The Absurdity of 'Thinking in Language'

Among the principal assumptions of major portions of philosophy in recent decades have been: (1) That philosophy somehow consists of (some sort of) logic, and (2) that logic is a study of and theory about (some sort of) language. There, of course, follows from these a third assumption: (3) That philosophy is a study of and theory about (some sort of) language--though this implication should not be taken as representing any phase of the historical development of recent philosophizing. Instead of listing these three points as assumptions, it would probably be more correct to regard them as categories or complexes of assumptions; or perhaps, more vaguely still, as ‘tendencies’ or proclivities of recent philosophical thinking. But precision of these points need not be put in issue here, as this paper does not seek any large-scale resolution of the problem area in question.

The aim here is to examine only one proposition which plays a role in the clearly existent tendencies referred to: Namely, the proposition that we think in or with language. I hope to show, first, that we do not always think in or with language; and then, second, that the very conception of thinking in or with language involves an absurdity. What implications this has for broader philosophical assumptions or tendencies will not be dealt with here, though the implications in question seem to me to be extremely important ones.

That human beings think in language is explicitly stated in such diverse places as ordinary newspapers, the more sophisticated popular magazines and journals, and serious discourse in the humanities and the social sciences, as well as in the technical writings of philosophers. To prove this broad range of consensus would be idle; but, in order to have the philosophical context clearly before us, we may give a few brief quotations. <126>

(1) Man, like every living creature, thinks unceasingly, but does not know it: the thinking which becomes conscious of itself is only the smallest part thereof. And, we may say, the worst part:--for this conscious thinking alone is done in words, that is to say, in the symbols for communication, by means of which the origin of consciousness is revealed. (Nietzsche, Joyful Wisdom, sub-sec. # 354)

(2) Let no one be contemptuous of symbols! A good deal depends upon a practical selection of them. Furthermore, their value is not diminished by the fact that after much practice, we no longer really need to call forth a symbol, we do not need to speak out loud in order to think. The fact remains that we think in words or, when not in words, then in mathematical or other symbols. (Frege, Mind, Vol. 73, p. 156)

(3) It is misleading then to talk of thinking as of a ‘mental activity’. We may say that thinking is essentially the activity of operating with signs. This activity is performed by the hand, when we think by writing; by the mouth and larynx, when we think by speaking; and if we think by imagining signs or pictures, I can give you no agent that thinks. If then you say that in such cases the mind thinks, I would only draw your attention to the fact that you are using a metaphor, that here the mind is an agent in a different sense from that in which the hand can be said to be an agent in writing. (Wittgenstein, Blue Book, pp. 6-7)

(4) ... The woof and warp of all thought and all research is symbols, and the life of thought and science is the life inherent in symbols; so that it is wrong to say that a good language is important to good thought, merely; for it is of the essence of it. (C. S. Peirce, Collected Papers, II, p. 129)

(5) Words only matter because words are what we think with. (H. H. Price, Aristotelian Society, Suppl. Vol. XIX, p. 7)

(6) Theorizing is an activity which most people can and normally do conduct in silence. They articulate in sentences the theories that they construct, but they do not most of the time speak these sentences out loud. They say them to themselves.... Much of our ordinary thinking is conducted in internal monologue or silent soliloquy, usually accompanied by an internal cinematograph-show of visual imagery.... This trick of talking to oneself in silence is acquired neither quickly nor without effort.... (Ryle, Concept of Mind, p. 27. See also pp. 282-83 and 296-97) <127>

(7)This helps to elucidate the well-known difficulty of thinking without words. Certain kinds of thinking are pieces of intelligent talking to oneself. Consider the way in which I ‘thinkingly’ wrote the last sentence. I can no more do the ‘thinking’ part without the talking (or writing) part than a man can do the being graceful part of walking apart from the walking (or some equivalent activity). (J.J.C. Smart, Philosophy and Scientific Realism, p. 89)

These quotations will suffice to establish the context within which philosophers speak of thinking in language (or with language). Many other quotations could be added from the literature.¹ It is not assumed here that the persons quoted all occupy the same position with reference to the relationship between thought and language. Yet it would be interesting to see what any of these thinkers, or others who suppose that human beings think in language, could save of their position from the critique which follows.

Uneasiness about the conception of thinking in or with language has been expressed by a number of writers, but only over limited aspects of it.² Here we shall consider arguments which purport to call the conception into question entirely and in principle. First, consider a reason for rejecting the view that we always think in language. It consists in the fact that thinking often occurs without the production, manipulation, or perception of sense-perceptible signs, without which there is no use of language. Such occurrences often provoke offers of ‘A penny for your thoughts.’

Thinking: Whatever we may decide to call them, and however it is that we are conscious of them, there are intentional states of persons, more or less fixed or fleeting, which do not require for their obtaining that what they are about or of be perceived by, or be impinging causally upon, the person involved. In order to think of³ Henry the Eighth, <128> of the first auto one owned, of the Pythagorean theorem, or of the Mississippi River, it is not required that they should disturb my nervous system. Such states (t-states) of persons are often called ‘thoughts’, especially in contrast with ‘perceptions’, and being in such a state is one of the things more commonly called ‘thinking’. One no more needs to be going through a change of such states in order to be thinking, than he needs to be changing his bodily position in order to be sitting or lying or sleeping. Rarely if ever--as is alleged in the case of mystic contemplation--are these t-states unchanging. Usually they flow, at varying rates, intermingled with person states of many sorts, governed by such transitional structures as inference, goal orientation, objective structures given in perception or in other ways, and elemental association of ‘ideas’, among others. In what follows, we shall use ‘thinking’ to cover both the single t-state and the flow of such states, without regard to how intermingled with other person states.

Language: Sense perceptible signs or symbols are an essential constituent of language. It is always false to say that language is present or in use where no signs are present or in use. And, whatever else a sign may be, it is something which is apprehendable via its sensible qualities. That is, it is something which can be either seen, heard, felt, tasted or smelled. Moreover, the use of language requires some level of actual sensuous apprehension of the signs which are in use on the occasion. (Confusion or distortion of this sensuous feedback can render a subject incapable of writing or speaking; and, of course, without perception of the sign-sequences emitted, one cannot understand the person emitting language.)

Now cases can be produced almost at will where thinking occurs without language being present or in use. This, of course, is something which everyone--including the proponent of thinking-in-language--very well knows. It is these cases which, together with the assumption that we always think in language, create what in (7) was called “the well-known difficulty of thinking without words.” If, as in (3), “thinking is essentially the activity of operating with signs,” then when there are no signs--and when, consequently, the means by which we produce, manipulate, or perceive signs are not functioning--we do have a difficulty. In fact, a difficulty so severe that it amounts to a proof that thinking is not essentially the activity of operating with signs, and that often we think entirely without language. One cannot operate with signs where there are no signs. <129>

As the above quotations indicate, the most common move made to save ‘thinking in language’ at this point is the shift to ‘silent soliloquy,’ as in (6), or to ‘pieces of intelligent talking to oneself,’ as in (7). These are latter-day shades of John Watson’s ‘sub-vocal language.’ Of course one can talk to oneself or write to onself. But talking and writing to oneself require the production and perception of sensuous signs just as much as talking and writing to another. The realization of this is what drives the thinking-in-language advocate to silent soliloquy or to nonvocal speaking--the written counterpart of which would be invisible writing. That is, they are driven to flat absurdities. A silent soliloquy--that is, silent speaking--is precisely on a par with a silent trumpet solo, for example, or silent thunder. A poet may say:

Heard melodies are sweet, but those unheard

Are sweeter; therefore, ye soft pipes, play on;

Not to the sensual ear, but, more endeared,

Pipe to the spirit ditties of no tone;...

(Keats, Ode to a Grecian Urn)

But there are in fact no unhearable melodies, no ears other than the “sensual,” no ditties of no tone.

What those who speak of silent discourse have in mind is, no doubt, the fact that interlaced with our thinking of or about things is a great deal of imaging of linguistic entities. (This is especially true of academics or intellectuals in general, because of their great concern with expression of thought. Probably an adequate phenomenology of thinking would exhibit great contrast between them and other classes of persons precisely at the relation between thinking and degree of activity in imaging linguistic entities and events.) But imaging a word is not using a word, any more than imaging a horse is using a horse. Moreover, imaging a word, phrase, or sentence is not producing or perceiving a word, phrase, or sentence any more than imaging a horse is producing or perceiving--or otherwise ‘having’--a horse. To image a linguistic sequence is not to have it in a special sort of place--the mind--nor is it to have a special sort of linguistic sequence. To image is to exemplify a certain sort of thinking or intentional state, and a sort which does have interesting relationships with other kinds of thinking. But there is no reason at all to suppose that all kinds of thinking necessarily involve or are accompanied by this kind of thinking (imaging) directed upon language segments. And if there were, it still would not follow that all thinking requires language, since this kind of thinking about language segments is not itself language at all. Nor does it require any <130> language present in order for it to come to pass, since intentional inexistence applies to mental events when language segments are the objects, as well as when sticks and stones and animals are.

Having considered a reason for rejecting the proposition that human beings always think in language, let us now consider whether they ever do. In fact, the difficulty is not, as Smart (above) and others have thought, in seeing how one can think without language, but in seeing how one would think with it. Thinking with or in language must consist in doing something with symbols, and so necessarily involves doing something to them--e.g., producing, altering, or perceiving them. If we would do something with the knife (e.g., cut the bread), we must do something to the knife, (e.g., clasp it in our hands). But, as we have seen, thinking occurs where nothing at all is being done to or with signs, there not being any signs in these cases. The power or act of having or changing t-states--that is, the power or act of thinking--is, then, not a power or act of having or altering linguistic symbols. (It is not, in fact, a power of doing anything with or in anything at all. The profound difference in kinds of powers and acts involved here is what Wittgenstein calls attention to in the last sentence of (3) above.) Thought is, of course, practical, in that it exercises an influence upon, or makes some difference in, the world of sense particulars. But it alone is not capable of acting with the sorts of particulars used in linguistic behavior as its immediate instruments. It is just this incapacity which makes it impossible for the advocates of thinking-in-language to give any account of the mechanisms or the ‘how’ by which the words in which we, allegedly, think are produced, manipulated, and gotten rid of--though they must be produced (or stored and hauled out), manipulated, and, in some sense, gotten rid of, if we are to think with and in them as our tools or instruments.

Merely to ask the question of how, in detail, this is done in the course of thinking reveals, I believe, the absurdity of ‘thinking in language’. Mere thinking can do nothing to signs which might be used in a language, and hence it can do nothing with such signs, or in the act of modifying the conditions of such signs. It is absurd to suppose that one can do x with y without in some way bringing about a change in the condition, state, relations, or properties of y. It is this and only this that I put by saying that it is absurd to suppose that one can do something with y while doing nothing to y.

If it is replied that, of course, the mind or thought does not do these things, but that when we write, speak, hear, see, and otherwise relate to actual words in the actual employment of language, we then are thinking, with bodily parts managing the symbols involved, then it <131> must be pointed out that, while we may indeed also be thinking in such cases, we are not simply thinking. The total event here, to which language certainly is essential, is not thinking. Correct use of language can even occur, as has been pointed out by Wittgenstein, without the occurrence of any peculiarly relevant t-states. On the other hand, thinking does occur without the use of hands, mouth, ears, eyes, fingers in any appropriately relevant manner. Hence, what can only occur by the use of these is not the same as thinking, though it may somehow involve or influence thinking.

Smart remarks in (7) that, when he thinkingly wrote the sentence, “Certain kinds of thinking are pieces of intelligent talking to oneself,” he could “no more do the ‘thinking’ part without the talking (or writing) part than a man can do the being graceful part of walking apart from the walking.” This may be true of thinkingly writing the sentence (whatever that means). But it does not follow that one cannot think that certain kinds of thinking are pieces of intelligent talking to oneself without the use of language, though Smart clearly thinks that it does. Of course one cannot thinkingly write without writing. But that is nothing to the point of whether or not we can and do think with or without words. Also, the comparison to graceful walking is not apt. We do, as above shown, sometimes think without words or symbols, while no cases of grace without behavior are known.

Now it is very certainly true that some processes clearly involving thinking as described above depend for their occurrence upon linguistic behavior and the sensible signs which it involves, for example, the processes of learning algebra or the history of the Basques, or learning how to counsel emotionally upset persons. But it is to be noted that these are not themselves processes of thinking, but rather are extremely complex processes involving all kinds of events and entities other than language and other than thinking--e.g., feelings, perceptions, buildings, other persons, days and nights, books, and so on. None of these processes is a process of thinking; and for that reason alone it is invalid to infer from them that thinking is linguistic behavior, or that one thinks with language. What is essential to things or events of a certain sort must be shown essential to them taken by themselves, not in combination with many other things. With reference to the involved processes in question, it might be more appropriate (though it would still be wrong) to say--as some have said in recent years--that we live in or with language. Nevertheless, it is certain that some kind of dependence relation--probably similar to feedback mechanisms--exists between linguistic processes and their sensuous signs, on the one hand, and certain sequences of t-states on the other. What, exactly, this relation <132> of dependence is continues to be veiled by, among other things, a priori assumptions about what thinking and language must be and do. One such assumption is that which holds thinking essentially to be an operation with signs or symbols, or doing something with--or in--linguistic processes or entities.

The view that we (necessarily) think without language is, today, regarded as so outlandish as not to merit serious consideration. But this is not due to a lack of arguments to support it. My object here has been to focus upon certain arguments purporting to show the absurdity of thinking in language. The main points in these arguments are: Thinking does occur without any accompanying language whatsoever, and thus shows itself not to be a power or act of managing linguistic signs, once it is clear what such a sign is. Thinking, as distinct from behavioral processes involving it, can do nothing to signs or symbols, and hence can do nothing with them.