Karl R. Popper

First published in Analysis, 15, 1955, as a reply to Professor Wilfrid Sellars. Reprinted in Karl R. Popper, Conjectures and Refutations: The Growth of Scientific Knowledge (New York: Basic Books, 1962).
Transcribed into hypertext by Andrew Chrucky, Sept. 12, 1997.

I am very grateful to Professor Wilfrid Sellars for bringing{1} my paper 'Language and the Body-Mind Problem',{2} to the attention of philosophers, and even more for his kindness in describing it as 'challenging', and as 'telling, if uneven'. Of its unevenness nobody can be more aware than I. I think I am more sensitive to it than Andersen's princess was to the pea. And although I am inclined to count its six leaves among my scanty laurels, I could not rest on them even if I wished to. But the small hard peas which bother me and keep me awake at night seem to have been well hidden, and in a spot far removed from Professor Sellars' two largish lumps of stuffing which I believe, are not at all hard to smooth out.


As to the first lump, Professor Sellars, after quoting me correctly at some length, proceeds to 'focus attention', as he puts it, 'On the statement [Popper's statement] quoted above, that ". . . if the two languages are not translatable, they deal with different sets of facts".' And Professor Sellars then goes on to say that a 'fact' may be either a 'descriptive fact' or else something like 'the "fact" that we ought to fulfil our undertakings', which I may be permitted to call a 'quasi fact'. And he says that my argument would be valid if only it would contain 'the premise that both languages in question have the business of describing', i.e. of stating 'descriptive facts'.

Now I agree with every word of this but I completely fail to see its relevance: in focusing attention upon one statement, Professor Sellars, understandably enough, got its context out of focus.

For (a), the premise which, according to Professor Sellars, makes my argument valid, was clearly enough indicated in my own argument which therefore is itself valid, according to Professor Sellars. Moreover, my argument has the form of a reductio ad absurdum of the 'two language theory', and the premise correctly demanded by Professor Sellars is not mine but part of that theory. It is, indeed, referred to in my argument as part of the 'two language solution' -- of 'the view that . . . the statements of physics and of psychology are . . . two ways of talking about the same facts' (which clearly indicates that these 'facts' are 'descriptive facts' in Professor Sellars' terminology). My own contribution consisted, simply, in pointing out that, once the two languages (of physics and of psychology) are admitted not to be translatable into each other, they cannot any longer be said to talk about the same facts, and must be admitted to talk about different facts -- where' facts' means whatever the two-language theorists meant when they said that physics and psychology talked about the same facts.

Thus the problem of 'quasi facts' simply does not arise.

All this can be verified by reading more closely the passage from my paper which Professor Sellars himself quotes at the beginning of his paper: it is the passage which gets out of focus once he focuses attention on part of it. (There is a not very important misquotation -- 'set' instead of 'kind' -- in the focused passage.)

So no hard core, no difference of opinion as far as I can see, underlies Professor Sellars' first lump -- although I seem to differ with him about the relevance of his comments.


Now to smooth out the second lump. 'In the later sections of his paper', Professor Sellars writes, 'Professor Popper makes a telling, if uneven, defense of the thesis that aboutness or reference cannot be defined in Behaviourese.' (Professor Sellars himself believes in the truth of this alleged thesis of mine.) I must confess that I was surprised when I read this. I was not aware of having ever tried to defend anything of the kind. It happens to be one of my oldest convictions that a thesis of the kind here attributed to me -- that such and such cannot be defined in somebody's language -- is nearly always irrelevant. (It is not irrelevant, of course, if the opponent's thesis was one out definability. Definability may be interesting in certain contexts, but to say a term is not definable never implies that it cannot be legitimately used; for it may be legitimately used as an undefined term.) There was no need for me to read through my paper in order to be sure that I never maintained anything like the 'thesis' attributed to me by Professor Sellars. But to make doubly sure I did read through my paper and I found no trace of such a thesis on definability. And to make trebly sure, I herewith publicly recant any theory I may ever have advanced based upon the thesis attributed to me by Professor Sellars: not because the thesis is false (I agree with Professor Sellars that it is true, and I even agree that my arguments might be used to support its truth -- which may perhaps explain the misunderstanding) but because I should hate the idea of philosophizing with the help of arguments about non-definability.

Professor Sellars goes on to say 'And he [Popper] is surely right [in holding the thesis I have just repudiated]. However, at this stage he [Popper] tacitly adds the premise " 'E is about x' is a descriptive assertion".'

It is hard for me to check whether or not I have added this premise tacitly at this stage, since 'this stage' is not indicated by Professor Sellars -- or only indicated with the help of a reference to that alleged thesis of mine which I fail to find anywhere in my paper. (I may here warn readers that seven of the passages in quotation marks in this second part of Professor Sellars' paper are not quotations from my paper, as some might think. Two others, 'Name relation' and 'Causal-physicalistic', did occur in my paper, but the former hyphened, the latter unhyphened.)

If, however, I have somewhere 'tacitly' and unconsciously added the l premise which Professor Sellars says I have added (I cannot detect any trace of it) then I wish. again, to recant. For I am in complete agreement with Professor Sellars' thesis that if a statement A says that another statement E is about something, then A usually does not, to use Professor Sellars' words play 'the same sort of a role as "The Moon is round" '. A need not be, and usually is not, 'descriptive' in the same sense as the statement about the moon (although it may be: 'What has your last lecture about?' -- 'It was a lecture about probability', is an instance of descriptive usage).

I also agree entirely with Professor Sellars concluding remark that from the fact, and it is a fact, that what Professor Popper calls the "name relation" (paragraph 5 ff.) is not definable in "causal-physicalistic" terms, we cannot conclude to the truth of Dualism'. Exactly. This is why I never said anything about definability. Indeed, had I no stronger arguments in favour of my dualistic faith than this completely irrelevant fact (for I agree that it is a fact, though completely irrelevant), then I should be ready -- nay, most anxious to give up dualism. As it happens, my arguments were quite different. They were about{3} the possible scope of deductive physical theories rather than about definability; and my thesis was that 'no causal physical theory of the descriptive and argumentative functions of language is possible'.

I wish to make it perfectly clear that I have no objection whatever to Professor Sellars' thesis that a statement such as 'E is about x' is (ordinarily, or frequently) 'a device whereby we convey to the hearer how a mentioned expression is used, by using an equivalent expression'. Nor do I deny that this thesis of Professor Sellars' is relevant to my own thesis. All I wish to say here is that my thesis is not based on the argument about definability which Professor Sellars ascribes to me, if it were, I should retract it.


There is a remark on Professor Ryle's views in Professor Sellars' paper which seems wrong to me. Professor Sellars writes: 'I also agree that "the idea of a mutual translatability" of . . . mind talk and behaviour talk "had to be given up long since", in spite of Ryle's valiant efforts to the contrary '

To this I should like to say that I am not aware of the fact that Professor Ryle has ever held what I call 'the two languages theory'. How could he, believing as he does that the problem arises out of category-mistakes within the one natural language? It is not to him I was alluding in that place.

At the same time, it is perfectly true that I had Professor Ryle in mind when, in another paragraph of my paper, I tried to chow briefly that the theory of 'category-mistakes' is also untenable.

If I might here add to my arguments another, then I should say this. Assuming that, by the usages of our language, expressions naming physical states are put in a category different from that in which expressions naming mental states are put, I should be inclined to see in this fact an indication, or a suggestion (not more than this, to be sure), that these two categories of expression name entities which are ontologically different -- or in other words, that they are different kinds of entities. Thus I should be inclined (not more than this) to entertain the opposite conclusion to the one drawn by Professor Ryle although, admittedly, the premises would be insufficient for a formal derivation of the conclusion.

However, I am not prepared to grant the truth of this assumption, quite apart from my (and from Professor Smart's{4}) objections to arguments based upon the idea of category-mistakes. I find very many of Professor Ryle's analyses most illuminating, but I can only say that ordinary English very often treats mental states and physical states on a par with each other; not only where it speaks of a 'mental disease', of a 'hospital for the mentally sick' or of a man who is 'both physically and mentally well balanced', etc. (these cases might be dismissed as deriving from a philosophical dualism) but especially where we say: Thinking of sheep always helps me to fall asleep' or 'Reading Mr Smith's novels always helps me to fall asleep' (which does not mean 'training my eyes on one of Mr Smith's novels always helps me to fall asleep' and yet is completely analogous to 'taking bromide always helps me to fall asleep'). There are countless similar examples. They certainly do not establish that ordinary English words describing mental states and physical states always belong to the same 'category (Professor Ryle has succeeded in showing that they don't). But my examples establish, I think, that the words are often used in ways which are strikingly alike. The uncertainty of the language-situation may be illustrated by an example of Professor Ryle's.{5} He says, rightly, that a child who has just watched the parade of all the battalions, batteries, and squadrons, which constitute a division, makes a mistake (in the sense that he has not quite got the meaning of the words) when he then asks 'And when will the division come?' -- 'He would', Professor Ryle says, 'be shown his mistake by being told that in watching the battalions, batteries and squadrons marching past he had been watching the division marching past. The march-past was not a parade of battalions . . . and a division; it was a parade of the battalions . . . of a division.' This is absolutely true. But are there no contexts, of perfectly good English usage, in which battalions are treated on a par with divisions? Could there not be a parade of, say, one division and three battalions and two batteries? I can imagine that this might be an outrage to military usage (although a battle in which a division attacks a battalion is, I suppose, perfectly good military usage). But is it an outrage to ordinary English usage? And if not, can the mistake which the child undoubtedly committed be a category mistake ? And if not, do we not commit a category mistake (assuming such a thing exists) if we wrongly diagnose that the child's mistake was a category mistake ?


{1} By way of his 'A Note on Popper's argument for Dualism'. Analysis, 15, pp 23 f. [Back]

{2} Not 'Mind-body problem' as Professor Sellars writes. [Back]

{3} This is another instance of an about-statement A which describes an argument E. [Back]

{4} See his excellent brief 'Note on Categories' in the British Journal for the Philosophy of Science, 4, 1953, pp. 227 f. [Back]

{5} The Concept of Mind, p. 16 f. The example of the Colleges and the university is precisely analogous the foreigner who wants to see the university asks, of course, for a University building (perhaps one like the Senate House in London); and this building would be of the same category as the college buildings. Is it not therefore a category mistake to suggest that he has made a category mistake? [Back]