsehood of propositions but only with their meaning" (Ayer, "The Case for Behaviourism", p. 231).

    6. Russell's theory of descriptions was first expounded in "On Denoting" (originally published in Mind, 1905, now in The Collected Papers of Bertrand Russell. Vol. IV: Foundations of Logic, 1903-05, London: Routledge, 1994, pp. 414-27), where the proposition about the author of Waverley is discussed.

    7. Harrod took up again the subject of the behavior of the "Other Selves" as a chapter of an unpublished book, to be titled The Known and the Unknown, written in 1940/41 (Harrod to Macmillan, 11 November 1942, in MaP. The manuscript, untitled and undated, is housed in HP V-70, with some typescript chapters in HP V-71. For further details and a reconstruction of the table of contents see D. Besomi, "Harrod and the `Time-lag Theories of the Cycle'", in G. Rampa, L. Stella and A. P. Thirlwall, Economic Dynamics, Trade and Growth. Essays on Harrodian Themes, 1998, pp. 141-42).

    8. The clinical case of Sally Beauchamp was discussed by M. Prince as an example of multiple personality: The Dissociation of Personality. A Biographical Study in Abnormal Psychology, London: Longmans, 1906.

    9. Ayer argues that while the usual answer to the problem of the existence of other minds than one's own is usually based on analogy, analogy can only render probable propositions that could be in principle established by direct observation, but not nonsensical (meaningless) conclusions: "The Case for Behaviourism", pp. 232-34.

    10. Ayer, "The Case for Behaviourism", p. 234.

    11. "In practice we do not go through the process of inferring `the existence of other minds'. We intuit it." ("The Case for Behaviourism", p. 235).

    12. G. E. Moore, "A Defence of Common Sense", in J. H. Muirhead, Contemporary British Philosophy. Personal Statements, second series, London: Allen & Unwin, 1925, pp. 193-223.

    1. a. ALS, seven pages on four leaves, with envelope initialled A. J. A. and addressed to R. F. Harrod, Esq., Ch. Ch., marked "Probability" in Harrod's post-war handwriting. In HP IV-22-65/2.

      b. Ms: there is no full stop, and «throughout» begins with small "t". But the new page begins with a word crossed out, whose first letter was capitalized. This suggests that the sentence may have been intended to stop here, which also makes grammatical sense.

      c. Ms: dashed underlining.

      d. Ms: «we are in such a concerned».

      e. Ms: «In all».

Welcome page

top of page

Return to index of this section

Go to previous page

Go to next page